The Longleaf Pine Ecosystem Ecology, Silviculture, and Restoration the Longleaf Pine Ecosystem Shibu Jose Eric J

Home , Liberation cutting, Longleaf pine ecosystem, Sarracenia minor, Sarracenia psittacina

Jokela Jose The Longleaf Pine Ecosystem Ecology, Silviculture, and Restoration The Longleaf Pine Ecosystem Shibu Jose Eric J. Jokela Deborah L. Miller Editors S PRINGER S ERIES ON E NVIRONMENTAL M ANAGEMENT

BRUCE N. ANDERSON ROBERT W. H OWARTH LAWRENCE R. WALKER Series Editors Springer Series on Environmental Management Volumes published since 1989

The Professional Practice of Environmental Marine Debris: Sources, Impacts, and Solutions Management (1989) (1997) R.S. Dorney and L. Dorney (eds.) J.M. Coe and D.B. Rogers (eds.)

Chemicals in the Aquatic Environment: Advanced Environmental Problem Solving: Psychosocial Hazard Assessment (1989) Barriers to Adaptive Change (1999) L. Landner (ed.) A. Miller Inorganic Contaminants of Surface Water: Research and Monitoring Priorities (1991) Rural Planning from an Environmental Systems J.W. Moore Perspective (1999) F.B. Golley and J. Bellot (eds.) Chernobyl: A Policy Response Study (1991) B. Segerstahl˚ (ed.) Wildlife Study Design (2001) Long-Term Consequences of Disasters: The M.L. Morrison, W.M. Block, M.D. Strickland, and Reconstruction of Friuli, Italy, in its International W.L. Kendall Context, 1976–1988 (1991) R. Geipel Selenium Assessment in Aquatic Ecosystems: A Guide for Hazard Evaluation and Water Quality Food Web Management: A Case Study of Lake Criteria (2002) Mendota (1992) A.D. Lemly J.F. Kitchell (ed.)

Restoration and Recovery of an Industrial Region: Quantifying Environmental Impact Assessments Progress in Restoring the Smelter-Damaged Using Fuzzy Logic (2005) Landscape near Sudbury, Canada (1995) R.B. Shepard J.M. Gunn (ed.) Changing Land Use Patterns in the Coastal Zone: Limnological and Engineering Analysis of a Managing Environmental Quality in Rapidly Polluted Urban Lake: Prelude to Environmental Developing Regions (2006) Management of Onondaga Lake, New York (1996) G.S. Kleppel, M.R. DeVoe, and M.V. Rawson (eds.) S.W. Efﬂer (ed.)

Assessment and Management of Plant Invasions The Longleaf Pine Ecosystem: Ecology, (1997) Silviculture, and Restoration (2006) J.O. Luken and J.W. Thieret (eds.) S. Jose, E.J. Jokela, and D.L. Miller (eds.) Shibu Jose Eric J. Jokela Deborah L. Miller (Editors)

The Longleaf Pine Ecosystem Ecology, Silviculture, and Restoration

With 92 Illustrations Shibu Jose Eric J. Jokela Deborah L. Miller School of Forest Resources and School of Forest Resources and Department of Wildlife Ecology Conservation Conservation and Conservation University of Florida University of Florida University of Florida Gainesville, FL 32611 Gainesville, FL 32611 Milton, FL 32583 USA USA USA sjose@ufl.edu ejokela@ufl.edu dlmi@ufl.edu

Series Editors: Bruce N. Anderson Robert W. Howarth Lawrence R. Walker Planreal Australasia Program in Biogeochemistry and Department of Biological Keilor, Victoria 3036 Environmental Change Sciences Australia Cornell University University of Nevada [email protected] Corson Hall Las Vegas, NV 89154 Ithaca, NY 14853 [email protected] [email protected]

Library of Congress Control Number: 2005936717

ISBN-10: 0-387-29655-7 ISBN-13: 978-0387-29655-5 eISBN: 0-387-30687-0

Printed on acid-free paper.

C 2006 Springer Science+Business Media, LLC All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, Inc., 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identiﬁed as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights.

Printed in the United States of America. (TB/EB)

987654321 springer.com Contributors

Janaki R.R. Alavalapati, School of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611

John Blake, Savannah River Site, USDA Forest Service, New Ellenton, South Carolina 29809

William D. Boyer, Southern Research Station, USDA Forest Service, Auburn, Alabama 36849

Dale G. Brockway, Southern Research Station, USDA Forest Service, Auburn, Alabama 36849

J. Bachant Brown, The Nature Conservancy, Jay Florida Ofﬁce, The Gulf Coastal Plain Ecosys- tem Partnership of Jay, Florida 32565

Douglas R. Carter, School of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611

Vernon Compton, The Nature Conservancy, Jay Florida Ofﬁce, The Gulf Coastal Plain Ecosys- tem Partnership of Jay, Florida 32565

Ralph Costa, U.S. Fish and Wildlife Service, Clemson Field Ofﬁce, Department of Forestry and Natural Resources, Clemson University, Clemson, South Carolina 29634

Roy S. DeLotelle, DeLotelle and Guthrie, Inc., 1220 SW 96th Street, Gainesville, Florida 32607

Cecil Frost, Adjunct Faculty, Curriculum in Ecology, University of North Carolina, 119 Pot Luck Farm Road, Rougemont, North Carolina 27572.

Dean Gjerstad, School of Forestry and Wildlife Sciences, Auburn University, Auburn, Alabama 36849

J.C.G. Goelz, USDA Forest Service, Southern Research Station, Pineville, Louisiana 71360

v vi Contributors

James M. Guldin, Arkansas Forestry Sciences Laboratory, Southern Research Station, USDA Forest Service, Monticello, Arkansas 71656

Timothy B. Harrington, USDA Forest Service, Paciﬁc Northwest Research Station, Olympia, Washington 98512

Larry D. Harris, Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, Florida 32611

M. Hicks, The Nature Conservancy, Jay Florida Ofﬁce, The Gulf Coastal Plain Ecosystem Part- nership Jay, Florida 32565

Thomas S. Hoctor, Department of Landscape Architecture, University of Florida, Gainesville, Florida 32611

Alan W. Hodges, Institute of Food and Agricultural Sciences, University of Florida, Gainesville, Florida 32611

Don Imm, Savannah River Site, USDA Forest Service, New Ellenton, South Carolina 29809

Steven B. Jack, Joseph W. Jones Ecological Research Center, Newton, Georgia 39870

Rhett Johnson, Solon Dixon Forestry Education Center, School of Forestry and Wildlife Sci- ences, Auburn University, Andalusia, Alabama 36420

Eric J. Jokela, School of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611

Shibu Jose, School of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611

John S. Kush, School of Forestry and Wildlife Sciences, Auburn University, Auburn, Alabama 36849

Peter E. Linehan, The Pennsylvania State University, Mont Alto Campus, Mont Alto, Pennsylvania 17237

Jagannadha R. Matta, School of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611

D. Bruce Means, Coastal Plains Institute and Land Conservancy, Tallahassee, Florida 32303

Deborah L. Miller, Department of Wildlife Ecology and Conservation, University of Florida, Milton, Florida 32583

Robert J. Mitchell, Joseph W. Jones Ecological Research Center, Newton, Georgia 39870

W. Keith Moser, USDA, Forest Service, North Central Research Station, Forest Inventory and Analysis, St. Paul, Minnesota 55108 Contributors vii

W. Leon Neel, Joseph W. Jones Ecological Research Center, Newton, Georgia 39870

Reed F. Noss, Department of Biology, University of Central Florida, Orlando, Florida 32816

Kenneth W. Outcalt, Southern Research Station, USDA Forest Service, Athens, Georgia 30602

Robert K. Peet, Department of Biology, University of North Carolina, Chapel Hill, North Carolina 27599-3280

P. Penniman, The Nature Conservancy, Jay Florida Ofﬁce, The Gulf Coastal Plain Ecosystem Partnership of Jay, Florida 32565

Andrea M. Silletti, USDA Forest Service, Southern Research Station, Clemson, South Carolina 29634

G. Andrew Stainback, School of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611

Joan L. Walker, USDA Forest Service, Southern Research Station, Clemson, South Carolina 29634

K. A. Whitney, Department of Urban and Regional Planning, University of Florida, Gainesville, Florida 32611

Richard A. Williams, School of Forest Resources and Conservation, University of Florida, Milton, Florida 32572. Preface

The history and development of the longleaf pine (Pinus palustris Mill.) ecosystem in the southeastern United States has intrigued natural resource professionals, researchers, and the general public for many decades. Prior to European settlement, longleaf pine forests were one of the most extensive ecosystems in North America. Most recent estimates suggest that only about 2.2% of the original area remains today, making it one of the most threatened ecosystems in North America. The reduction in land area of the longleaf pine ecosystem has been attributed to a number of factors, including: (i) extensive harvesting in the early 1900s that significantly reduced growing stock levels; (ii) an inadequate understanding of the biophysical factors influencing regeneration dynamics such as seeding habits and fire management; (iii) general intolerance of longleaf pine to shade and understory competition; and (iv) conversion of longleaf pine sites to other commercially important species such as loblolly (P. taeda L.) and slash pine (P. elliottii Engelm.) Over the last decade, considerable interest has grown in conserving and restoring the longleaf pine ecosystem. For example, it provides habitat for a wide variety of wildlife species including the endangered red-cockaded woodpecker (Picoides borealis Vieillot) and gopher tortoise (Gopherus polyphemus Daudin). Similarly, interest in longleaf pine regeneration and management systems has been high among land managers, ecologists, the forest products industry, and the general public. One example of this is the formation of the Longleaf Alliance based in Andalusia, AL, which is a partnership of private landowners, forest products industries, state and federal agencies, university researchers, and others interested in promoting a regionwide recovery of longleaf pine forests for their ecological and economic benefits. A variety of conventional and alternative management systems are being studied (e.g., single and multiple cohort stands) with regard to achieving these goals. Restoration efforts in the longleaf pine ecosystem have focused on expanding areas of critical habitat. Ecosystem restoration efforts, however, require effective training of natural resource practitioners. For example, knowledge regarding past history of the southeastern landscape, current status of the longleaf pine ecosystem, its potential economic and associated biodiversity values, and the role of fire in maintaining the system is of critical importance. The idea for this book, therefore, was conceived originally as a textbook for undergraduate and graduate students because the time-tested classic of Wahlenberg (1946; Longleaf Pine: Its Use, Ecology, Regeneration, Protection, Growth and Management) was out of print. To achieve that aim we desired a text with ecosystem-level coverage on topics related to the ecology, management, and restoration

ix x Preface of longleaf pine. In addition to the biophysical aspects, we desired coverage on the historical, social, and political aspects as well. It quickly became apparent that a book serving not only students, but also practitioners, scientists, policymakers, and the general public was needed. The skills required to effectively manage natural resources have changed considerably over the past two decades. In addition to managing ecosystems for products and services, increasing emphasis has been placed on ecosystem restoration. This has become particularly important in promoting the recovery, management, and ecological integrity of disturbed and degraded ecosystems. The authors who contributed to this multidisciplinary book have diverse backgrounds. As editors, we endeavored to accommodate their ideas, experiences, and interpretations over a broad range of topics. We wanted to treat each chapter as a standalone manuscript. As a result, a certain degree of overlap between some of the chapters was inevitable. However, each chapter addresses unique aspects of the longleaf pine ecosystem. The book is not intended to be viewed as a practical guide or prescription handbook for students and managers. The focus, rather, is on providing a foundation to relate information on processes to ﬁeld problems and their solutions using innovative management approaches. We hope that this book will be particularly useful to students, practitioners, and scientists seeking a broader perspective on the biophysical and social dimensions of managing and restoring the various components of the longleaf pine ecosystem. We are grateful to a large number of individuals for assistance in accomplishing this task, particularly the authors for their commitment to the project and their synthesis of the current knowledge. Also, the invaluable comments and suggestions made by the referees signiﬁcantly improved the clarity and content of the chapters. In addition to many of the chapter authors who served as reviewers for other chapters, we thank: Robert Abt, Larry Bishop, Lindsay Boring, Andre Clewell, Kenn Dodd, Kevin Enge, Dennis Hardin, Nancy Herbert, Katherine Kirkman, David Maehr, Michael Messina, Jaroslaw Nowak, Scott Roberts, Kevin Robertson, Linda Roth, Wayne Smith, George Tanner, Morgan Varner, and Jeff Walters. We are grateful to Larry Schnell who served as our copy editor during this project and wish to extend our sincere thanks to Janet Slobodien and her staff at Springer Science for their timely efforts in publishing this book.

Shibu Jose Eric J. Jokela Deborah L. Miller Contents

Contributors ...... v

Preface ...... ix

Section I Introduction

1. The Longleaf Pine Ecosystem: An Overview ...... 3 Shibu Jose, Eric J. Jokela, and Deborah L. Miller

2. History and Future of the Longleaf Pine Ecosystem ...... 9 Cecil Frost

Box 2.1: The Naval Stores Industry ...... 43 Alan W. Hodges

Section II Ecology

3. Ecological Classiﬁcation of Longleaf Pine Woodlands ...... 51 Robert K. Peet

4. Longleaf Pine Regeneration Ecology and Methods ...... 95 Dale G. Brockway, Kenneth W. Outcalt, and William D. Boyer

5. Plant Competition, Facilitation, and Other Overstory–Understory Interactions in Longleaf Pine Ecosystems ...... 135 Timothy B. Harrington

6. Vertebrate Faunal Diversity of Longleaf Pine Ecosystems ...... 157 D. Bruce Means

xi xii Contents

Section III Silviculture

7. Uneven-Aged Silviculture of Longleaf Pine...... 217 James M. Guldin

Box 7.1: The Stoddard–Neel Approach ...... 242 Steven B. Jack, W. Leon Neel, and Robert J. Mitchell

Box 7.2: The Stoddard–Neel System—Case Studies ...... 246 W. Keith Moser

8. Longleaf Pine Growth and Yield...... 251 John S. Kush, J. C. G Goelz, Richard A. Williams, Douglas R. Carter, and Peter E. Linehan

Section IV Restoration

9. Restoring the Overstory of Longleaf Pine Ecosystems...... 271 Rhett Johnson and Dean Gjerstad

10. Restoring the Ground Layer of Longleaf Pine Ecosystems...... 297 Joan L. Walker and Andrea M. Silletti

Box 10.1: Prescribed Burning for Understory Restoration ...... 326 Kenneth W. Outcalt

Box 10.2: Restoring the Savanna to the Savannah River Site ...... 330 Don Imm and John Blake

11. Reintroduction of Fauna to Longleaf Pine Ecosystems: Opportunities and Challenges ...... 335 Ralph Costa and Roy S. DeLotelle

12. Spatial Ecology and Restoration of the Longleaf Pine Evosystem...... 377 Thomas S. Hoctor, Reed F. Noss, Larry D. Harris, and K. A. Whitney

13. Longleaf Pine Restoration: Economics and Policy...... 403 Janaki R. R. Alavalapati, G. Andrew Stainback, and Jagannadha R. Matta

14. Role of Public–Private Partnership in Restoration: A Case Study ...... 413 Vernon Compton, J. Bachant Brown, M. Hicks, and P. Penniman

Index ...... 431 Section I Introduction

1 Chapter 1 The Longleaf Pine Ecosystem An Overview

Shibu Jose, Eric J. Jokela, and Deborah L. Miller

An Ecosystem in Peril? to coastal regions along navigable streams. Land clearing and open range cattle and feral The longleaf pine (Pinus palustris Mill.) ecosys- hogs that fed on longleaf pine seedlings in tem once occupied an estimated 37 million nearby woods were characteristic features of hectares in the southeastern United States these early domesticated landscapes. Commer- (Frost this volume). These forests dominated cial logging had little impact until introduc- the Coastal Plain areas ranging from Virginia tion of the water-powered sawmill in 1714, to Texas through central Florida, occupying a but by the 1760s hundreds of these mills variety of sites ranging from xeric sandhills to were turning out sawn lumber. Still, deforesta- wet poorly drained flatwoods to the montane tion was limited to narrow dendritic patterns areas in northern Alabama. The extent of the defined by streams and rivers. By this time longleaf pine ecosystem has greatly declined much of the eastern Piedmont was fully set- since European settlement. At present, it oc- tled and the frontier had passed on toward the cupies less than 1 million hectares, making it Appalachians. one of the most threatened ecosystems in the By the Civil War, all of the best land on the United States. Will this ecosystem always be in Atlantic slope was in fields and pasture, but peril? Maybe not! The objective of this chapter much virgin forest remained along the Gulf is to provide an overview of the book’s content Coast. The naval stores industry that caused that will examine the historical, ecological, sil- further decline in the area of longleaf pine vicultural, and restoration aspects of longleaf stands is also discussed in detail by Hodges pine ecosystems. in Box 2.1. This crude turpentine industry, In the second chapter in Section I, Frost de- which began in Virginia in 1608, was prac- scribes the historic context of the decline of the ticed through the Colonial Period. By that longleaf pine ecosystem and examines the cur- time, there had been little impact farther to the rent status and future outlook. Longleaf pine south, with exception of stands found along was exploited from first settlement; however, rivers in North Carolina. Then, in 1834, adap- before 1700 travel and trade limited impacts tation of the copper whiskey still for turpentine

r Shibu Jose and Eric J. Jokela School of Forest Resources and Conservation, University of Florida, Gainesville, r Florida 32611. Deborah L. Miller Department of Wildlife Ecology and Conservation, University of Florida, Milton, Florida 32583.

3 4 I. Introduction distillation made the fledgling forest indus- The first chapter in Section II (Chapter 3) by try vastly more efficient and profitable. Tur- Peet illustrates how complex the plant asso- pentining, along with the communities and ciations can be in longleaf pine forests. Based jobs it supported, moved south into Georgia on data from his own work and other pub- and then west along the Gulf Coast. Even- lished sources, Peet has classified the seem- tually, the turpentine industry reached virgin ingly homogenous expanse of longleaf pine stands in Texas by around 1900. Steam tech- woodlands into 135 vegetation associations. nology mushroomed by 1870, with prolifer- Recognizing the considerable variation that oc- ation of logging railroads, steam log skidders, curs in longleaf pine communities with simple and steam sawmills. An intensive era of logging geographic distance and subtle environmen- activities occurred in the South from 1870 to tal changes is of particular importance in mak- 1920. The 1920s also saw the beginning of ing management decisions. The vegetation as- commercial pine plantations, now more than sociations described in Chapter 3 could serve 20% of southern uplands. as a benchmark for classifying longleaf pine The presettlement range of longleaf pine forests for conservation and providing targets was estimated at 37 million hectares, of which for restoration. 23 million were longleaf dominant and 14 mil- One of the significant reasons for the re- lion had longleaf in mixtures with other pines duction of longleaf pine regeneration was the and hardwoods. By 1946, longleaf pine had interruption of natural fire cycles in the un- dwindled to one-sixth its original area. This derstory. Understanding the role of fire and decline has continued, such that only about the autecology of longleaf pine is vital for the 2.2% of the original area remains today. Of the restoration of this ecosystem. The chapter by original range, only about 0.2% of the land in Brockway et al. (Chapter 4) discusses the ecol- 2000 was being managed with fire sufficient ogy of longleaf pine and the silvicultural re- to perpetuate the open structure and species production methods commonly used for this diversity represented by the hundreds of fire- species. Longleaf pine is a very intolerant pi- dependent plant and animal species of the lon- oneer species (Landers et al. 1995) and does gleaf pine ecosystem. not compete well for site resources with other more aggressive species (Brockway and Lewis 1997; Harrington this volume). Compared to Ecological Significance other pine species, longleaf pine is not a prolific seed producer. Longleaf pine seeds require The longleaf pine ecosystem plays a prominent over 3 years for their physiological develop- role in the ecology and economy of the south- ment. Thus, good seed crops are infrequent eastern United States. These ecosystems have and may arise only once every 6–8 years. The one of the richest species diversities outside seeds are large and heavy and do not disperse the tropics. Although the overstory is domi- great distances. The short dissemination dis- nated by one species, the understory is host tances of the seeds prevent longleaf pine from to a plethora of plant species. The diversity colonizing and establishing in areas far from among the herbaceous plants is the main con- the seed source. Longleaf pine requires an ex- tributor to its high biodiversity. In general, the posed mineral soil seedbed that is free of sur- composition of the understory is site specific, face litter. Fire exclusion results in accumula- but is mainly dominated by grass species. In tion of forest litter that hinders proper germi- the western Gulf Coastal Plain, the understory nation of longleaf pine seeds (Croker 1975). is comprised mainly of bluestem (Andropogon With the removal of fire, the less fire adapted and Schizachyrium spp.) grasses. In Florida and shrub species can spread into the understory. along the Atlantic Coast wiregrass (Aristida The encroaching hardwoods compete for site beyrichiana) is dominant, with Aristida stricta resources and light with the longleaf seedlings occurring from central South Carolina through and hinder their growth and regeneration. North Carolina. Longleaf pine seedlings undergo an extended 1. The Longleaf Pine Ecosystem: An Overview 5 stemless phase without height initiation un- loblolly (P. taeda L.)–shortleaf (P. echinata Mill.) der competition from surrounding vegetation. pine in the upper west Gulf Coastal Plain. In This phase, also known as the “grass stage,” the first chapter in Section III (Chapter 7), varies in length depending on site resources Guldin presents an overview of lessons learned and competition and may last as long as 10– from loblolly–shortleaf uneven-aged manage- 25 years. These competitive interactions are ment and explains the underlying principles of the subject of Chapter 5 by Harrington. applying the same approach in longleaf pine Chapter 6 by Means explores the past and ecosystems. Described in detail are reproduc- present vertebrate faunal diversity of the lon- tion methods, stand-level regulation, and de- gleaf pine ecosystem. The highest species rich- velopmental dynamics. The Stoddard–Neel ap- ness of turtles, frogs, and snakes in the United proach to uneven-aged management is also States and Canada (Kiester 1971), as well as a described in detail by Jack et al. and Moser large salamander fauna (Means this volume), in Boxes 7.1 and 7.2, respectively. Available occurs on the Coastal Plain of the south- literature on the growth and yield of both plan- eastern United States. However, bird species tation and natural stands of longleaf pine is richness (Stout and Marion 1993) is not summarized in Chapter 8 by Kush et al. particularly high and mammal fauna is depauperate. With a number of threatened and endangered species and loss of over 97% of Ecological Restoration their habitat, these vertebrates still represent one of the largest vertebrate faunas in tem- The Society for Ecological Restoration (SER) perate North America. There are 212 resident defines restoration as an intentional activity vertebrate species in longleaf pine savannas of that initiates or accelerates the recovery of an which 38 are specialists occurring exclusively ecosystem with respect to its health, integrity, or primarily in longleaf pine savannas. and sustainability (SER 2004). The ecosystem The gopher tortoise and red-cockaded that requires restoration may be degraded, woodpecker are keystone species in this damaged, transformed, or entirely destroyed ecosystem that enable increased species rich- as the direct or indirect result of anthropogenic ness by providing shelter for many species activities. The vast majority of the remaining through their specialized activities. The gopher longleaf pine ecosystems fall into one of the tortoise is a longleaf pine specialist, which ex- above-mentioned categories. Most have been cavates extensive underground burrows used altered beyond their resiliency; therefore, it is by more than 300 species of other verte- nearly impossible for them to revert back to the brates and invertebrates (Jackson and Milstrey predisturbance state or historic developmental 1989). The red-cockaded woodpecker is the trajectory without human intervention. only woodpecker to make cavities in living Ecological restoration attempts to return trees. Because the longleaf pine trees are alive sites formerly occupied by longleaf pine when cavities are excavated, the latter persist ecosystems to their historic trajectory. Historic for up to 400 years and are used by many other conditions are therefore the ideal starting point animals over the lifetime of the tree. for restoration design. Restoration of longleaf pine ecosystems requires identifying important reference communities that have conditions Silvicultural characteristic of a “historic” state. However, us- Considerations ing a static image for restoring a dynamic forest ecosystem, is not only difficult to achieve, Uneven-aged silviculture of longleaf pine has but may not be an appropriate goal (Hobbs received considerable attention in the recent and Harris 2001). There is a need to discuss past. This reproduction method and manage- in detail ecological indicators for restoration ment system has been successfully applied assessments. These indicators should be iden- in other southern pine stands such as mixed tified for their influence on determining the 6 I. Introduction dynamics of plant community succession and native fauna into longleaf pine ecosystems in soil productivity (Burger and Kelting 1999). Chapter 11. The focus is on rare species, includ- In the past, ecological restoration has been ing those considered “sensitive,” “of special practiced using a retrospective approach, try- concern,” or “candidates” for listing by con- ing to capture the properties of an ecosystem servation groups, or state or federal agencies. that existed during some designated period Their discussion also includes federally listed of the past (Hobbs and Harris 2001). Current species as either “threatened” or “endangered” planning augments historical information by under the Endangered Species Act. Special characterizing ecosystem composition, struc- emphasis is also placed on the red-cockaded ture, function, biodiversity, and resilience from woodpecker. an existing system that is free of degrada- The importance of a landscape approach tion and located within a reasonable distance in restoring the longleaf pine ecosystem is (Harris 1999). This neighboring system is used the topic covered in Chapter 12 by Hoctor as a model or reference for comparison. The et al. Given the distinctive ecology and cur- advantage is that these reference systems can rent condition of longleaf pine communities, be studied over time and space. Sources of in- landscape ecology and regional reserve design formation that can be used in describing the principles are crucial for guiding restoration reference ecosystem include (SER 2004): efforts. Chapter 13 by Alavalapati et al. explores the socioeconomic and policy aspects of 1. Ecological descriptions, species lists, and restoration. Incentive programs in place to pro- maps of the project site prior to damage mote restoration activities are also discussed. 2. Historical and recent aerial and ground- An example regional approach is presented in level photographs Chapter 14 by Compton et al. The success- 3. Remnants of the site to be restored, indicat- ful Gulf Coastal Plain Ecosystem Partnership ing previous physical conditions and biota is emerging as a model for restoring longleaf 4. Ecological descriptions and species lists of pine across its former range. similar intact ecosystems; herbarium and museum specimens 5. Historical accounts and oral histories by persons familiar with the project site prior to damage Are We There Yet? 6. Paleoecological evidence, e.g., fossil pollen, Restoration Ecology, the art and science be- charcoal, tree ring history, rodent middens hind ecological restoration, is not an exact sci- Based on the lessons learned from several ence. Because ecosystems are dynamic, it is operational restoration projects, Section IV ex- difficult to identify exact values to determine plores the current status of restoration of the restoration success (van Diggelen et al. 2001). longleaf pine ecosystem. Restoring the over- Instead, a range of values are used to iden- story is the focus of the first chapter (Chapter 9) tify restoration trajectories and “thresholds” by Johnson and Gjerstad. The authors outline (SER 2004; Suding et al. 2004). An ecosystem restoration strategies for 10 scenarios, repre- is considered to have reached a restored state senting 10 degraded conditions commonly en- when the system has been shifted across recov- countered within the natural range of longleaf ery thresholds and has returned to the gen- pine. Walker and Silletti (Chapter 10) discuss eral direction and boundaries of the historic the techniques employed in restoring the un- trajectory. Exceeding recovery thresholds be- derstory community. The importance of fire for comes an important goal in the restoration pro- understory restoration is further explained by cess. An ecosystem is restored when it contains Outcalt in Box 10.1. Imm and Blake narrate sufficient biotic and abiotic resources to con- a success story of putting savanna back to the tinue its development (trajectory) without fur- Savanna River Site in Box 10.2. ther assistance. It will sustain itself structurally Costa and DeLotelle discuss the reintroduc- and functionally. The Society for Ecological tion and augmentation, via translocation, of Restoration has identified nine attributes for 1. The Longleaf Pine Ecosystem: An Overview 7 determining when restoration has been ac- rounding landscape have been eliminated complished (SER 2004). They are: or reduced as much as possible. 8. The restored ecosystem is sufficiently re- 1. The restored ecosystem contains a charac- silient to endure the normal periodic stress teristic assemblage of the species that occur events in the local environment that serve in the reference ecosystem so that it pro- to maintain the integrity of the ecosystem. vides an appropriate community structure. 9. The restored ecosystem is self-sustaining to 2. The restored ecosystem consists of indige- the same degree as its reference ecosystem, nous species to the greatest extent possible. and has the potential to persist indefinitely In restored cultural ecosystems, allowances under existing environmental conditions. can be made for domesticated alien species Nevertheless, aspects of its biodiversity, and for noninvasive ruderal (plants that col- structure, and functioning may change as onize disturbed sites) and segetal (plants part of normal ecosystem development, that grow intermixed with crop species) and may fluctuate in response to normal species that presumably co-evolved with periodic stress and occasional disturbance them. events of greater consequence. The species 3. All functional groups necessary for the con- composition and other attributes of a re- tinued development and/or stability of the stored ecosystem may evolve as abiotic con- restored ecosystem are present or, if they are ditions change. not, the missing groups have the potential to colonize by natural means. A monitoring and evaluation program 4. The physical environment of the restored should be in place to track the success of the ecosystem is capable of sustaining viable re- restoration efforts. A good monitoring pro- producing populations of the species neces- gram should be focused on a few key indica- sary for its continued stability or develop- tors in order to provide for statistically sound ment along the desired trajectory. information (Lindenmayer 1999). Monitoring 5. The restored ecosystem functions normally should be conducted in a systematic manner, for its ecological stage of development, and designed to provide the needed information. signs of dysfunction are absent. The following steps have been recommended 6. The restored ecosystem is integrated into a to ensure a functional monitoring plan (Block larger ecological matrix or landscape, with et al. 2001): (a) Set monitoring goals, (b) which it interacts through abiotic and biotic identify the resources to monitor, (c) establish flows and exchanges. threshold points, (d) develop a sampling de- 7. Potential threats to the health and integrity sign, (e) collect and analyze data, and (f) eval- of the restored ecosystem from the sur- uate results (Fig. 1).

Identify Collect Set Goals Variables Data

Establish Analyze Thresholds Data

Develop Feedback Evaluate FIGURE 1. Flow diagram of monitor- Study Results to Results ing process. Modiﬁed from Block et al. Design Goal Setting 2001. 8 I. Introduction

Of the listed steps, identifying the ecosys- Burger, J.A., and Kelting, D. 1999. Using soil quality tem variables to monitor can be the most dif- indicators to assess forest stand management. For ficult. A small group of interrelated proper- Ecol Manage 122:155–166. ties of a community can be used to develop a Croker, T.C. 1975. Seedbed preparation aids nat- range of values instead of any single attribute ural regeneration of longleaf pine. USDA Forest such as an indicator species or species rich- Service, Southern Forest Experiment Station, Re- search Paper SO-112, New Orleans, LA. ness index. This will help avoid identification Harris, R.R. 1999. Defining reference conditions for of a false threshold based on a single commu- restoration of riparian plant communities: Exam- nity attribute or a single threshold point (Block ples from California, USA. Environ Manage 24:55– et al. 2001). Finally, monitoring should provide 63. a feedback mechanism whereby the researcher Hobbs, R.J., and Harris, J.A. 2001. Restoration ecol- or manager can make adjustments to the mon- ogy: Repairing the Earth’s ecosystems in the new itoring program based on the analyzed data. millennium. Restor Ecol 9:239–246. Since monitoring provides data about the dy- Jackson, D.L., and Milstrey, E.G. 1989. The fauna namics of a community over time, a model can of gopher tortoise burrows. In Proceedings of the be developed from the results of monitoring Gopher Tortoise Relocation Symposium, eds. J. E. the preselected group of community proper- Diemer, D.R.Jackson, J.L. Landers, J.N. Layne, and D.A. Woods, pp. 86–98. Florida Game and ties (indicators), which can illustrate how the Fresh Water Fish Commission Nongame Wildlife community functions on a continuum. Program Technical Report 5. Tallahassee, FL. Longleaf pine still occurs over most of its for- Kiester, A.R. 1971. Species density of North Amer- mer natural range. By restoring degraded, de- ican amphibians and reptiles. Syst Zool 20:127– stroyed, damaged, or transformed tracts and 137. by expanding these pockets, it should be feasi- Landers, J.L., Van Lear, D. H., and Boyer, W.D. 1995. ble to gradually increase longleaf pine acreage The longleaf pine forests of the southeast: Re- in the Southeast (Landers et al. 1995). As quiem or renaissance? J For 93:39–44. pointed out by Van Lear et al. (2005), restor- Lindenmayer, D.B. 1999. Future directions for bio- ing the longleaf pine ecosystem is a daunt- diversity conservation in managed forests: Indi- ing task that raises many questions. Identifi- cator species, impact studies and monitoring programs. For Ecol Manage 115:277–287. cation and removal of critical constraints to SER 2004. The SER International Primer on Ecologi- moving the system across recovery thresholds cal Restoration. Society for Ecological Restoration is the most important step. However, once International, Tucson, AZ. the desired condition is achieved, it can be Stout, I.J., and Marion, W.R. 1993. Pine flatwoods maintained with adaptive management using and xeric pine forests of the southern (lower) proven silvicultural practices (Van Lear et al. Coastal Plain. In Biodiversity of the Southeastern 2005). United States: Lowland Terrestrial Communities, eds. W.H. Martin, S.G. Boyce, and A.C. Echternacht, pp. 373–446. New York: John Wiley & Sons. References Suding, K.N., Gross, K.L., and Houseman, G.R. 2004. Alternative states and positive feedbacks Block, W.M., Franklin, A.B., Ward, J.P., Ganey, J.L., in restoration ecology. Trends Ecol Evol 19:46– and White, C. 2001. Design and implementa- 53. tion of monitoring studies to evaluate the success Van Diggelen, R., Grootjans, A.P., and Harris, J.A. of ecological restoration on Wildlife. Restor Ecol 2001. Ecological restoration: State of the art or 9:293–303. state of the science? Restor Ecol 9:115–118. Brockway, D.G., and Lewis, C.E. 1997. Long-term Van Lear, D.H., Carroll, W.D., Kapeluck, P.R., and effects of dormant-season prescribed fire on plant Johnson, R. 2005. History and restoration of community diversity, structure and productivity the longleaf pine–grassland ecosystem: Implica- in a longleaf pine wiregrass ecosystem. For Ecol tions for species at risk. For Ecol Manage 211:150– Manage 96:167–183. 165. Chapter 2 History and Future of the Longleaf Pine Ecosystem1

Cecil Frost

Introduction much has been made of the loss of some 10% to 30% of wetlands in the region (Hefner and From Virginia to Texas, much of the coastal Brown 1985), the elimination of natural veg- plain landscape was once covered by a “vast etation on 97% of uplands (Table 1) has gone forest of the most stately pine trees that can be largely unnoticed. imagined . . . ” (Bartram 1791 [1955]). Long- leaf pine could be found from sea level, on the margins of brackish marshes, to around 2000 feet on the Talladega National Forest in Presettlement Vegetation of Alabama (Harper 1905; Stowe et al. 2002). the Longleaf Pine Region The spectacular failure of the primeval longleaf pine forest (Fig. 1) to reproduce itself The presettlement range of longleaf pine has after exploitation is a milestone event in the been estimated at 37 million hectares, of which natural history of the eastern United States, 23 million were longleaf dominant and 14 mil- even greater in scale and impact than the lion had longleaf in mixtures with other pines elimination of chestnut (Castanea dentata) from and hardwoods (Frost 1993). States bordering Appalachian forests by blight. This chapter the Atlantic, and some of the Gulf Coast region, discusses presettlement extent and summa- lack the systematic database of witness trees rizes major events in the decline of the long- that were recorded when lands were surveyed leaf pine ecosystem and its displacement after 1790 under the township, range, and sec- from more than 97% of the lands it once tion system in the rest of the country. Thus, occupied. there can be no easy reconstruction of virgin Land uses ranging from 100 to 400 years of forests from such data. Even where histori- agriculture; open range grazing by hogs and cal survey records are available, interpretation other livestock; logging; production of turpen- is compromised because surveyors routinely tine, and elimination of naturally occurring failed to distinguish the various species of pine, wildﬁres have left less than 3% of the upland just lumping them as “pine” on records and landscape in entirely natural vegetation. While survey plats. There is, however, an exceptional

r Cecil Frost Adjunct Faculty, Curriculum in Ecology, University of North Carolina, 119 Pot Luck Farm Road, Rougemont, North Carolina 27572.

9 10 I. Introduction

FIGURE 1. Virgin longleaf pine savanna 10 miles east of Fairhope, Baldwin County, Alabama, August 13, 1902. Note the absence of woody understory and the classic bilayered structure of ﬁre-resistant canopy over a rich herbaceous layer under a natural ﬁre regime (estimated at 1–3 years at this site). Roland Harper commented that “...itmaynever be possible to take such a picture in Alabama again.” Photo from Harper (1913). 2. History and Future of the Longleaf Pine Ecosystem 11

TABLE 1. Distribution of natural vegetation and land use categories in presettlement forests, in 1900, and in 2000 for the 412 counties of the original longleaf pine ecosystem. Percent of Percent of uplands region ha × 1000 a × 1000 Presettlement 1. Longleaf pine (dominant) 52.0 36.0 22,852 56,430 2. Longleaf (mixed)a 33.2 23.0 14,606 36,064 3. Mixed (w/o longleaf) 9.0 6.3 4,001 9,878 4. Upland slash pine 3.3 2.3 1,440 3,555 5. Beech-magnolia 2.5 1.7 1,108 2,735 11. Wetlands 0 30.7 19,496 48,137 100.0 100.0 63,503 156,799 1900 1. Longleaf pine (natural) 24.2 17.5 11,109 27,430 2 + 3. Mixed pyrophytic spp. 20.7 15.1 9,581 23,657 4. Upland slash pine 1.7 1.2 775 1,914 5. Beech-magnolia 0.4 0.3 166 410 6. Successional forests 25.0 18.1 11,501 28,399 7. Pine plantation 0 0 0 0 8 + 9. Pasture and cropland 27.0 19.6 12,448 30,733 10. Developed 1.0 0.7 460 1,137 11. Wetlands 0 27.5 17,463 43,119 100.0 100.0 63,503 156,799 2000 1. Longleaf pine (natural) 2.1 1.7 1,017 2,510 2 + 3. Mixed pine-hardwood 0.5 <0.4 250 618 4. Upland slash pine 0.4 0.3 222 547 5. Beech-magnolia 0.4 0.3 222 547 6. Successional forests 44.0 34.6 20,104 49,639 7. Pine plantation (all species) 15.2 12.0 11,077 27,350 8. Pasture 6.4 5.0 3,456 8,534 9. Cropland 20.8 16.3 6,027 14,882 10. Developed 10.2 8.0 7,538 18,616 11. Wetlands 0 21.4 13,590 33,556 100.0 100.0 63,503 156,799 Vegetation and Land Use Categories 1. Natural, fire-maintained communities dominated by longleaf pine 2. Longleaf-dominant patches and longleaf pine in fire-maintained mixed species savanna and woodland having longleaf, shortleaf, loblolly, pond pine, and sometimes hardwoods in various combinations 3. Pyrophytic woodlands without longleaf pine 4. Natural, fire-maintained slash pine on uplands 5. Southern mixed hardwood forest (nonpyrophytic, fire-refugial beech-magnolia) 6. Successional mixed pine-hardwood forests resulting from logging, old field abandonment, and fire exclusion 7. Pine plantation (all species) 8. Pasture 9. Cropland 10. Cities, towns, roads, industry 11. All wetlands: types wetter than hydric longleaf pine savanna a Of the combined area of longleaf-dominant and longleaf-mixed species stands with patches of pure longleaf, I estimated the total original area of longleaf-dominant stands at 30 million hectares. 12 I. Introduction

FIGURE 2. Presettlement range and major divisions of the longleaf pine ecosystem, showing the transition region between frequent fire communities of the Coastal Plain and the fire communities of the Piedmont described by Sargent (1884). Reprinted from Frost 1993 with permission from the Tall Timbers Research Station. narrative literature on the longleaf pine forests, of longleaf pine as well as maps, lumbering dating from 1608 when Captain John Smith records, and calculations of acreage and board exported the first barrels of pitch and tar feet by state, allowing a reasonable approxi- made from pines near the new settlement at mation of its original range and abundance. Jamestown, Virginia (Smith 1624). Figure 2 is a reconstruction of the original Because of its primacy as the commercial range of longleaf pine, using as a base a com- tree of the South, longleaf pine became in pilation of the state maps prepared by Sargent. the 1880s the first forest species to be stud- Range maps and numerous locations provided ied in detail by botanists and early professional by Ruffin (1861), Lockett (1870), Hale (1883), foresters. Major studies by Sargent (1884), State Board of Agriculture (1883), Ashe Mohr (1896), Ashe (1894a), and Harper (1913, (1894a,b), Harper (1905, 1906, 1911, 1913, 1928) include literally hundreds of locations 1914, 1923, 1928), Sudworth (1913), Mattoon 2. History and Future of the Longleaf Pine Ecosystem 13

(1922), Wakeley (1935), Wahlenburg (1946), Fire Relations of the Original and Little (1971) were also useful. In addi- Forests tion, numerous historical references and remnant locations for longleaf were used to fill In the pastoral landscapes of Britain, domesti- in areas unknown to Sargent and reconstruct cated since Roman times, wildfire was an alien its original northern range in North Carolina concept. A British traveler in South Carolina and Virginia. The resulting map includes all in 1829 was astonished to discover a recently areas known to have once supported long- burned stand of longleaf pine: leaf pine. In all, in the presettlement range of longleaf pine there were 412 counties in There was no underwood properly so-called, while nine states. Sources of statistics and meth- the shrubs had all been destroyed a week or two be- ods for reconstructing the original range are fore by a great fire. The pine-trees, the bark of which discussed further in Frost (1993). Figures for was scorched to a height of about 20 feet, stood on pine plantation (all species) were updated ground as dark as if it had rained Matchless Black- ing for the last month. Our companions assured us using a projection for 2000 by McWilliams that although these fires were frequent in the for- (1987), and corrected for the area of each state est, the large trees did not suffer. This may be true, lying outside the original range of longleaf but certainly they did look very wretched, though pine. their tops were green as if nothing had happened. (Hall 1829, p. 137) Amount of Longleaf Pine Historically, agents of fire included light- Remaining in 2000 ning, Native Americans, and European settlers. Agents of fire suppression were bodies of wa- According to data of the 1995 Forest Inventory ter, topography (steep slopes, islands, penin- and Analysis (FIA), there were some 1.02 mil- sulas [Harper 1911]), a few plantation owners lion hectares of longleaf pine remaining at that (Gamble 1921, p. 27), and government agen- time. About 15% of this, or 178,200 hectares, cies (Sherrard 1903). Varying effects of fire consisted of pine plantation, mostly on old field in the landscape mosaic have been attributed or mechanically prepared sites, so about 85%, to fire frequency, fire intensity, and season of or 841,800 hectares, of naturally regenerated burn (Garren 1943; Komarek 1974). Given longleaf pine having some degree of under- that lightning fires would mostly have been story integrity persist (Outcalt and Sheffield growing season fires, fire frequency must have 1996). There are a variety of factors of un- been the most important fire variable in pre- certainty in the estimate of remaining longleaf settlement vegetation. pine. The FIA data are based only on stands Mattoon (1922) commented that longleaf with at least 50% longleaf pine canopy cover, lands experienced fire at an average of every 2– so will be an underestimate of the total re- 3 years over millions of hectares. There is evi- maining. On the other hand, longleaf pine in dence that fire frequency is proportional to fire FIA permanent sample plots declined by 22% compartment size: the larger the fire compart- from 1985 to 1995 (Kelly and Bechtold 1990; ment the higher the fire frequency, and in the Outcalt and Sheffield 1996): the data were al- largest fire compartments (over 1000 km2), ready 9 years out of date as of January 2004 the original fire frequency averaged 1–3 years and so will be an overestimate of the longleaf (Frost 2000). On the Pamlico Terrace and dominant natural stands remaining in 2005. other terraces of the lower Coastal Plain from We would expect these under- and overesti- Virginia to Texas, there were numerous tracts mates to partially cancel each other, making of land from several hundred to over a thou- the figure of 841,800 hectares a reason- sand square kilometers in size without a single able estimate of naturally regenerated long- natural firebreak. In Florida, Komarek (1965) leaf in all stands in 2000. This is about 2.2% reported that 99 wildfires were started by of the presettlement extent of longleaf pine. lightning on a single summer day. On the 14 I. Introduction

Chesapeake 1–3 years Flat plains, some rolling plains, Bay local relief mostly less than 100 ft.

4–6 years Irregular plains and tablelands, local relief mostly 100-300 ft.

7–12 years Plains with hills and open low mountains, Albemarle local relief 300-3,000 ft. Sound

>12 years Wet swamps, high mountains where less than 20% of area is gently sloping, local relief near 0 or up to 6,000 ft. Pamlico Sound

Wilmington

Charleston

Savannah

Jacksonville

Mobile Bay Galveston Bay

Tampa Bay

Miami

FIGURE 3. Presettlement fire regimes of the southeastern United States. Frequencies are for the most fire- exposed parts of the landscape. Each region contains variously fire-protected areas with lower incidences of fire (revised from Frost 1995, 2000). Revised from Frost 1995 with permission from the Tall Timbers Research Station.

Pamlico Terrace, where a single ignition might 12,000 to 13,000 years ago, essentially all burn 1000 km2, a few ignitions in each state fires would have been caused by lightning. might be sufficient to burn most of the land- E V. Komarek marshaled evidence to support scape. On the other hand, fire frequency the idea that lightning alone is adequate to should decrease inland on the more dissected account for evolution of pyrophytic vegeta- upper Coastal Plain and Piedmont, where nu- tion, the antiquity of which far exceeds the merous separate ignitions would be required appearance of aboriginal peoples on the scene. to burn the decreasingly smaller fire compart- This provided a basis for thinking about fire ments. The resulting decrease in fire frequency, as a ubiquitous environmental parameter, as along with clayey soils, colder winter tem- influential as slope, aspect, rainfall, and temperatures, and increased topographic variation perature on shaping vegetation structure and should explain the admixtures of other pine the species composition of plant communities species and hardwoods with longleaf in the (Komarek 1964, 1965, 1966, 1967, 1968, transition regions (Sargent 1884). 1972, 1974). His paper on ancient wildfires Figure 3 shows generalized presettlement (1972) seems to have had particular im- fire frequencies of the longleaf pine region. Be- pact on paleoecologists, and opened a door fore immigration of Indians into the Southeast into inquiries concerning the role of fire and near the end of the Wisconsin glaciation some vegetation through geologic and evolutionary 2. History and Future of the Longleaf Pine Ecosystem 15 time (Cloud 1976; Cope and Chaloner 1980; with prairies and open woodlands maintained Scott 1989; Scott and Jones 1994). by fire. These open landscapes were almost The emerging picture suggests that terres- certainly the result of burning by Native trial vegetation has evolved with fire from Americans (Barden 1997). its very beginning in the early Devonian Era, some 400 million years ago. Some species, such as Venus flytrap (Dionaea muscipula), have been Distribution of Major Vegetation shown to require a mean fire frequency of at least 3 years to survive (Frost 2000). Venus fly- Types in Presettlement Forests trap is a highly evolved species with a suite of Sargent (1884) divided the range of longleaf adaptations far too complex to have evolved pine into two regions, the larger having long- in the short time since Native Americans ap- leaf as the most common dominant tree, and peared on the scene. Further, the requirement a second region around the margins of the for fire for reproduction in species such as first, in which longleaf occurred in patches longleaf pine, and even a fire frequency of or in mixed stands transitional to other types 1–3 years for species such as Venus flytrap, outside its range. Each of these two was fur- may be millions of years old. Such adaptations ther divided in Fig. 2. In the flat-to-gently may have developed along with evolution of rolling lands Sargent described longleaf as the species themselves, rather than represent- the “prevailing growth” on the uplands and ing adaptations to fire in the mere 12,000 or F. A. Michaux reported that “Seven-tenths 13,000 years that humans have been using of the country are covered with pines of fire in the Western Hemisphere. Such species one species, or Pinus palustris . . . ” (Michaux, indicate that some parts of the landscape— 1805 [1966]). This longleaf-dominated land- the largest fire compartments—experienced scape included a diverse mosaic of pine savan- a natural fire frequency of 2–3 years long nas, sandhills, and flatwoods, with variants in before immigration of man into the Western other habitats, such as riparian sand ridges, Hemisphere, and before man, the only agent Carolina bay sand rims, coastal scarps, and that could have provided a frequent ignition dunes (Peet and Allard 1993; Harcombe et al. source was lightning. 1995). On the other hand, burning by Native Boundaries of the primary region were com- Americans did transform vegetation in many piled almost exactly as drawn on Sargent’s in- parts of the southeastern landscape. Accounts dividual state maps. In Fig 2, I divided this from the Colonial Period describing Indian first region into two, depending on presence burning practices indicate that use of wild- or absence of wiregrass. Wiregrass in North fire by Indians in the Southeast peaked in fall Carolina and the northern third of South and winter when fires were set to drive game Carolina is Aristida stricta, that from south- (Smith 1624; Lawson 1709; Byrd 1728; Martin ern South Carolina to Mississippi is Aristida 1973). On the outer Coastal Plain, where beyrichiana (Peet 1993). Vegetation type 1 indi- annual spring and summer lightning fires pre- cates the portion of the known historical range empted fuel, the effect of any Indian burning of wiregrass that occurs within the longleaf may have been only a slight increase in burn pine ecosystem, based on herbarium records area resulting from the inclusion of peninsu- (Parrott 1967; Peet 1993). las and isolated patches of uplands that otherwise were naturally protected from fire. On the other hand, Indian influence may have been Transitional Communities much more significant on dissected inland terraces and the Piedmont, where their pri- Sargent’s second major assemblage of commu- mary effect, in compartments missed by light- nities included the mosaic of forest types tran- ning, would have been a net increase in fire sitional between coastal plain regions dom- frequency. Early explorers described some inated by nearly pure stands of longleaf, regions of the Piedmont that were dotted and the oak–hickory–shortleaf pine pyrophytic 16 I. Introduction woodlands of the Piedmont. Sargent described transition regions and replaced them with pure the transition regions as “long leaved pine (Pi- longleaf. nus palustris) with hardwoods in about equal proportion” in the Gulf states and “short leaved (Pinus echinata) and loblolly pine (P. taeda) in- Hardwoods in Presettlement termixed with hardwoods and scattered long Forests leaved pine” in the Atlantic states. I added the Several types of natural hardwood communi- transitional woodlands around the northern ties occur interspersed in the longleaf pine up- and eastern sides of the primary longleaf range lands. Besides longleaf pine stands with un- in Virginia and North Carolina. Not described derstory turkey oak (Quercus laevis), there are by Sargent, these stands included variants in stands of mixed scrub oaks (Quercus laevis, which pond pine (Pinus serotina) was added to Quercus marilandica, Quercus incana and Quercus the mixture (Ashe 1894a). margaretta); pyrophytic woodland with mixed longleaf, post oak, southern red oak, and Mixed Patches versus Mixed Species mockernut hickory (Carya tomentosa); and patches of post oak savanna (Quercus stellata), The importance of natural mixtures of long- the importance of which has been mostly over- leaf pine with other fire-resistant trees has looked. been generally overlooked. In Sargent’s tran- In contrast to the dominant fire communi- sition regions we can further distinguish the ties, small areas of nonpyrophytic types such as difference between mixed longleaf-dominant Southern Mixed Hardwood Forest, dominated patches in a landscape with other forest types, by beech, magnolia, semievergreen oaks, and and true mixed-species stands. The first was a other hardwoods, may have been confined to patch mosaic having nearly pure stands of naturally fire-sheltered sites within the range longleaf pine on south slopes and upland of longleaf pine (Harper 1911). Old-growth ridges. Both Mohr (1896) and Harper (1905, stands of beech and other mesophytic hard- 1923, 1928) described pure stands as well woods can be found on steep slopes, islands as mixed stands. In the second group, they in swamps, and a few upland flats on penin- pictured the mixed pyrophytic types as open sulas. In many places, species such as beech woodland with a geographically varying mix- (Fagus grandifolia) are now escaping from these ture of the dominant trees, which were long- fire refugia onto the uplands (Ware 1978). leaf, shortleaf pine, loblolly pine, post oak, Studies by Delcourt and Delcourt (1977) in white oak, southern red oak, hickories, and the Apalachicola bluffs region of the Florida various scrub oaks. From historical photos, Panhandle suggest that fire-refugial Southern these were bilayered communities, having a Mixed Hardwood Forest occupied less than 1% tree canopy and a savannalike grass–forb un- of the presettlement landscape. derstory, indicative of a frequent fire regime. The existence of natural mixed species stands has been overshadowed by the remarkable pure longleaf stands that dominated most of Landscape Changes 1565 the southern uplands, and by the fact that the to 1900 mixed stands occurred on the moister and finer textured, more fertile soils, the preponderance Ecosystem Changes in the Early of which were cleared for farming long ago Colonial Period (Williams 1989). These diverse communities, with all their geographic variation, have never While the landscape that greeted the first two been adequately described. With rising inter- major groups of European settlers held as- est in restoring longleaf pine, well-intentioned tonishing forest resources, neither the English individuals have in some cases eliminated nat- nor the Spanish were well equipped to ex- ural mixed longleaf–shortleaf savanna in the ploit them, and the two cultures used radically 2. History and Future of the Longleaf Pine Ecosystem 17

FIGURE 4. Pattern of settlement in the Southeast to 1890. Note the three small centers of population in Florida, which comprised most of its sparse population until 1821. With exception of the new cotton plantation regions, most virgin forest of the interior of the six Gulf states remained intact in 1850. Map redrawn from Hammond Inc., Maplewood, NJ. Reprinted from Frost 1993 with permission from the Tall Timbers Research Station. different approaches in exploitation of the New United States in 1821—the Spanish blocked World. settlement of the Gulf Coast interior, leav- DeSoto set out in 1539 to explore the Gulf ing longleaf pine forests of much of the re- Coast interior, an epic overland journey com- gion in pristine condition well into the nine- plete with army, horses, and droves of hogs, teenth century. Curiously, with the exception that took him as far inland as the Cherokee of a handful of coastal villages such as St. towns of North Carolina and west beyond the Augustine and Pensacola, they never pursued Mississippi River (Bakeless 1961). While the immigration and settlement of the land. In Spanish, disappointed with the scarcity of in- 1821, at the end of their occupation, the en- teresting targets for conquest and pillage, lost tire European population of Florida was barely interest in the north Gulf interior, they contin- more than 20,000 people, scarcely enough for ued to control access to much of that vast re- a reputable town. Note the contrast in settle- gion from Florida to Texas. What is signiﬁcant ment patterns between Spanish lands and En- for landscape history is that during their 256- glish settlements along the Atlantic in Fig. 4. year tenure—from establishment of St. Augus- Unlike the Spanish military outposts, En- tine in 1565 until cession of Florida to the glish settlements were commercial ventures 18 I. Introduction

financed by corporations of wealthy stock- for fencing. Of great importance to natural sa- holders. Backers of the 1607 Jamestown, VA, vanna and woodland communities, though lit- expedition under John Smith promoted settle- tle remarked historically, was the introduction ment and domestication of the land in order of swarms of hogs, cattle, horses, mules, sheep, to establish a productive populace from which and goats onto open range in all of the settled they could harvest taxes, agricultural produce, areas. Of these ravening herds, hogs in partic- and whatever natural products the land could ular would play a major part in the decline of supply (Smith 1624). longleaf pine. For the first 150 years, dependence on water for travel and trade limited settlement to Naval Stores and the Original Northern the nearest high lands along coastal sounds, Range of Longleaf Pine to the bays, and the tidal portions of major and mi- Virginia/Maryland Border nor streams (Hart 1979). The tidewater area included at least 10,000 miles of shoreline Tar, pitch, rosin, and turpentine were collec- from Virginia to Texas, and until the coastal tively called naval stores (Ashe 1894a; Mohr zone was thoroughly populated there was lit- 1896) and were produced in the Southeast al- tle incentive to push inland. Domestication of most exclusively from longleaf pine, although this easily accessible landscape resulted in land smaller amounts were made from slash pine, clearing and establishment of saturation den- shortleaf, and sometimes even loblolly pine sities of open-range hogs and other livestock (Michaux 1871) (see box by Hodges in this that fed on longleaf pine seedlings in nearby chapter). There were five substances com- woods. monly produced from longleaf pine gum: At that time, in the absence of machinery, crude turpentine, spirits of turpentine, tar, timber was worthless except for local use in pitch, and rosin. Crude turpentine was just fencing and log cabin construction. The only the fresh gum exuded from the tree when a milled boards were laboriously pit sawed by section of bark was removed. Spirits of tur- hand with crosscut saws, using one man in a pit pentine was the aromatic fraction produced and another above (Hindle 1975). A very early by distilling crude gum, and rosin was the exception, a water-powered sawmill built at dense, waxy residue left over from distillation. Henrico on the James River in Virginia in 1611, These materials were produced from the living was destroyed by the Indians a few years later tree. Tar was the product of distillation of dead (Hindle 1975). Port records from the British “lightwood,” the resin-rich heartwood from Public Records Office from the early 1600s old stumps, or gathered from partly decayed show that while lumber was a frequent item trunks on the forest floor and distilled in tar in ship’s cargoes, the quantities were small. kilns. The black, much thicker pitch was simply Cooperage stock—barrel staves and wooden tar that had been burned down in iron “pitch water pipes made from oak and white cedar— kettles” to about one-third its original volume. supplied practically the only manufactured The early history of naval stores and long- items for export for the first hundred years leaf pine has been all but lost, since the species (British Public Records 1607–1783). was commercially extirpated from much of its At the onset of agriculture, timber was lit- northern range by 1850. Even Mohr (1896) tle more than an obstruction. Settlers simply states that the naval stores industry began in killed trees by girdling them, and the land was North Carolina. Such was not the case, how- then burned and grazed, or planted in corn and ever; it had been carried on earlier for over 200 other crops beneath the dead timber (Beverley years in Virginia. Longleaf once extended to 1705 [1947]). Since most livestock were al- within a mile of the Maryland border (Fig. 5), lowed to graze on open range in the woods, and likely continued into that state. I exam- it was necessary to fence them out of the small ined a herbarium specimen of longleaf pine crop patches (Beverley 1705 [1947]). As a re- collected near Sinnickson, VA, in 1925. I also sult, the principal early demand for timber was visited the site and interviewed the collector 2. History and Future of the Longleaf Pine Ecosystem 19

FIGURE 5. Documentation of the original range of longleaf pine in Virginia. Circles indicate herbarium specimens or living trees seen from 1960 to 2004, or reported to me by local foresters (also includes two tar kilns visited in Suffolk and Chesapeake). Squares denote clear historical records, some as early as 1608, but lacking herbarium specimens. Triangles are used for naval stores place names like Pitch Kettle Road, Lightwood Swamp, Tar Pit Swamp, and Tar Bay. Reprinted from Frost 1993 with permission from the Tall Timbers Research Station.

before his death (Moldenke 1979, personal on the north side of the James River, where communication). He reported that he collected Smith reported finding only a tree here and the specimen from a natural stand growing on there “fit for the purpose” [of making naval the ridges of forested coastal sandhills that lie stores]. on the scarp that forms the eastern uplands Tar and pitch were absolutely essential com- before dropping down into the coastal marshes modities until the development of petroleum- on the Atlantic side of the Eastern Shore. These based substitutes in the mid-1800s. Wagons low sandhills continue into Maryland only 2 could not move without tar to grease the axles. miles from this site. This area is part of a large, Ships could not sail without tar and pitch for unbroken fire compartment, and it is almost waterproofing cordage and sails, for caulking certain that longleaf pine once extended at leaks, and for coating hulls to prevent destruc- least into Worcester County, MD. This state, tion by shipworms (Wertenberger 1931). Dur- however, was not included in the presettle- ing the Revolutionary War, Captain H. Young ment range map for lack of a verifiable record. wrote to his colonel “. . . let me entreat you Tar and pitch were produced in Virginia once more to lay before the Council my dis- for over 200 years before the boom in North tressed situation for the want of two Barrels of Carolina that gave the Tarheel State its nick- Tar.” “I have offer’d Brown (who is the only name. We know of the early trade, the extent one that has Tar) his price in specie, or two of which has never been thoroughly investi- barrels of Tar for one, both of which2 offers gated, only through disparate and widely scat- he has refused. Our waggons can’t run for the tered records. The southern naval stores indus- want of tar” (Young 1781 [Calendar of State try began in 1608 when John Smith exported Papers (Virginia) 1881], 2:619). Colonel Davies the first “tryalls of Pitch and Tarre” (Smith had his own problems with the recalcitrant 1624). The settlement was founded in 1607 Mr. Brown, while trying to ship 30 cannon and the next year the Jamestown, VA, colony to prevent their capture by the British: “Our exported some three or four dozen barrels to own vessels are all in readiness, except for England. To all indications, longleaf was sparse some slight repairs, for the finishing of which 20 I. Introduction some small quantity of tar is necessary, tho’ not Anne County and part of Norfolk County. The more than a barrel at the utmost—we cannot disposition was split three ways: local con- procure this quantity under some time unless sumption, sale to ship’s masters, and export we obtain it from Mr. Brown, who will not to the West Indies. Customs records on file part with it upon any other terms than for for ports from around the Chesapeake Bay specie,2 of which the State has none to pay” list barrels of naval stores as one of the most (Davies 1781 [Calendar of State Papers (Vir- common exports from the colony from the ginia) 1881], 2:599). late 1600s until the Revolution (British Pub- Early naval stores production concentrated lic Records).3 In a typical entry, the customs on burning tar kilns for tar and pitch. Tar official at Hampton, VA, noted on April 12, kilns were earth-covered mounds of sev- 1745, “Cleared at Hampton, the snow John eral cords of collected dead pine “lightwood” and Mary, Thomas Bradley, for Liverpool with that were burned under controlled condi- 106 hhd. tobacco, 500 bbl tar, 60 walnut stocks tions by carefully regulating the amount of and 5600 staves” (a snow [pronounced like air let into the mound. This sometimes dan- “now”] was a square-rigged sailing vessel, one gerous process took up to 2 weeks of con- of the most frequently mentioned trading-ship tinuous management—from the first drops designs in early eighteenth century). The ex- which might not appear for several days, un- act point of origin of the goods is seldom de- til the tar ceased to flow into the barrels terminable since ships often stopped at planta- placed below (Catesby 1731, 1743). The sections up and down the rivers to pick up cargo, ond, more destructive practice involved box- sailing on to be cleared through customs at the ing of live trees for the crude gum that was ports of Accomack, Hampton, or Norfolk. then shipped to New England or Europe for Twenty-five years later, the export trade distillation of spirits of turpentine in crude iron had increased such that, from March 25 to retorts. While boxing was practiced as early September 29, 1726, 17 vessels were cleared as 1608 (Smith 1624), the necessity of ship- from Hampton, only one of the ports, with ping the bulky crude gum long distances lim- 1194 barrels of pitch and 6004 barrels of tar. ited the price and demand for the first hundred One ship alone carried 1580 barrels of tar and years. 130 of pitch (British Public Records 1726). By While tar and pitch were made from 1608 1791 the port at Norfolk exported 29,376 tons on, most seem to have been consumed lo- of naval stores (La Rochefoucauld 1799). By cally until around 1700. In 1697, Governor Sir 1803, the number of ships cleared for foreign Edmund Andros said that Virginia produced ports from Norfolk and Portsmouth reached no naval stores for sale except along the Eliz- 484, and it was reported that Virginia was no abeth River [Norfolk County], where about longer able to meet the export demand for yel- 1,200 barrels of tar and pitch were made an- low pine (Wertenberger 1931). The designa- nually (Pierce 1953). This would have had tion “yellow pine” most often meant lumber ready market at the port of Norfolk just a few from longleaf pine in the early trade. miles downstream. The industry was carried Early channels of trade in tar and pitch in on by poor men who built their kilns unas- Virginia were the Elizabeth and Nansemond sisted by servants or slaves, and considered Rivers, with their tidal tributaries interpene- a few dozen barrels a year an excellent out- trating the lands in the interiors of Norfolk put (Wertenberger 1931). F. A. Michaux, writ- and Nansemond counties. Not a single longleaf ing about his own observations made around pine remains within the watersheds of these 1802, notes that “toward the north, the Long- two stream systems today, and not a single leaved Pine first makes its appearance near tree remains in the former longleaf counties of Norfolk, in Virginia, where the pine-barrens Norfolk and Princess Ann. The only evidence begin” (Michaux 1871). remaining in the three counties east of the In 1704 Jenings (1704 [1923]) reported Nansemond River are a few remnant tar kilns some 3000 barrels of tar produced in Princess and a handful of isolated trees in Suffolk. Most 2. History and Future of the Longleaf Pine Ecosystem 21 of the remainder of Colonial production of tar, counties, Virginia, and Gates and Chowan pitch, and turpentine originated from counties counties, North Carolina. The first record of along the south side of the James River, where naval stores produced in North Carolina was there is evidence of once-extensive longleaf in 1636, 17 years before the first settler set up pine forests (Frost and Musselman 1987). a house and trading post in 1653. A visitor from There were as much as 600,000 hectares Bermuda sailing up the Chowan River was sur- in the original range of longleaf pine in Vir- prised to discover a large number of men there ginia, based on the extent of suitable soils in busily producing “sperrits of rosin” (Clay et al. the original range defined in Fig. 5. Longleaf 1975). This was in the vicinity of the “Sand pine forests in Virginia appear to have been Banks” of western Gates County. The crew had largely exhausted by 1840, after which no fur- apparently come south, overland from the set- ther naval stores production was listed (U.S. tlements, only a few years old, along the James Census Office 1841). The Census of Manufac- River in Virginia. From 1980 to 1990 I was tures for that year listed 5012 barrels produced only able to locate about 25 old longleaf trees from five counties. The species no longer oc- in the Sand Banks region. I counted annual curs in two of these and I was able to find fewer rings when some of these were logged around than 200 mature native trees left in this state— 1980: the largest was 308 years old and only enough to stock perhaps 5 hectares—where 23 inches (60 cm) in diameter on the stump once there were more than 4000 km2 domi- when cut. nated by longleaf pine. In 1893, forester B. E. Schoepf (1788 [1911]) traveling down the Fernow concluded that “[i]n Virginia the long- coastal plain from Virginia to South Carolina leaf pine is, for all practical purposes, extinct.” observed that “. . . the greatest and most im- In Southampton County, Virginia, I met a portant part of the immense forests of this farmer, 84 years old—born around 1896— fore-county consists of pine . . . ”, and com- whose recollection went back to the days of mented on “. . . the opportunity for consider- “longstraw” pine as it was known there in the able gain from turpentine, tar, pitch, resin and past. Perhaps the last person in the state to re- turpentine-oil.” In the northern tier of North member that term from daily use, he took me Carolina counties, as mentioned above, some to see three trees that he had ordered to be left 20 mature trees remain in Gates County, only when his land was logged. Longleaf pine has 2 trees are known in Hertford County, and been completely extirpated from 11 of the orig- a single tree in Perquimans County. The last inal 15 counties of its range in Virginia. Rem- stand of longleaf in Northampton County was nant trees can now be found only in Isle of logged about 1980, and longleaf pine has also Wight, Southampton, Suffolk, and Greensville been extirpated from Currituck, Pasquotank, counties. Washington, and Tyrrell counties. Fernow (1893) observed that “in North Carolina, in the division of mixed growth and Southward Migration of the in the plain between the Albemarle and Pam- Naval Stores Industry, North lico Sound, the long-leaf pine has likewise been almost entirely removed and is replaced Carolina to Texas with the loblolly.” In the central part of the In 1622, John Pory traveled overland from state, there was considerable turpentining ac- Jamestown to the Indian town of Chowanoc, tivity along the Tar River in the central Coastal passing through a “great forest of Pynes 15. or Plain by 1732, and by 1850 the state was the 16. myle broad and above 60. mile long, which world’s leading supplier of naval stores (U.S. will serve well for Masts for Shipping, and for Censuses of Agricultural and of Manufactures pitch and tarre, when we shall come to ex- 1841, 1853, 1864, 1872, 1883, 1895). Agricul- tend our plantations to those borders” (Powell turalists complained that the entire labor force 1977, p. 101). These were the great pine bar- of the Coastal Plain was employed in the tur- rens of western Isle of Wight and Nansemond pentine orchards, to the neglect of agriculture 22 I. Introduction

by Merrens (1964). Gamble (1921), Croker (1987), and Earley (2004) have reviewed the history of naval stores for the rest of the South. Few mature trees escaped the turpentine boxing procedure. Large trees were boxed on three or even four sides (Schoepf 1788), with deep wedges cut into the base to collect the resin (Fig. 6). Crude gum was dipped from the box six to eight times a season and transported by cart or boat to the nearest still (Figs. 7 to 9). Casks of distilled spirits of turpentine and barrels of rosin, the residue after distillation, then were shipped downstream to the nearest port (Fig. 10). Using nineteenth- century methods, virgin stands often produced for only about 4 years (Mohr 1896). Weak- ened trees in abandoned turpentine orchards often were blown over or killed when the

FIGURE 6. Boxing trees for turpentine. Bark and cambium were removed and large boxes were chopped into the base to collect the crude gum. Photo courtesy of U.S. National Archives.

(Ruffin 1861). By 1900 longleaf had been decimated in North Carolina and the industry had passed on to the south. Ashe (1894b) commented: “In North Carolina most of the trees which now bear seed are boxed and have been in this condition for 50–100 years . . . .” Introduction of the copper still in 1834 allowed concentration of the final product into distilled “spirits of turpentine” making the process highly efficient, slashing shipping costs, and touching off a wave of commercial exploitation which swept south from North Carolina to Texas decade by decade, decimating the longleaf pine region within 80 years (Mohr 1896). Sargent’s state maps (1884) for Louisiana and Texas show the ex- FIGURE 7. Gum was collected every few weeks by tent of turpentine orcharding being carried dipping with large spoons. Barrels were crafted lo- into the virgin pine forests. The history of naval cally from white oak. Photo courtesy of U.S. Na- stores in North Carolina has been reviewed tional Archives. 2. History and Future of the Longleaf Pine Ecosystem 23

FIGURE 8. Barrels of crude gum were taken by boat or wagon to the nearest still. Photo courtesy of Forest History Society. next ground ﬁre set the residue ablaze in the Thomas Gamble (1921, p. 35) summarized boxes (Fig. 11). Much of the virgin timber the wave of turpentining that decimated the thus was wasted until around 1870, when virgin longleaf forests: narrow-gauge logging railroads were extended into upland forests. As forests of each state The exhaustion of the South Carolina pine forests so far as heavy supplies of naval stores were con- were exhausted the industry moved south and cerned, was astoundingly rapid. Such a thing as by 1890 foresters raised the alarm that with- conservation was undreamed of. The vast forests of out provision for reforestation the turpentine Georgia and Alabama and Florida were too invit- industry would soon come to an end (Ashe ing to promote the thought of care in the use of 1894b). what remained of the Carolina pine forests that had

FIGURE 9. Introduction of the copper still into the woods in 1834 permitted reduction of crude gum to spirits of turpentine, saving shipping costs and making the process immensely more PROFITABLE. Photo courtesy of U.S. National Archives. 24 I. Introduction

FIGURE 10. The rosin yards at Savannah, GA, in 1893. Every 50-gallon barrel of distilled turpentine contained the entire life’s production of 33 virgin longleaf pine trees, with a by-product of 4 barrels of rosin. Net profit per tree was about 32 cents (Mohr 1893). Photo courtesy of U.S. National Archives. evoked the admiration of the early discoverers and ture of ruin and desolation painful to behold, explorers. No section of the primeval longleaf pine and in view of the destruction of the seedlings forests was more quickly or more effectively oblit- and the younger growth all hope of the re- erated than that through which the “Tar Heelers” forestation of these magnificent forests is expressed on their way from North Carolina to Geor- cluded.” This grim prediction was largely ful- gia. A very few years and they had cut their last filled when the last of the virgin forests were boxes, hacked their last trees, gathered their last depleted in the 1920s. crops of crude gum, and, like an army of locusts leaving a Kansas wheat farm, moved on to fields new and pastures green. The Spread of Agriculture in the Mohr (1896) described the situation in most Longleaf Pine Region of the South by 1896: “. . . the forests invaded by turpentine orcharding present, in five or six Indians were the first farmers, and the full ex- years after they have been abandoned, a pic- tent of Indian agriculture in the South has 2. History and Future of the Longleaf Pine Ecosystem 25

FIGURE 11. This virgin longleaf stand in Beaufort County, SC, had been boxed for turpentine. Fires further weakened the trees by setting the boxes ablaze and in coastal areas, hurricanes often finished the job. Photo, Sherrard 1903. never been delimited. Bartram (1791) de- terrupted agriculture for a number of years scribed “tallahassees” or abandoned Indian and in 1795 it was reported that “all Tidewater old fields in north Florida. To the north, the Virginia was full of ‘old fields’ reverting to tim- hunter-gatherer cultures of North Carolina ber” (Wertenberger 1922). and Virginia farmed on a smaller scale in The longleaf pine region was fully settled patches adjacent to villages, while much of the by 1750 with the exception of Florida, Texas, diet came from fishing and hunting (Harriott and the interiors of Alabama and Mississippi 1590 [1972]; Smith 1624). In the Creek coun- (Fig. 12). As late as 1820 the vast longleaf try of Alabama, however, Bartram traversed forests of the interior of Alabama, Missis- a region of Indian farmland broken only by sippi, Louisiana, and east Texas remained un- small tracts of woods between the outlying touched. In 1821, however, cession of Florida agricultural lands of one village and the next to the United States by Spain, along with major (Bartram 1791 [1955]). Clearly a portion of the land purchases from the Creek and Choctaw longleaf pine region had already been domesti- Indians, opened this region to settlement. By cated long before arrival of the first Europeans. 1850 the fertile Black Belt region of central Along the Atlantic slope, settlers finally be- Alabama and Mississippi had been plowed gan expanding out of the tidewater region in and converted to cotton plantations by large the 1730s (Clay et al. 1975) and, with later slave-holding planters. A map compiled from waves of immigrants, settled the Piedmont, the Census of 1840 (Williams 1980) shows reaching the foothills of the Appalachians by the distribution of major cotton plantations the 1790s (Fig. 4). During the period 1750– in three dense regions: coastal South Carolina 1850 virtually all longleaf communities of the and Georgia, the lower Mississippi River valley, more fertile soils were converted to farm- and the Black Belt. land and pasture (Williams 1989). Both the By the Civil War, nearly all lands optimally American Revolution and the Civil War in- suitable for agriculture were in production. By 26 I. Introduction

FIGURE 12. Virgin longleaf stands of the interior hills of the Piedmont and southern tip of the Appalachians were nearly as open as those of the Coastal Plain. Fresh boxes had just been chopped into the bases of these trees for the turpentine process, which had just reached the hills in 1905. Bibb or Coosa Co., Alabama. Photo, Reed 1905.

1900, 12.5 million hectares, or about 27% of where the harvest could be transported. While the uplands in the former range of longleaf waterpower was tried as early as 1611 in pine upland was listed as “improved” farm- Virginia, this technology did not take hold land, a category that included pasture, roads until around a century later, with introduc- and buildings as well as cropland (U.S. Census tion of water-powered sawmills in Louisiana Office 1902). While there were no separate fig- about 1714 (Hindle 1975) and the Cape Fear ures for land in pasture in 1900, it was nec- region of North Carolina in the 1730s. In essary to maintain pasture or range on ev- 1732, Governor Burrington reported that an ery farm for horses, mules, and oxen used for abundance of sawmills was being constructed plowing and transportation, and until around along the Cape Fear River. In 1764 Gover- 1880 much livestock was still maintained on nor Dobbs reported that 40 sawmills had been open range in the woods. completed on branches of the Cape Fear, and Governor Tryon reported that the number had risen to 50 two years later in 1766 (Merrens History of Logging: Hand Power, 1964). Waterpower opened up the first possibility Waterpower, and Steam of a commercial lumber industry. Steel saw Effects of timbering were minor through the blades were imported from Holland where the early Colonial Period (beginning in 1607 in technology had been developed, and sawmills Virginia, 1565 in Florida) to the mid-1730s, proliferated rapidly along streams in settled when logging was done by hand, using horses, areas. Still, these were slow, straight-bladed re- mules, and oxen to drag the logs. Commercial ciprocating saws (slash saws), with an up and logging was limited to the vicinity of streams down action, mimicking the human-powered 2. History and Future of the Longleaf Pine Ecosystem 27

FIGURE 13. A “carry-log” drawn by mules. Economical range of this kind of transport was less than 4 miles (Croker 1987). Photo courtesy of U.S. National Archives. pit saws: the circular saw and band saw were commercial exploitation was limited to narrow still 100 years away, not coming into general zones along navigable streams. use until after the Civil War (Hindle 1975). Prosperous South Carolinians were fasci- Many of these small mills operated only part nated by steam power and in 1833 constructed time—when there was enough water in the the first railroad in the United States, connect- mill pond in winter and spring to turn the ing Charleston on the coast to the vicinity of wheel. Many were plantation-owned, produc- Augusta on the Savannah River. The entire ing boards for local use, with only a small sur- route lay through longleaf pine country, and plus shipped downstream to coastal towns. As on some of the first runs the engine slowed to a late as 1826, a few decades before the appear- crawl from lack of steam and had to stop while ance of steam-powered sawmills, Mills (1826) hands ran to chop longleaf pine lightwood for commented that the pine timber was still used fuel (Derrick 1930). In 1856, the first steam- mostly for local construction. powered dredges were used in Norfolk County, While waterpower helped the clapboard VA, to build the Albemarle and Chesapeake house replace the log cabin, lumber produc- Canal (Ruffin 1861), and the period 1850– tion remained a minor industry from 1730 1870 saw explosive proliferation of steam tech- to around 1850. Most logging occurred along nology for logging railroads, steam skidders, streams where logs were skidded out by horses, and steam-powered sawmills (Anon. 1907). mules, and oxen. The giant wheeled “carry- By the end of the Civil War, with resump- log” (or “caralog,” Fig. 13) was important tion of intensive turpentining throughout the from this time until the late nineteenth cen- longleaf forests of North and South Carolina, tury when it was supplanted by logging rail- and with steam logging methods perfected, the roads and steam skidders. Logs were dragged stage was set for cataclysmic decimation of the this way to the nearest water and then rafted longleaf ecosystem. downstream to mills. The maximum effective After the war, huge tracts of southern lands distance for this kind of overland transport were bought by railroad companies (Fig. 14). was only 3 or 4 miles (Croker 1987) and so After construction the railroads sold surplus 28 I. Introduction

FIGURE 14. Clearing right-of-way through virgin longleaf forest in Mississippi for the Natchez, Columbia and Mobile Railroad in 1907. All timber was soon cut within several miles of railroads and more distant lands were sold to logging companies. Photo, American Lumberman 1907. lands to logging companies. Lands sometimes 1907). By 1880, all commercial timber had changed hands at the rate of 40,000 hectares or been removed from lands within a few miles of more, at prices of $3 per hectare (Napier 1985). streams and railroads. Tapping of virgin forests The decade 1880 to 1890 saw standardization of the interior had just begun, but huge vol- of track sizes and concatenation of isolated umes of lumber were being produced. Sargent railroad lines, making overland transport of reported an annual cut of over a billion board lumber cheap and efﬁcient (Hale 1883; Anon. feet in 1884 (Table 2), increasing to 3.7 billion

TABLE 2. Virgin longleaf pine remaining in 1880 and annual cut in 1880 (board feet)a Merchantable longleaf pine Annual cut Virginia No reported commercial production North Carolina 5,229,000,000 108,411,000 South Carolina 5,316,000,000 124,492,000 Georgia 16,778,000,000 272,743,000 Florida 6,615,000,000 208,054,000 Alabama 18,885,000,000 245,396,000 Mississippi 18,200,000,000 108,000,000 Louisiana 26,588,000,000 61,882,000 Texas 20,508,000,000 66,450,000 Totals 118,119,000,000 1,194,428,000

a Figures are only for major longleaf pine regions and major logging companies. While virgin growth had been depleted in Virginia and exhaustion in the Carolinas was imminent, stands in Louisiana and Texas still were largely untouched (Sargent 1884). 2. History and Future of the Longleaf Pine Ecosystem 29 board feet by 1896 (Mohr 1896). This phase of second growth, whether after culture, after mere intensive logging, from 1870 to 1930, saw re- cutting down the first growth for fuel, the second moval of virtually all remaining virgin forests growth pines are of the “loblolly” or “old-field” kind, in the South. By 1900, it was apparent that of mean sized appearance. (Ruffin 1843, p. 60) many cutover longleaf areas, particularly those on better soils, were being occupied by scrubby Mohr (p. 64) commented, “on the lowlands second growth of other species, while some re- of the Atlantic coast toward its northern limit mained open and nearly treeless. To the grow- this pine is almost invariably replaced by the ing concern of foresters, longleaf pine replaced Loblolly Pine.” “In the stronger soil of the up- itself only sporadically in a small percentage of per division of the maritime pine belt, the re- its former landscape (Mohr 1896). Given the gion of mixed growth, where seedlings of the vast extent of longleaf once reproducing nat- Longleaf Pine spring up simultaneously with urally in primeval forests, what could explain the hard wood trees and the seedlings of the its failure to do so now? Shortleaf Pine, these latter will eventually gain the supremacy and suppress those of the Lon- gleaf Pine.” “It is evident that the offspring of The Disappearance the Longleaf Pine is rarely seen to occupy the place of the parent tree, even in the region of Longleaf Pine most favorable to its natural renewal, and that final extinction of the forests of the Longleaf Failure of Longleaf Pine Pine is inevitable unless proper forest man- Regeneration after Logging agement is applied.” To Mohr’s mind proper management meant eliminating all fire, en- Historical records suggest that two factors com- couraging 15 to 20 years later, shade-tolerant bined to explain the final disappearance of tree species below the longleaf to build up a longleaf pine after initial exploitation for tur- humus layer “to secure improvement and per- pentine and lumber. First was the fondness of manency of favorable soil conditions.” These feral livestock, especially hogs, for the seed- sentiments were echoed by Sherrard (1903). lings (Mohr 1896; Hopkins 1947a,b,c). Unlike Unfortunately, this was a prescription for ex- other pines, longleaf seedlings have a non- tirpation of longleaf pine. resinous, carbohydrate-rich meristem, which, The question that dogged foresters was, why while in the grass stage, is vulnerable to grazing did longleaf not reproduce, at least on those for 5 to 7 years or more. Hogs have been ob- lands where nothing else was done other than served to feed heavily on longleaf seedlings, logging of the virgin timber? Contemporary consuming up to 400 each in a day (Hopkins with Mohr, one of the first foresters to wres- 1947a,c). The second and final nail in the coffin tle with this problem was W. W. Ashe, who was twentieth-century fire suppression. noted that not only was the longleaf seed crop By the 1890s foresters saw clearly that, over produced in irregular mast years, but also that large expanses of the landscape, longleaf was the seeds were descended upon by a vari- not replacing itself after logging (Ashe 1894a,b; ety of predators: “. . . its large and sweet seeds Mohr 1884, 1896). On the road on the ridge are eaten in large quantities by fowls of vari- between the Cooper and Ashley rivers out of ous kinds, rats, squirrels, and by swine, which Charleston, Edmund Ruffin observed changes prefer them to all other kinds of mast, and in the forest, on lands long settled: when there is enough long leaf pine mast become very fat on it” (Ashe 1894b, p. 57). This The trees are nearly all pine, & generally of second growth, the land having been formerly cultivated & had been noticed as early as 1728 by William afterwards turned out. Byrd during the survey of the Virginia–North The pines of original forest are mostly of the ‘long Carolina line, and Ruffin (1861) commented leaf’ species, & many of the great size & beauty for that “[t]hey are so eagerly sought for by hogs which that kind is distinguished. But whenever of that scarcely any are left on the ground to 30 I. Introduction germinate” (p. 255). Ashe was one of the first was walled off were not sufficient to keep the to report the fondness of hogs for the larger hogs out of the streets of Jamestown, Virginia. seedlings. “No sooner, however, has the young Capt. Samuel Argall and James Rolfe on land- pine gotten a foot high and its root an inch in ing there in May of that year, only 10 years diameter than the hog attacks it, this time eat- after settlement, commented on the “innumer- ing out the roots, which until two inches in able numbers of swine” (Smith 1624). diameter, are very tender and juicy, pleasantly flavored and free of resinous matter” (Ashe 1894b, p. 57). Evidence for Early Saturation of Like most foresters of his time, Ashe re- the Landscape by Hogs in garded fire as the unrelenting enemy of forest regeneration, even going so far as to insist Coastal Regions that in North Carolina “. . . the burnings of the Both the Spanish and English experiences present and future, if not soon discontinued, demonstrated the potential of hogs to in- will mean the final extinction of the long leaf crease from a handful to thousands in a few pine in this state” (Ashe 1894b). This opinion years under conditions of complete neglect on echoed that of Mohr (1884) and others on the open range. By 1702 a Swiss visitor to coastal destructive nature of fire. The groundwork for Virginia declared that “pigs are found there in the field of fire ecology had clearly not yet been such numbers that I was astonished” (Michel laid. 1702 [1916]). This was corroborated by Ashe concluded that the chief agencies pre- Beverley (1705 [1947]) who stated that “hogs venting regrowth of longleaf pine were fire and swarm like Vermine upon the Earth. . . . The hogs. In contrast, later authors asserted the ac- Hogs run where they list, and find their own tual dependence of the species upon fire to pre- Support in the Woods, without any Care of vent site appropriation by shade-tolerant pines the Owner; and in many Plantations it is and hardwoods (Harper 1913). When some of well, if the Proprietor can find and catch the the early assertions were tested in southeast- Pigs, or any part of a Farrow when they are ern South Carolina, longleaf pine was found young, to mark them. . . . ” A few years later, to be replaced by slash pine when both fire Brickell (1737 [1968]) reported similar condi- and hogs were excluded (Sherrard 1903), and tions in northeastern North Carolina where he studies in 1935 showed only 8% fire mortality saw “. . . swine, breeding in vast numbers. . . . ” in 2-year-old longleaf plantations in Louisiana, A considerable meat packing business had versus 53% for 7-year-old loblolly (Wakeley sprung up in Norfolk, VA, the major seaport 1935). If fire is excused as one of the two prin- in the mid-Atlantic region, to supply salt pork ciple culprits, that leaves hogs conspicuously to sailing ships. The first direct evidence that in need of closer scrutiny. hogs had reached saturation density in North In 1539, DeSoto made the first introduction Carolina is provided by the report of Gov- of swine to the South (Bakeless 1961). Later, ernor Barrington in 1733, that about 50,000 English settlements brought with them starter hogs were driven annually to the Norfolk livestock (Strachey 1610 [1964]; Smith 1624). market from the Albemarle region of North Hogs showed an astounding reproductive po- Carolina (Wertenberger 1931). The first live- tential, and demonstrated an ability to fend stock census figures from these six small coun- entirely for themselves in the woods with no ties showed no increase in hog numbers from attention from their owners (Beverley 1705 1840 to the Civil War, indicating that satura- [1947]; Blakeley 1812 [1910]). The capacity tion density had been reached, with an average of the landscape to support open range hogs of 14,800 hogs on open range in each of the six has never been investigated, but the evidence counties south of the state line within hog driv- suggests that they quickly reached saturation ing range of Norfolk. This gives an average of density within a few decades after settlement. 4.3 hectares per hog (U.S. Census 1841). For By 1617 the log palisades with which the town the 1890 census only, supplementary figures 2. History and Future of the Longleaf Pine Ecosystem 31 were kept for hogs consumed or hogs that died. In Alabama, which still had hogs on open range, an annual number equal to 45% of the total hogs alive were consumed and 23% died. This gives us an approximation for surplus hogs that could be harvested when populations were near capacity (U.S. Census Office 1902: U.S. Census of Agriculture for 1902). If the total number of hogs in the six North Carolina counties mentioned above were at carrying capacity in 1750, the numbers should be nearly the same as in 1840 (88,850 hogs), then the surplus should have been 45%, or 40,000 hogs. The fact that the reported surplus of 50,000 fully grown hogs driven to Virginia FIGURE 15. Evidence for saturation of the land- exceeds our estimate of 40,000 strongly sug- scape by feral hogs. The lower curves represent sta- gests that carrying capacity had been reached ble hog populations in coastal regions long-settled by 1840—more than 200 years for coastal Virginia in this region sometime before 1733. These (bottom line) and over 100 years for coastal Al- counties were settled between the years of abama (middle line). The vast regions in central 1655 and 1700 so there had been from 35 Alabama, only opened to settlement in 1821, had to 78 years, easily sufficient for hogs to reach just reached carrying capacity in 1850, with over a saturation density. million hogs on open range. The dip from 1860 to While hogs spread inland from southeast- 1870 was the result of overconsumption of all live- ern Virginia and northeastern North Carolina, stock during the starvation that accompanied the other introductions were made along the Civil War. Data from U.S. Censuses of Agriculture, Atlantic and Gulf coasts. In 1663, explorers 1840–1890. Reprinted from Frost 1993 with per- stepping ashore on the barrier island at Cape mission from the Tall Timbers Research Station. Fear, NC, were astonished at being offered pork for sale by the Indians, livestock having been regularly kept figures, however, were not placed on the islands a few years earlier by available until a century later with the 1840 stockmen from New England (Lawson 1709 Census of Agriculture. The lower line in Fig. 15 [1967]). On the Gulf Coast, Mobile, founded shows the total number of hogs from the 15 in 1711, was the first permanent city (Hamil- Virginia counties within the original range of ton 1910 [1976]), and in 1812, free-ranging longleaf pine from 1840 to 1900. The plunge in hogs were kept on three islands of about 1600 numbers occasioned by famine during the Civil hectares each at the head of Mobile Bay. Josiah War is characteristic of all the southern states Blakeley, the owner, wrote that “[c]attle and and is closely paralleled by figures for cattle and hogs do well upon them, and no expense. other livestock (U.S. Census Office 1841, 1853, Upon them I have about 30 head of cattle 1864, 1872, 1883, 1895, 1902). Note that the and hundreds of hogs, the hogs wild. I shoot population curve for the decades preceding the or catch them with a dog” (Blakeley 1812 Civil War is flat, and recovers to a relatively flat [1910], p. 405). There is no evidence, however, slope within two or three decades afterward. that hogs spread very far beyond the frontier, This supports the notion that carrying capac- where Indians and other predators likely kept ity had been reached some time before such them under control. records were kept. From the descriptions above, it seems likely In contrast, figures for Alabama indicate that that tidewater Virginia was saturated with hogs only the coastal region was saturated by 1840. by around 1700, and the whole coastal plain The middle line in Fig. 15, which parallels that of Virginia and the portion of North Carolina for Virginia, represents the number of hogs north of Albemarle Sound by 1730. The first in the seven old, long-settled coastal counties 32 I. Introduction near Mobile. The upper line represents the part of the landscape in which longleaf pine middle counties. The interior remained Span- flourished. Wakeley (1954, p. 151) considered ish territory until 1821, when settlers from hogs by far the most serious threat to longleaf: Georgia and the coast were poised for entry “Where there are many hogs it is foolhardy (see Fig. 4). By 1840, only 19 years after open- to plant longleaf pine without fencing . . . . To ing of the territory, immigration was in full this species hogs are infinitely more destructive swing but the country was still sparsely settled. than fire.” Figure 15 shows increasing numbers of hogs There are several hog-and-fire exclusion in the central counties, but leveling off after studies to back up this assertion, two of which 1850, within 19 years of the 1821 opening of reported complete failure of stand regenera- the land to settlement (the large numbers for tion on tracts where feral hogs were present. the central counties reflect the much greater Two experimental tracts at Urania, LA, after land area). The flattening of the curve again 5 years of protection against hogs, contained suggests carrying capacity had been reached, an average of 16000 longleaf saplings per ha, and demonstrates the capacity of hogs to satu- as compared with an average of only 20 per ha rate a vast landscape in less than 20 years. on two unprotected tracts (Mattoon 1922). In Hogs were not the only competitor for forage an area with free-ranging hogs in Georgetown on open range, however. While hogs were con- County, SC, hogs were fenced out of 32 0.04 ha sistently the most abundant livestock species plots. After two growing seasons the fenced ar- reported in the agricultural censuses, the range eas contained 1200 large seedlings (those with was shared by cattle, horses, mules, sheep, root collar diameters of 1.3 cm or larger) per and goats, whose numbers collectively equaled ha, while unfenced areas contained only 20 per those of hogs (U.S. Censuses 1841, 1853, 1864, ha (Lipscomb 1989). The hogs largely ignored 1872, 1883, 1895, 1902). small first-year seedlings but focused on those One writer estimated that 5–10 hectares of large enough to have accumulated starchy root unmanaged southern woodland was required content. Density of hogs was not controlled but to support one cow, while 0.8 hectare of good was estimated to be about three to six animals pasture would suffice (Gardner 1979). No fig- on the 24-hectare study area, or 4–8 hectares ures were ever determined for carrying capac- per hog. This is comparable to the hog densi- ity of southern range for hogs (Grelen 1980). ties of 4.3 hectares per hog reported above, on As noted above, the apparent saturation den- open range in colonial North Carolina, which sity of hogs in 1840 in northeastern North we have suggested may represent carrying Carolina was 4.3 hectares. While this might capacity. seem an abundance of land per hog, keep Ashe (1894b) and Mohr (1896) both com- in mind that there was an equal number of mented on the palatability of longleaf pine other open range livestock competing for food, seedling roots in the 1.5 to 5 centimeter diam- the county acreage included water, areas from eter range. Wakeley (1954) reported hog con- which hogs were fenced out, and large areas sumption of 200 to 1000 longleaf seedlings per of upland forests where there may have been day, at rates of up to 6 per minute. Hopkins little forage except for the fall mast crop of (1947a,b,c), after observing hogs rooting up acorns and pine seeds. There was also stiff com- hundreds of seedlings a day, analyzed the petition for the mast crop from birds and na- root starch content and found them to be as tive animals (Ashe 1894b; Wahlenburg 1946). nutritious as corn. Little wonder then that Longleaf pine seedlings, on the other hand, hogs would be drawn to longleaf seedlings, were available and vulnerable all year round. which, in the grass stage, are highly conspic- While birds have been observed to consume uous and vulnerable for 3 to 7 years. With from 8% to up to 42% of the longleaf seed 10,000 to 40,000 hogs on open range in ev- crop (Wahlenburg 1946), they do not molest ery settled county in the longleaf region (U.S. the seedlings, and this much predation must Censuses 1841, 1853, 1864, 1872, 1883, 1895, have been tolerable, since birds were a natural 1902), all that would be required to eliminate 2. History and Future of the Longleaf Pine Ecosystem 33 reproduction would be for a drove of hogs to of livestock on open range. For the first three happen upon a regenerating plot once every centuries, crops had been fenced in to protect 3 or 4 years to largely eliminate the species them from livestock, which had free run of the from the landscape. land. Even if a farmer had little stock of his Hogs on open range were completely depen- own, he had no choice but to fence his crops dent on natural forage. If carrying capacity had against the animals of his neighbors. As more been reached, survival would be tenuous and land came into agriculture, demands for fenc- occasional disasters could be expected when ing increased until the timber shortage made mast crops or other wild foods failed. A cu- it apparent that it would more economical to rious example occurred in Illinois when hogs fence in the livestock rather than the crops. starved in winter after passenger pigeons un- In response, fence laws (stock laws) were expectedly descended on a local area and ate all passed throughout the South, beginning in the the fall mast of acorns, beechnuts, and chest- 1870s. In 1883 a statewide law was passed in nuts (Bakeless 1961). This raises the ques- South Carolina making it incumbent upon the tion about the reverse situation, that satura- owners of livestock to see that they do not tres- tion of the landscape with hogs contributed to pass on the lands of others. A respondent to the extinction of the passenger pigeon. Their an 1880 timber survey, from Anson County, summer breeding range extended only as far NC, commented that “every man who owns south as Virginia but from late September to cattle, hogs, sheep, goats or horses in Anson early November the flocks migrated to the County is now compelled to pasture them on winter range from South Carolina to Florida his own land. None are allowed to run at large (Bent 1932 [1963]). This coincided with long- on the range. This system came into effect in leaf seed fall, and it has been observed that re- our county about two years ago, and so much lated birds like mourning doves and quail have is it esteemed already that a return to the old their crops “crammed” with longleaf seeds dur- style of fencing the crops against the incursions ing this time (Wahlenburg 1946). The distinct of stock is next to impossible. This is regarded parallel between the decline of longleaf pine, a as the most important single step taken in this major winter food source, and that of the pas- county in the last twenty years” (Hale 1883). senger pigeon may not be a coincidence. In the The process took decades to become effective South, memory of the species persists only in over the whole South and there are still some place names like “Passager Swamp” in Isle of areas where hogs run wild (Lipscomb 1989). Wight County, VA.4

The End of Open Range Landscape Changes from The effects of hogs on longleaf pine were not 1900 to 2000 noticed until the massive wave of logging that followed the Civil War physically removed Fire Suppression and the the forest. Most of the timber cut in the pe- Decline of Fire as a Natural riod 1870–1900 was still virgin forest (Mohr Determinant of Vegetation 1896), where the effects of hogs in eliminating seedlings could be overlooked as long as the The end of open range should have been a trees stood. Note that longleaf had indeed been boon to longleaf pine, but while three cen- extirpated from much of the northern range turies of open range were drawing to a close, a a hundred years before, but the process had new threat was in the making. Fire was still taken 200 years, while decimation of the for- widespread, but by the Civil War, much of est using steam-logging technology seemed to the landscape had been fragmented by agri- occur overnight. This precipitated an immedi- culture, reducing the size of fire compart- ate shortage of lumber for fencing (Hale 1883), ments. In central South Carolina there were an and forced landowners to look at the problem average of 20.3 hectares per farm cleared and 34 I. Introduction tilled (State Board of Agriculture 1883). As this condition with a vengeance in the next long as raising stock was the primary source 30 years. In 1919, Virginia passed laws creat- of income the remaining woodlands were ing the position of State Forester and provided burned by the residents to green up forage for forest wardens. The act also imposed fines for livestock. This practice may have perpet- and a minimum penalty of a year in prison for uated longleaf pine and its associated flora of maliciously starting a forest fire, a far cry from wiregrass and savanna herbs, in a landscape the days when burning was a casual manage- where roads, plowed fields, and other man- ment practice. made firebreaks fragmented the fire landscape Few of the early foresters cared to acknowl- and eliminated landscape-scale fires ignited by edge the role of lightning as an ignition source. lightning. When cattle grazing declined in im- In South Carolina, Sherrard (1903) blamed all portance after the Civil War, the practice of fires on humans, stating that fires were “care- spring burning was abandoned in major agri- lessly set to improve grazing, to clear land, cultural areas. Describing the resultant vege- and to protect woods where turpentine is be- tation changes in South Carolina, one writer ing gathered.” Burning in this case was done noted that “the uplands were covered, as they after first raking pine straw away from the still are, with a large growth of yellow pine, flammable boxes in the bases of the trees. Ashe but a deer might then have been seen, in the even believed that one of the reasons longleaf vistas made by their smooth stems, a distance pine was being replaced by loblolly was that it of half a mile, where now, since the discon- was more sensitive to fire: tinuance of the spring and autumn fires, it could not be seen fifteen paces, for the thick The loblolly pine is less injured by fire because its growth of oak and hickory that has taken bark is thicker and so offers more protection to the the land” (State Board of Agriculture 1883, growing wood, –the bark, too, lying closer to the p. 79). wood in firmly appressed layers, does not so easily On all but the drier lands, longleaf reproduc- take fire. tion is completely eliminated by other pines The chief agencies, then, which prevent a re- and hardwood, and shrub invasion within growth of long leaf pine on the high sandy lands, a few years after fire exclusion (Sherrard are the hogs and the fires . . . the burnings of the 1903). Nowhere in the South can longleaf be present and future, if not soon discontinued, will seen reinvading the mesophytic mixed pine– mean the final extinction of the long leaf pine in hardwood succession that has replaced it. this State. (Ashe 1894, p. 58, writing about North Modern fire laws and the state apparatus for Carolina) prevention and suppression of wildfire did not come into being in most of the South until In contrast, Sherrard observed that “the Long- the period 1910–1930. This left a window of leaf Pine may rightly be called a fireproof some 50 years, between the end of open range species in so far as the survival of scattered around 1880 and the beginning of twentieth- groups and patches of second growth and indi- century fire suppression, in which longleaf viduals is concerned.” Still, he was one of the pine had a safe opportunity to reproduce. first to call for a public campaign: “the peo- Many of the stands that did result have now ple must be educated to a sentiment against been logged and the oldest of those naturally fires.” regenerated stands still remaining, date to the The first voice to clearly distinguish the nat- end of this window of opportunity. ural role of fire was Roland Harper, who stated, Fernow (1893) was one of the first to ar- “it can be safely asserted that there is not and gue for governmental involvement in forestry: never has been a long-leaf pine forest in the “there exist some legislative provisions regard- United States . . . which did not show evidences ing forest fires in almost every State, but they of fire, such as charred bark near the bases of are rarely if ever carried into execution for lack the trees; and furthermore, if it were possible of proper machinery.” Most states remedied to prevent forest fires absolutely the long-leaf 2. History and Future of the Longleaf Pine Ecosystem 35

FIGURE 16. The first documented study showing the effects of exclusion of fire and hogs from longleaf pine. A dense forest of slash pine is regenerating in a fenced plot after exclusion of fire and hogs for several years. Old boxed longleaf survivors and scattered slash pine make up the canopy (Sherrard 1903). Sherrard aspired to produce a similar forest on all pine lands in the two counties being studied and fire exclusion became the general forest prescription for the South. Southeastern South Carolina. Photo, Sherrard 1903. pine—our most useful tree—would soon be- legitimate concern, and foresters were hungry come extinct” (Harper 1913, p. 16). for solutions. While most were convinced that If not recognized by early foresters, it was both hogs and fire were inimical to longleaf well known to inhabitants of the longleaf pine regeneration, the first real demonstration was region as early as the 1830s that lightning was conducted in 1903. Sherrard (1903) examined often responsible for fires in the “turpentine a fenced plot from which fire and hogs were orchards.” On a large estate in Onslow County, excluded. Within a few years a dense stand of NC, damage to the turpentine crop was pre- slash pine had established itself beneath the vented by providing log cabins free of rent longleaf (Fig. 16). Sherrard was pleased with to poor white families, whose duties included the result. Never mind that the new forest fighting summer lightning fires: would be composed of a new dominant species and of entirely different structure than the These men are required to do three things: first, they open longleaf forests. And curiously, neither are to guard the orchards from fire, and if a small he nor Ashe nor Mohr ever questioned that if fire occur, as it often does in the summer time by fire were the enemy of longleaf, why did its ex- lightning striking and igniting a resinous pine tree, clusion lead to an entirely different forest type? they and their families must extinguish it. If it gets While it must have been apparent that this beyond their control they are to blow horns, sum- moning the neighboring tenants, sending all around kind of succession would eventually lead to refor help, fight the fire until it is put out . . . (Gamble placement of longleaf, it was sufficiently good 1921, p. 27) news in a landscape recently denuded of its primeval forest cover, that within a few years, The slow and patchy reproduction charac- fire exclusion and a program of educating the teristic of unmanaged longleaf under condi- public “to a sentiment against fires” became tions of frequent growing season fires was a the general forest prescription for the South. 36 I. Introduction

TABLE 3. The ﬁrst pine plantations: 1892–1931a 1892–1928 1928 1929 1930 1931 Total Virginia 136 19 141 128 162 587 North Carolina 618 124 220 109 190 1,261 South Carolina 1,308 — 45 195 301 1,850 Georgia 608 2 324 1,030 62 2,026 Florida 391 0 14 595 756 1,756 Alabama 36 20 133 108 14 311 Mississippi — — — 217 241 457 Louisiana 7,914 3,756 4,286 4,298 1,002 20,018 Texas — — — 105 — 105 28,371 ha

a Figures are from Wakeley (1935), with exception of the area planted before 1928 in South Carolina, from Boyce (1979).

Pine Plantation loblolly and slash pine, but also small amounts of shortleaf and longleaf (based on figures and Pine plantations scarcely existed in 1900. The projections in Boyce 1979; McWilliams 1987; earliest plantations of record in the South Outcalt and Sheffield 1996). were three small plots established by farmers in 1892, 1896, and 1907 (Wakeley 1935). The first large attempt at plantations by the Expansion of Agriculture and U.S. Forest Service, 365 hectares on the Choctawhatchee and Ocala National Forests in Developed Land 1911, proved a failure. Wakeley knew of only While much mixed pine–hardwood is now 200 hectares successfully established by 1919. converted to plantation after logging, some is Problems with technique were soon worked also cleared and converted to cropland or pas- out, however, and Table 3 shows the extent ture. While commercial dairy operations have of pine plantation in the nine states within proliferated since 1900, total pasture and crop- the range of longleaf pine by 1931.5 By this land have declined. After World War II, mules time more than 20 lumber and paper compa- and horses were retired by tractors, and sur- nies were involved and they accounted for at plus pasture lands went into cropland or suc- least 78% of the area planted. ceeded to loblolly pine and hardwoods (Boyce Fire was a threat to pine plantations, but and Knight 1980). The relative percentages of establishment of increasingly large areas pro- land in cropland and forest are the net result tected from fire in the 1930s and 1940s made it of a complexity of changes that include for- seem feasible to plant loblolly and slash pine as est succession of abandoned cropland on small commercial crops. Pine planting was expanded uncompetitive farms between 1940 and 1965, by large timber corporations in the 1940s and clearing of new cropland from woodland and 1950s, and there were 12,460,000 acres by large farming operations. Agricultural land (5,046,300 hectares) established in the years area peaked in 1930 and has been reverting to 1965 to 1967 (Boyce 1979). Forced into more forest and other land uses ever since (Williams marginal lands by development pressures, tim- 1989). The 1997 Census of Agriculture re- ber companies found it increasingly desirable ported 3,456,000 hectares in pasture (7% of to produce pine pulpwood and sawtimber us- the uplands) and 6,027,000 hectares in crop- ing intensive management. In the former long- land (12% of the uplands) in the portions of leaf region, there are at present about 11 mil- the 412 counties included in the former long- lion hectares of pine plantations, primarily leaf pine region (Table 1). 2. History and Future of the Longleaf Pine Ecosystem 37

TABLE 4. Fire regime conditions in 785 stands of longleaf pine in the northern range of the speciesa CC1 CC2 CC3 CC4 CC5 Virginia 012218 North Carolina 42 60 60 77 116 South Carolina 17 16 114 83 104 Northeast Georgia 12 4 17 17 23 Totals 71 81 193 179 261

a Condition Class 1 contains stands that had been burned often enough to have retained at least 70% of their original plant species diversity. Condition Class 2 indicates longleaf dominant stands with loss of more than 30% of species but that had been burned recently enough to retain conspicuous fire char on their trunks. CC3 stands were also dominated by longleaf but fire had been excluded long enough for the understory to fill in with dense woody vegetation: with no other treatment, these stands, when next logged for longleaf pine, will largely convert to hardwoods and early successional pines such as loblolly, with a few residual stems of longleaf. This successional process also accounts for most of the CC4 stands which consisted of scattered longleaf pines in fire-suppressed stands dominated by other species. CC5 were former stands in which fewer than 10 longleaf pines were found, in some cases consisting of a single ancient boundary line tree in a forest completely converted to other types.

Fire Regimes Today and the full complement of plant species that they once Condition of Remnant Longleaf supported under the natural fire regimes indicated in Fig. 3. The stands in Condition Class 2 Pine Communities had experienced some reduction in fire fre- The few substantial, well-maintained rem- quency and many were stands to which fire nants of longleaf pine communities are now had been reintroduced after a long period of found primarily on military bases whose man- fire exclusion in the mid to late twentieth agers have sufficient fire staff to maintain ef- century. Of these, most had lost more than fective fire regimes. Smaller, fire-maintained 50% of their understory species diversity. In examples can be found locally on national most natural longleaf pine communities, more forests and other public lands and private pre- than 90% of the plant species diversity is found serves, and fire programs are now gearing up in the herb layer, as a rich assortment of native for restoration of natural stands suffering from grasses and forbs. Most of the rare species are various stages of fire regime alteration. also found in this layer (Walker 1995). In the Over the 25-year period 1978–2003, I ex- worst case, during the initial stages of reintro- amined 785 stands of longleaf pine ranging duction of fire, I saw several stands during this from the northernmost remaining tree in Isle survey with not a single herbaceous species in of Wight County, VA, to stands in north- a study plot. east Georgia (Table 4). I evaluated each in Longleaf pine has been extirpated from all terms of its departure from the natural fire but about 2.2% of its original range (exclud- regime. Stand investigation ranged in intensity ing recent plantations), or about 1,050,000 from detailed 1/10-hectare study plots to 100- hectares. Of that fraction, only about 19%, square-meter plots or quick visual evaluations. or 193,000 hectares, is currently being main- By 2000, only 19% of remnant stands in tained with fire, and only 9% has escaped sig- the northern range of longleaf pine were be- nificant loss of species diversity resulting from ing maintained with fire (Classes 1 and 2), and episodes of fire suppression. Fire-suppressed even this interpretation is optimistic. Only 9% stands typically were invaded by hardwoods, of stands retained something approaching the loblolly pine, sweetgum, or slash pine. Instead 38 I. Introduction of the two-layered structure typical of natural being bought or protected, the dense thick- longleaf communities, there were heavy shrub ets resulting from decades of fire exclusion and midstory layers. The resulting shade, along are being subjected to midstory thinning, and with deep pine needle litter and duff accumu- fire is being restored to the land. The newest lation, had completely eliminated wiregrass efforts include reintroduction of grasses and and most of the rest of the herb layer on many other herb layer species. Recent government sites. With only 9% of remnant stands in Con- actions mandate the determination of original dition Class 1, that means that less than 96,572 fire regimes and Fire Regime Condition Class hectares, or less than 0.2% of the original (FRCC)—the degree from which current fire extent of the longleaf pine ecosystem, remains regimes and stand conditions have departed in condition good enough to support most of from that in nature (Hann 2002). As FRCC its native plants and animals. is determined for lands across the country we The logging boom of the late nineteenth will then have targets for restoration of the fire century left in its wake cutover lands and regimes that thousands of species rely upon. dense, scrubby second growth, and efforts of This gives us some cause to hope that 2000 crusading fire exclusionists guaranteed that represented the low point for the longleaf pine over much of the region, the sunny, open, ecosystem. fire-maintained woodlands would be seen no Within the 2.2% (1.01 million hectares) of more. For the inhabitants who lived during the the landscape that still supports natural long- first decades, seeing the forest of centuries fall leaf pine today, there is a remarkable galaxy around them was often a disheartening expe- of sites large and small, only one generation rience that transformed their world. One re- away from logging and turpentining, some of spondent to a timber survey in 1882 in Cur- which have recovered nicely. These we may rituck County, NC, noted bitterly: still be able to maintain, and perhaps we can restore more of Bartram’s “. . . expansive, The avaricious and insatiable saw mills, together with the desire of every man who could buy a pair of airy pine forests . . . of the great long-leaved oxen and “Carry-Log”, have demolished and trans- pine . . . the earth covered with grass, inter- ported nearly all of our pine ....This certainly looks spersed with an infinite variety of herbaceous like a gloomy report, but more truth than poetry. plants, and embellished with extensive savan- (Hale 1883, p. 222) nas, always green . . . .”

The Future of the Longleaf Endnotes Pine Ecosystem 1. An earlier version of this chapter was first published in Proceedings of the Tall Timbers Fire Ecol- If less than 0.2% of the original extent of long- ogy Conference, No. 18, The Longleaf Pine Ecosys- leaf pine remains in condition good enough to tem: ecology, restoration and management, ed. support a significant diversity of their native Sharon Hermann, Tall Timbers Research Station, plants and animals, then the few areas that Tallahassee, FL, 1993. have been burned often enough to have re- 2. Gold or other coin, “hard cash.” tained their full complement of species are ex- 3. I am grateful to the staff of the Colonial Williams- ceedingly valuable—as refugia for species, and burg Foundation Library, Williamsburg, VA, for as reference communities for setting restora- access to the original records on microfilm. 4. The term “passenger” pigeon is a pejoration of tion targets for the rest of the longleaf pine the original word “passager,” as used in Colonial landscape. With so few remnants, we are now times. Most people only saw them as birds of pas- compelled to make every effort to get fire back sage since the great flocks, except for a few weeks into all remaining longleaf stands. while nesting, were constantly on the move in Encouraging signs for the future are now ap- search of food. As a consequence they were called pearing on public lands. Remnant stands are “passager pigeons.” 2. History and Future of the Longleaf Pine Ecosystem 39

5. A small amount of planted trees were hardwood, Clay, J.W., Orr, D.M., Jr., and Stuart, A.W. 1975. something under 5%. North Carolina Atlas. Chapel Hill: University of North Carolina Press. Cloud, P. 1976. Beginnings of biospheric evolution References and their biogeographical consequences. Palaeo- biology 2:351–387. Anon. 1907. A trip through the varied and extensive Cope, M.J., and Chaloner, W.G. 1980. Fossil char- operations of the John L. Roper Lumber Com- coal as evidence of past atmospheric composition. pany. Am Lumberman [n.v.n.], April: 51–114. Nature 283:647–649. Ashe, W.W.1894a. The long leaf pine and its struggle Croker, T.C., Jr. 1987. Longleaf pine, a history of for existence. J Elisha Mitchell Soc 11:1–16. man and a forest. USDA Forest Service, Southern Ashe, W.W. 1894b. The Forests, Forest Lands and For- Forest Experiment Station, Forestry Report R8- est Products of Eastern North Carolina. Raleigh, NC: FR 7. Josephus Daniels, State Printer. Davies, W. 1781 [1881]. Letter to D. Jamieson. Cal- Bakeless, J. 1961. The Eyes of Discovery. New York: endar of state papers [Virginia] 2:599. Dover Publications. Delcourt, H.R., and Delcourt, P.A. 1977. Pre- Barden, L.S. 1997. Historic prairies in the piedmont settlement magnolia-beech climax of the Gulf of North and South Carolina, USA. Nat Areas J Coastal Plain: Qualitative evidence from the 17:149–152. Apalachicola River bluffs, north-central Florida. Bartram, W. 1791 [1955]. Travels Through North Ecology 58:1085–1093. and South Carolina, Georgia, East and West Florida. Derrick, S.M. 1930. Centennial History of the South Car- Reprint. New York: Dover Publications. olina Railroad. Columbia, SC: The State Co. Bent, A.C. 1932 [1963]. Life Histories of North Amer- Earley, L. 2004. Looking for Longleaf: The Fall and Rise ican Gallinaceous Birds. Reprint. New York: Dover of an American Forest. Chapel Hill: University of Publications. North Carolina Press. Beverley, R. 1705 [1947]. The History and Present Fernow, B.E. 1893. Results on investigations on State of Virginia. Chapel Hill: University of North long-leaf pine. In Timber physics, Part II. Forestry Carolina Press. Division, Bulletin No. 8. Washington, DC: US De- Blakeley, J., dated February 28, 1812, at Mobile. In partment of Agriculture. Hamilton, P.J. 1910 [1976]. Colonial Mobile. Uni- Frost, C.C. 1993. Four centuries of changing land- versity: University of Alabama Press. scape patterns in the longleaf pine ecosystem. In Boyce, S.G. 1979. Prospective ingrowth of southern The Longleaf Pine Ecosystem: Ecology, Restoration and pine beyond 1980. USDA Forest Service, South- Management, ed. S.M. Hermann, pp. 17–43. Pro- eastern Forest Experiment Station Research Pa- ceedings Tall Timbers Fire Ecology Conference per SE-200. No. 18. Boyce, S.G., and Knight, H.A. 1980. Prospective Frost, C.C. 1995. Presettlement fire regimes in ingrowth of southern hardwoods beyond 1980. southeastern marshes, peatlands and swamps. In USDA Forest Service, Southeastern Forest Exper- Fire in Wetlands: A Management Perspective, eds. S.I. iment Station, Research Paper SE-203. Cerulean and R.T. Engstrom, pp. 39–60. Proceed- Brickell, J. 1737 [1968]. The Natural History of North ings Tall Timbers Fire Ecology Conference No. 19. Carolina. Reprint. Murfreesboro, NC: Johnson Frost, C.C. 2000. Studies in landscape fire ecology Publishing Company. and presettlement vegetation of the southeastern British Public Records, Colonial Office. 1607–1783. United States. Doctoral dissertation. Chapel Hill: Williamsburg, VA: Original records on microfilm University of North Carolina. at Colonial Williamsburg Foundation Library. Frost, C.C., and Musselman, L.J. 1987. History and Byrd, W. 1728 [1967]. Histories of the Dividing Line vegetation of the Blackwater Ecological Preserve. Betwixt Virginia and North Carolina. Reprint. New Castanea 52:15–46. York: Dover Press. Frost, C.C., Legrand, H.E., Jr., and Snyder, R.E. Catesby, M. 1731. The Natural History of Carolina, 1990. Regional inventory for critical natural ar- Florida and the Bahama Islands. Vol. I. London: M. eas, wetland ecosystems, and endangered species Catesby. habitats of the Albemarle–Pamlico estuarine re- Catesby, M. 1743. The Natural History of Carolina, gion: Phase 1. Raleigh, North Carolina: U.S. En- Florida and the Bahama Islands. Vol. II. London: M. vironmental Protection Agency and N.C. Dept. Catesby. of Environment, Health and Natural Resources. 40 I. Introduction

Albemarle–Pamlico Estuarine Study, Project No. Harriott, T. 1590 [1972]. A Briefe and True Report of 90-01. the New Found Land of Virginia. Reprint. New York: Gamble, T., ed. 1921. Naval Stores: History, Produc- Dover Press. tion, Distribution and Consumption. Savannah, GA: Hart, J.F. 1979. The role of the plantation in south- Review Publishing & Printing Co. ern agriculture. Proc Tall Timbers Ecol Manage Conf Gardner, A. 1979. A cow in a woodlot is as welcome 16:1–19. as a bullfrog in a punchbowl. Westvaco CFM News Hefner, J.M., and Brown, J.D. 1985. Wetland trends [n.v.n.] summer 1979. in the southeastern United States. Wetlands 4:1– Garren, K.H. 1943. Effects of fire on vegetation of 11. the southeastern United States. Bot Rev 9:617– Hindle, B. 1975. America’s Wooden Age—Aspects of Its 654. Early Technology. Tarrytown, NY: Sleepy Hollow Grelen, H.E. 1980. Letter dated January 14, 1980, Restorations. in response to my inquiry about carrying capacity Hopkins, W. 1947a. Perhaps the hog is hungry. of southern forest range for feral hogs. USDA Forest Service, Southern Forest Experi- Hale, P.M. 1883. The Woods and Timbers of Eastern ment Station, Southern Forestry Notes No. 50. North Carolina. New York: E.J. Hale & Son. Hopkins, W. 1947b. Pigs in the pines. Forest Farmer Hall, B. 1829. Travels in North America in the Years 1827 7:3,8. and 1828, Vol. III. Edinburgh, UK: Cadell and Hopkins, W. 1947c. Hogs or logs? South Lumberman Co. 175:151–153. Hamilton, P.J. 1910 [1976]. Colonial Mobile. Univer- Jenings, E. 1704 [1923]. Letter to Her Majesties sity: University of Alabama Press. Lords Commissioners for Trade & Plantations. Hammond. 1980. United States Atlas. Maplewood, William Mary Quart 3:209–210. NJ. Kelly, J.F., and Bechtold, W.A. 1990. The longleaf Hann, W.J. 2002. Mapping fire regime condition pine resource. In Proceedings of Symposium on the class: A method for watershed and project scale Management of Longleaf Pine, Long Beach, MS 1989, analysis. In Proceedings of the 22nd Tall Timbers Fire pp. 11–22. USDA Forest Service Gen. Tech. Rep. Ecology Conference: Fire in Temperate, Boreal and Mon- SO-75. Southern Forest Experiment Station, New tane Ecosystems. Tallahassee, FL: Tall Timbers Re- Orleans, LA. search Station. Komarek, E.V. 1964. The natural history of light- Harcombe, P.A., Glitzenstein, J.S., Knox, R.G., ning. Proc Tall Timbers Fire Ecol Conf 3:139– Orzell, S.L., and Bridges, E.L. 1995. Western 183. Gulf coastal plain communities. In The Longleaf Komarek, E.V. 1965. Fire ecology—Grasslands and Pine Ecosystem: Ecology, Restoration and Management, man. Proc Tall Timbers Fire Ecol Conf 4:169– ed. S.M. Hermann. Proceedings Tall Timbers Fire 220. Ecology Conference No. 18. Komarek, E.V. 1966. The meteorological basis for Harper, R.M. 1905. Some noteworthy stations for fire ecology. Proc Tall Timbers Fire Ecol Conf 5:85– Pinus palustris. Torreya 5:55–60. 125. Harper, R.M. 1906. A phytogeographical sketch of Komarek, E.V. 1967. The nature of lightning fires. the Altamaha Grit Region of the Coastal Plain of Proc Tall Timbers Fire Ecol Conf 6:5–41. Georgia. Ann NY Acad Sci 17:1–415. Komarek, E.V. 1968. Lightning and lightning fires Harper, R.M. 1911. The relation of climax vegeta- as ecological forces. Proc Tall Timbers Fire Ecol Conf tion to islands and peninsulas. Bull Torrey Bot Club 8:169–197. 38:515–525. Komarek, E.V. 1972. Ancient fires. Proc Tall Timbers Harper, R.M. 1913. Economic botany of Alabama. Fire Ecol Conf 12:219–240. Part 1. University, Alabama: Geological Survey of Komarek, E.V. 1974. Effects of fire in temper- Alabama, Monograph 8. ate forests and related ecosystems: Southeast- Harper, R.M. 1914. Geography and vegetation of ern United States. In Fire and Ecosystems, eds. T.T. northern Florida. Sixth Annual Report, pp. 163– Kozlowski and C. Ahlgren, pp. 251–277. New 437, Florida Geological Survey. York: Academic Press. Harper, R.M. 1923. Some recent extensions of the La Rochefoucauld. 1799. Voyages dans les Etats-Unis. known range of Pinus palustris. Torreya 23:49–51. Vol. IV. Paris. Harper, R.M. 1928. Economic botany of Alabama. Lawson, J. 1709 [1967]. A New Voyage to Carolina. Part 2. University, Alabama: Geological Survey of Reprint. Chapel Hill: University of North Carolina Alabama, Monograph 9. Press. 2. History and Future of the Longleaf Pine Ecosystem 41

Lipscomb, D.J. 1989. Impacts of feral hogs on lon- at Sinnickson, VA, 1 mile south of the Maryland gleaf pine regeneration. South J Appl For 13:177– line. 181. Napier, J.H., III. 1985. Lower Pearl River’s Pineywoods. Little, E.L. 1971. Atlas of United States Trees. Vol. 1. Center for the Study of Southern Culture. Univer- Conifers and Important Hardwoods. Miscellaneous sity: University of Mississippi. Publication 1146. Washington, DC: USDA Forest Outcalt, K.W., and Sheffield, R.M. 1996. The lon- Service. gleaf pine forest: Trends and current conditions. Lockett, S.H. 1870. Louisiana as it is. Baton Rouge: USDA Forest Service, Southern Research Station, Louisiana State University Press. Resource Bulletin SRS-9. Martin, C. 1973. Fire and forest structure in the abo- Parrott, R.T. 1967. A study of wiregrass (Aristida riginal eastern forest. The Indian Historian 6:38–42, stricta Michx.) with particular reference to fire. 54. Master’s thesis, Duke University, Durham, NC. Mattoon, W.R. 1922. Longleaf pine. USDA Bul- Peet, R.K. 1993. A taxonomic study of Aristida stricta letin No. 1061. Washington, DC: US Government and A. beyrichiana. Rhodora 95:25–37. Printing Office. Peet, R.K., and Allard, D.J. 1995. Longleaf pine McWilliams, W.H. 1987. Developing timber-supply vegetation of the southern Atlantic and eastern projections for the South’s fourth forest. In Pro- Gulf Coast regions: A preliminary classification. ceedings: Current Challenges to Traditional Wood Pro- In The Longleaf Pine Ecosystem: Ecology, Restoration curement Practices. Atlanta, GA, September 20–30, and Management, ed. S.M. Hermann. Proceedings pp. 16–21. Madison, WI: Forest Products Re- Tall Timbers Fire Ecology Conference 18. search Society. Pierce, A.M. 1953. Tobacco Coast. A Maritime history of Merrens, H.R. 1964. Colonial North Carolina in the Chesapeake Bay in the Colonial Era. Newport News, Eighteenth Century. Chapel Hill: University of VA: The Mariners’ Museum. North Carolina Press. Powell, W.S.1977. John Pory, 1572–1636. Chapel Hill: Michaux, F.A. 1805 [1966]. Travels to the west of University of North Carolina Press. the Alleghany Mountains. In Early Western Travels, Reed, F.W. 1905. A Working Plan for Forest Lands Vol. III, ed. R. Thwaites. Reprint. New York: AMS in Central Alabama. Washington, DC: US Govern- Press. ment Printing Office. Michaux, F.A. 1871. North American Silva. Ruffin, E. 1843, 1844. Report of Survey of South Philadelphia: W. Rutter & Co. Carolina. Charleston, SC: Southern Agriculturist. Michel, F.L. 1702 [1916]. Report of the journey of Ruffin, E. 1861. Sketches of Lower North Carolina. Francis Louis Michel from Berne, Switzerland, Raleigh: North Carolina State Printer. to Virginia, October 2, 1701–December 1, 1702. Sargent, C.S. 1884. Report on the Forests of Wm. J. Hinkle, trans. Virginia Mag Hist Biog 24:1– North America. Washington, DC: USDI Census 43. Office. Mills, R. 1826. Statistics of South Carolina, Including a Schoepf, J.D. 1788 [1911]. Travels in the Confedera- View of its Natural, Civil and Military History. Charles, tion. Philadelphia: William J. Campbell, Pub. SC: Hurlbut and Lloyd. Scott, A.C. 1989. Observations on the nature of fu- Mohr, C. 1884. Quoted in Report on the forests sain. Int J Coal Geol 12: 443–475. of North America, ed. C.S. Sargent. Washington, Scott, A.C., and Jones, T.J. 1994. The nature DC: USDI Census Office. and influence of fires in Carbonaceous ecosys- Mohr, C. 1893. Turpentine orcharding in America. tems. Palaeogeogr Palaeoclimatol Palaeoecol 106:91– In Report of the Chief of the Division of Forestry for 112. 1892, ed. B.E. Fernow, pp. 342–346. Washington, Sherrard, T.H. 1903. A Working Plan for Forest Lands DC: US Government Printing Office. in Hampton and Beaufort Counties, South Carolina. Mohr, C. 1896. The timber pines of the southern Washington, DC: US Government Printing Office. United States. USDA Division of Forestry, Bulletin Smith, J., Capt. 1624. The general historie of Virginia, No. 13. Washington, DC. New England and the Summer Isles. Reprint, n.d. Mohr, C. 1901. Plant life of Alabama. USDA Division (circa 1970). Murfreesboro, NC: Johnson Publish- of Botany, Contrib. U.S. National Herbarium, Vol. ing Co. VI. Washington, DC. State Board of Agriculture. 1883. South Carolina: Re- Moldenke, H. 1979 (letter and personal communi- sources and Population, Institutions and Industries. cation). Observations and field notes from 1925, Charleston, SC: Walker, Evans and Cogswell, with longleaf pine herbarium specimen collected Publishers. 42 I. Introduction

Stowe, J.P., Jr., Varner, J.M., III, and McGuire, census: 1890. Washington, DC: US Government J.P. 2002. Montane longleaf pinelands. Tipularia Printing Office. 17:9–14. U.S. Census Office. 1902. Twelfth census of the U.S. Strachey, W. 1610 [1964]. A true repertory of the Vol. 5. Agriculture. Part 1. Washington, DC: US wracke, and redemption of Sir Thomas Gates Government Printing Office. Knight; upon, and from the Ilands of the Bermu- Wahlenburg, W.G. 1946. Longleaf Pine. Washington, das: his comming to Virginia, and the estate of that DC: Charles Lathrop Pack Forestry Foundation. Colonie then, and after, under the government of Wakeley, P.C. 1935. Artificial reforestation in the the Lord La Warre, July 15, 1610. In A Voyage to southern pine region. US Department of Agri- Virginia, ed. L.B. Wright. Charlottesville: Univer- culture Technical Bulletin No. 492. Washington, sity Press of Virginia. DC. Sudworth, G.B. 1913. Forest Atlas: Geographic Dis- Wakeley, P.C. 1954. Planting the southern pines. tribution of North American Trees, Part 1. Pines. Agriculture Monograph No. 18. Washington, DC: Washington, DC: USDA Forest Service. USDA Forest Service. U.S. Census Bureau. 1982, 1984, 1997. Census of Walker, J. 1995. Regional patterns in the vascu- agriculture. Vol. 1, Geographic area series. Part 1, lar flora and rare plant component of longleaf Alabama; Part 9, Florida; Part 10, Georgia; Part pine communities. In The Longleaf Pine Ecosys- 18, Louisiana; Part 24, Mississippi; Part 33, North tem: Ecology, Restoration and Management, ed. S.M. Carolina; Part 40, South Carolina; Part 43, Texas; Hermann. Proceedings Tall Timbers Fire Ecology Part 46, Virginia. Washington, DC: US Govern- Conference No. 18. ment Printing Office. Ware, S. 1978. Vegetational role of beech in U.S. Census Office 1841. Compendium of the sixth the Southern Mixed Hardwood Forest and the census of the United States. U.S. Census of Man- Virginia Coastal Plain. Va J Sci 29:231–235. ufactures. Washington, DC: Department of State. Ware, S., Frost, C.C., and Doerr, P.D.1993. Southern U.S. Census Office. 1853. The seventh census of mixed hardwood forest: The former longleaf pine the United States: 1850. Washington, DC: Robert forest. In Biodiversity of the Southern United States: Armstrong, Public Printer. Lowland Terrestrial Communities, eds. W.H. Martin, U.S. Census Office. 1864. Agriculture of the United S.G. Boyce, and A.C. Echternacht, pp. 447–493. States in 1860 compiled from the original returns New York: John Wiley & Sons. of the eighth census. Washington, DC: US Gov- Wertenberger, T.J. 1922. Planters of Colonial Virginia. ernment Printing Office. Princeton, NJ: Princeton University Press. U.S. Census Office. 1872. A compendium of the Wertenberger, T.J. 1931. Norfolk: Historic Southern ninth census (June 1, 1870). Washington, DC: US port. Durham, NC: Duke University Press. Government Printing Office. Williams, M. 1980. Products of the forest: Mapping U.S. Census Office. 1883. Report on the produc- the census of 1840. J For Hist 24:4–23. tions of agriculture as returned at the tenth census Williams, M. 1989. Americans and Their Forests: A His- (June 1, 1880). Washington, DC: US Government torical Geography. New York: Cambridge Univer- Printing Office. sity Press. U.S. Census Office. 1895. Report on the statistics of Young, H. 1781 [1881]. Letter to Col. Davies. Cal- agriculture in the United States at the eleventh endar of state papers [Virginia] 2:619. 2. History and Future of the Longleaf Pine Ecosystem 43

BOX 2.1 resins having been dated to the Carbonif- erous Era, over 200 million years before The Naval Stores Industry present (Langenheim 1969). Terpenes are synthesized through the basic malevonic– Alan W. Hodges pyruvate biochemical pathway, and occur in Institute of Food and Agricultural Sciences, many different groups of plants, as well as all University of Florida, Gainesville, conifers. Longleaf pine, slash pine (P. elliot- Florida 32611 tii), and other subtropical pine species are especially rich in terpenes because of high The vast forests of longleaf pine (Pinus palus- year-round pest pressures. tris) in the southeastern United States were Terpenes are produced in pine trees by once the basis of a very large industry specialized cells that form a network of mi- for producing pine tar, rosin, turpentine, croscopic ducts with interconnected longi- pine oil, and other products derived from tudinal and radial segments, as illustrated the natural oleoresin of the tree. The term in Box Fig. 1. The epithelial cells of the “naval stores” came about from the use of resin ducts arise from the parenchyma tis- these products for building and maintain- sue, and lack the rigid cell wall of normal ing wooden ships during the seventeenth wood fibers. As terpenes are secreted into through nineteenth centuries. Pine tar was the lumen of the resin ducts, where they used together with various fibrous materi- are stored, the elastic membrane of the ep- als to seal the seams between the wooden ithelial cells maintains a relatively high pres- planking. It was an important strategic com- sure (300 psi) on the fluid oleoresin such modity for the naval-based European em- that it can be mobilized in case of an in- pires of Britain, Holland, Spain, and Por- jury. Oleoresin is present in all parts of the tugal. Although the term “naval stores” is tree—leaves, branches, stem, roots, bark— now antiquated, it is still frequently used in and typically represents about 3 to 5% of to- historical contexts. However, the preferred tal tree biomass (dryweight basis). In older contemporary terminology in the indus- trees, oleoresin accumulates in the stumps try is “pine chemicals” or “gum and wood and heartwood, and in the wood around the chemicals.” base of major branches. Oleoresin is not to Naval stores production is based upon ex- be confused with sap, the nutrient solution ploiting the terpene chemical defense sys- that is carried in a separate system of vascu- tem of the pine tree, which protects against lar tissues. wood-decaying fungi and insect pests such Methods for harvesting or extraction of as bark beetles (Dendroctonus, Ips sp.). When oleoresin from pines have evolved sig- a tree suffers injury to the bark and cam- nificantly over the past 400 years, due bium layer, oleoresin is secreted to prevent to changes in technology and the forest the establishment and spread of pathogens, resource. Beginning in the early 1600s, acting as a natural biocide and preserva- colonists in North America made pine tar tive. Pine oleoresin is a complex mixture of for export to Europe by a pyrolysis pro- about 30 to 50 different terpene molecules, cess, i.e., by slowly burning resinous wood comprised chiefly of diterpene resin acids in earthen kilns. This activity reached its and monoterpene essential oils, which im- peak in North Carolina, where tar makers, part different physical and biotic proper- known as “tarheels,” exploited the abun- ties (Zinkel and Russell 1989). The terpene dant longleaf pine forests (Butler 1998). chemical defense system came about very They gathered naturally occurring resinous early in the evolution of higher plants on wood from old-growth stumps, heartwood, earth, with some ambers from fossilized tree and branch knots, and also deliberately 44 I. Introduction

FIGURE 1. Anatomy of the oleoresin duct system of longleaf pine. Courtesy of USDA Forest Service. Illustration by Susan Trammell, Willisiton, FL.

injured trees to cause resinosis by remov- or weakened the trees to other mortality ing bark and applying ﬁre. Resin-saturated factors. The exploitative resin harvesting pine is known as “lightwood,” because the was usually followed by clear-cut logging, settlers used the wood for torches. resulting in widespread deforestation. The During the 1800s and early 1900s, as set- laborious process of hacking and dipping tlers moved into the southeast U.S. coastal was done mostly by black workers, many plain region, the gum naval stores indus- of whom were slaves, descendants of slave try developed for tapping of living longleaf families, or prisoners, who lived in isolated and slash pine trees (Box Fig. 2). Trees camps. Most gum naval stores operators were repeatedly wounded using a hook- practiced annual controlled burning of the shaped cutting tool known as a “hack” to forest stands to improve accessibility, and cause the natural defensive response and carefully prepared the stand by raking away bring about oleoresin exudation. The ole- litterfall around each tree to avoid scorching oresin was collected in cavities (“boxes”) the tapped face. chopped into the base of the tree, and In the twentieth century, better meth- special “dipping” tools were used to pe- ods were developed for collecting gum riodically remove the accumulated oleo- oleoresin in manufactured containers at- resin. This destructive practice often killed tached to the tree. Clay cups introduced by 2. History and Future of the Longleaf Pine Ecosystem 45

FIGURE 2. Map of the naval stores belt, and location of processing plants (1934). Courtesy of USDA Forest Service. Southern Forest Experiment Station, forest survey (1934).

Dr. Charles Herty in 1906 led to a resur- practiced (Gerry 1935). With appropriate gence of the industry. Typically, the cups conservative practices, trees may be tapped were hung from a nail underneath a pair for about 12 years on two sides or “faces,” of sheet metal gutters that channeled the and can then normally be used for the full oleoresin into the cup (Box Fig. 3). Many range of wood products (Box Fig. 4). of these old clay cups and tins can still be Chemical treatments are also used for found in the woods today. The standard increasing oleoresin yields from tapping contemporary method of oleoresin tapping, pines. Sulfuric acid is applied as a spray known as bark chipping, was developed by or paste solution to the freshly exposed U.S. Forest Service researchers during the cambium to destroy cells surrounding the 1940s. Strips of bark and cambium approx- opening of severed resin ducts, preventing imately 2 inches (5 cm) wide are removed premature occlusion of the ducts, and pro- across one-third of the tree’s circumference, longing oleoresin ﬂow for several weeks. at intervals of 3 to 4 weeks throughout the A new generation of chemical stimulation March through October period in the south- has been developed using plant regulators eastern United States. This method causes such as ethylene, which acts as a general signiﬁcantly less damage to trees than the stressor, stimulating biosynthesis of oleo- previous practice of deeply chipping into resin. With the best available method, the the wood. It was also discovered that chip- expected annual oleoresin yield from a 35- ping stimulates a roughly a sevenfold in- year-old longleaf pine, 10 inches (25 cm) crease in the number of resin ducts in new DBH, is approximately 11.2 pounds (5.1 kg) wood formed above the chipped face, which (McReynolds and Kossuth 1984). For an enables higher yields if light chipping is 8-year production period, in a typical stand 46 I. Introduction

the 1930s, the primitive fire stills were replaced by about 30 large central steam processing plants. Rosin is an amber-colored crystalline solid material at room temperature, and is used for making adhesives, sealants, coatings, fluxes, printing inks, emulsifiers, and food products such as chewing gum. Turpentine is used in solvents, cleaners, antiseptics, in- secticides, flavors and fragrances, and syn- thetic resins. The rosin fraction typically represents about 70% of the original crude oleoresin, turpentine about 15%, and foreign material such as dirt, litter, and water the remaining 15%. Rosin is traded commercially based upon color and chemical composition, which is determined by the pine species, and methods of production and processing. “American” rosin from longleaf and slash pines makes an excellent grade that is recognized worldwide as the standard of quality. A characteristic of the rosin from longleaf pine is that it crystal- lizes very rapidly, due to the particular mix

FIGURE 3. The Herty cup and gutter system for of diterpene resin acids. A significant por- resin collection, circa 1906. Courtesy of USDA tion of the oleoresin becomes crystallized Forest Service. Image scanned from A Pictoral and dried and on the face, and it is neces- Album of the Naval Stores Industry (1937), Univer- sary to remove this material with a special sity of Florida library. scraping tool. At its peak in 1910, the United States of 50 tappable trees per acre, the revenue produced nearly 600,000 metric tonnes of from a gum naval stores operation would rosin and turpentine, which is the high- be over $1500 per acre. Oleoresin yields est of any country in history, and is un- are strongly influenced by the age, size, and likely to ever be achieved again. At this vigor of trees, and stand density and canopy time, there were about 27,000 workers em- development. ployed in the industry, with over 10,000 The crude oleoresin extracted from pines independent producers operating on about is distilled to separate the principal con- 8 million acres. Since the 1930s, the gum stituents of diterpene resins and monoter- naval stores industry in the United States pene essential oils, known commercially has steadily declined due to the high cost of as rosin and turpentine. Originally, this labor, scarcity of suitable timber, and com- was done at small farm-scale distilleries us- petition from foreign producers and substi- ing fire-heated copper kettles adapted from tute materials. The last remaining process- Scottish whiskey stills, and at one time there ing plant in the United States closed in 2001. were over 1800 such stills in the southeast- However, gum naval stores is still an im- ern United States (Box Fig. 5). A few of these portant industry in many developing coun- old stills have been preserved for public tries, such as China, Indonesia, India, Brazil, demonstrations of the process. Beginning in and Mexico, with global production of gum 2. History and Future of the Longleaf Pine Ecosystem 47

FIGURE 4. Typical woods scene of gum naval stores harvesting operation, circa 1900. Courtesy of USDA Forest Service. State archives of Florida, call number Rc 02612.

FIGURE 5. Turpentine still building and loading ramp. State archives of Florida, call number PR 12612. 48 I. Introduction

rosin and turpentine in excess of 700,000 recovered from pulp digestors, while non- metric tons annually (Hodges 2002). About volatiles extracted from wood pulp are re- 10 major pine species are used around the covered from the black liquor stream as world, including Pinus elliottii established crude tall oil, then fractionated into rosin, from seedstock in the United States. fatty acids, and a variety of other com- Wood naval stores is another facet of the pounds. Both the turpentine and rosin industry based upon extraction of rosin and are contaminated by sulfates, but indus- oils from the resin-saturated stumps of first- trial users have adapted these materials to growth longleaf pine trees that died or were their needs at lower cost than other sources, harvested decades ago. This process was de- and they compete with petroleum hydro- veloped in the 1920s, and continues today at carbon chemical feedstocks. There is over a single plant in Brunswick, GA, operated by 800,000 metric tons of crude tall oil pro- Hercules Inc. Weathered stumpwood con- duced in the United States and Canada tains about 40% extractives, by weight, (Hodges 2002). However, longleaf pine does and is highly decay-resistant. Stumps are not contribute substantially to this indus- recovered from cut-over forest lands with try because it is seldom used for making heavy equipment, then transported to the paper. plant by truck or railcar, where they are finely shredded before extraction with solvents such as gasoline or hexane. The pro- References cess produces an excellent grade of pale Blount, R. 1995. Spirits of Turpentine: A History of rosin, wood turpentine, and pine oil. Lon- Florida Naval Stores, 1528–1950. Florida Heritage gleaf pine stumpwood is essentially a non- Journal Monograph No. 3. Tallahassee: Florida renewable resource, since trees are gener- Department of Agriculture and Consumer Ser- ally not allowed to grow to the age when vices. heartwood formation occurs, after about 80 Butler, C.B. 1998. Treasures of the Longleaf Pine: years. Nevertheless, the remaining supply Naval Stores. Shalimar, FL: Tarkel Press. of stumps is expected to last many more Gerry, E. 1935. A Naval Stores Handbook Dealing with the Production of Pine Gum or Oleoresin. years at the current rate of recovery, to pro- Miscellaneous Publication 209, US Forest Ser- duce about 20,000 metric tons of wood rosin vice, Washington, DC. annually. A system for artificially inducing Hodges, A., ed. 2002. Forest Chemicals Review In- resinosis in young southern pines by treat- ternational Yearbook, 2001. New Orleans, LA: ment with herbicides such as paraquat was Kreidt Publications. developed during the 1970s (Stubbs et al. Langenheim, J. 1969. Amber: A botanical in- 1984). This process achieved a 15-fold inquiry. Science 163:1157–1169. crease in total whole-tree extractives over a McReynolds, R.D., and Kossuth, S.V. 1985. period of 2 years, under the optimal treat- CEPA in liquid sulfuric acid increases oleoresin ment. However, it was not commercialized yields. South J Appl For 9(3):170–173. because of difficulties with high tree mor- Stubbs, J., Roberts, D., and Outcalt, K. 1984. Chemical stimulation of lightwood in southern tality and insect pests. pines. General Technical Report SE-25. USDA The most important source of naval stores Forest Service, Southeast Forest Experiment products in the United States today is a Station, Asheville, NC. by-product of the sulfate or “Kraft” pulping Zinkel, D. and Russell, J., eds. 1989. Naval Stores: process (Zinkel and Russell 1989). Turpen- Production, Chemistry, Utilization. Atlanta, GA: tine volatilized by cooking of wood chips is Pulp (Pine) Chemicals Association. Section II Ecology

49 Chapter 3 Ecological Classiﬁcation of Longleaf Pine Woodlands

Robert K. Peet

Introduction bilayered communities with the physiognomy maintained by frequent, low-intensity surface When Europeans first settled in southeastern fires that removed most small woody plants North America and began to explore their and thereby kept the canopy open. However, new homeland, they found a landscape that this caricature obscures the remarkable floris- was to a large extent dominated by open, sa- tic diversity of these systems. At a within- vannalike longleaf pine woodlands. The pines community scale, longleaf vegetation can be were typically widely spaced, affording the among the most diverse in North America with traveler opportunities to see for long dis- some examples having 40 or more species of tances without obstruction by undergrowth. higher plants per square meter (Walker and The ground layer was dominated by grasses Peet 1983) or 170 per 1000 m2 (Peet, Carr and with a great diversity of showy forbs. Vege- Gramling 2006; W. J. Platt personal commu- tation of this character occurred from south- nication). But even more impressive is the di- eastern Virginia southward deep into penin- versity reflected in the change in composition sular Florida and west to western Louisiana of longleaf vegetation with subtle changes in and eastern Texas (Frost et al. 1986; Harcombe environmental conditions, or with geographic et al. 1993; Peet and Allard 1993; Ware et al. distance. This diversity is particularly conspic- 1993; Platt 1999; Christensen 2000; Frost this uous in the floristic richness and endemism of volume). the region. There are on the order of 6000 vas- Early descriptions and superficial treatment cular plant taxa that occur on the southeast- in textbooks have created and maintained ern Coastal Plain, which represents almost a the inaccurate perception, widespread out- quarter of all plant species that occur in North side the Southeast, that the vegetation of America north of Mexico. Moreover, 1630 taxa the original longleaf pine ecosystem was a are endemic to the Coastal Plain, and with homogeneous and monotonous bilayer com- 1306 full species included (Sorrie and Weakley munity with pines above and grass below. 2001, 2006). The region falls just short of While this is clearly an oversimplification, the qualifying as one of the top 25 biodiversity original longleaf ecosystems were generally hotspots on the globe (see Myers et al. 2000).

r Robert K. Peet Department of Biology, University of North Carolina, Chapel Hill, North Carolina 27599-3280.

51 52 II. Ecology

A large proportion of the endemics occur in this volume). For many types of sites, noth- the longleaf-dominated vegetation (Sorrie and ing is left that might be used as a template Weakley 2005). for restoration efforts. Many sites where long- Although longleaf ecosystems once domi- leaf does persist have been significantly al- nated the southeastern Coastal Plain, most of tered by fire suppression. Further, the great this vegetation was gone by 1920 and today diversity, endemism, and spatial heterogene- less than 3% of the original extent of longleaf ity of the longleaf system means that any vegetation remains in natural conditions (Frost reserve system that aspires to preserve bio- 1993; Ware et al. 1993; Frost this volume). To- diversity needs to incorporate the range of day, natural longleaf stands occur primarily on environmental conditions at numerous places lands characterized by soils too wet or too dry scattered over the original range of the for agriculture (Ashe 1897; Mohr 1901; Harper species. 1906; Frost 1993). From North Carolina to Vegetation description and classification Mississippi the inner coastal plain with its pre- play a key role in many areas of conservation, dominantly finer-textured soils has lost essen- land management, and scientific research. A tially all of its original, vast extent of longleaf primary role of vegetation classification is to vegetation through conversion to agriculture. delimit natural communities so as to provide Gone also are nearly all longleaf populations a framework for identifying, understanding, from Virginia and from North Carolina north managing, and restoring the natural vegeta- of the Neuse River, the region where longleaf tion. Managers of conservation lands require was first exploited for naval stores. What re- accurate and detailed descriptions of the veg- mains of the original longleaf ecosystem is a etation attributes they need to preserve or small and biased sample of what was once one re-create. Without a well-formulated vege- of the most extensive and diverse biomes of tation classification and description, a qual- North America. ity template for management or restoration Simple removal of longleaf is far from the is often impossible. In short, future conserva- whole story of the demise of the longleaf tion and restoration must be based on knowl- ecosystem. Just as important has been the loss edge of vegetation composition across a broad of the original fire regime. On all but the most range of sites selected to represent that range sterile sites, a significant decrease in fire fre- of natural conditions and geographic varia- quency quickly leads to a dense growth of tion. Unfortunately, the diversity of longleaf woody plants, followed by a competitive fail- vegetation has been relatively little studied. ure of the original ground cover and thus most Documentation of compositional variation can of the original biodiversity of the longleaf sys- be found in the scientific literature for small tem (see Frost 2000). Although conservation- portions of this system over limited ranges ists working in other ecological systems often of soil conditions. Vast areas of the longleaf identify critical tracts for preservation by the region have not been subjected to rigorous persistence of old-growth trees, this approach ecological study. Although for some regions is generally inappropriate for longleaf systems. it is now too late, efforts to document the For longleaf ecosystems where the biodiversity compositional variation of the remnants of is concentrated in the ground layer, tree age this remarkable ecosystem are underway. My is relatively unimportant compared to the in- goal in this chapter is to combine informa- tegrity of the ground-layer vegetation, which tion in the nascent U.S. National Vegetation in turn depends on the long-term persistence Classification (NatureServe 2005; see Ander- of a regime of frequent, low-intensity surface son et al. 1998) with available quantitative fires. data to create a preliminary classification of Conservation and restoration of the nat- natural, fire-maintained, longleaf-dominated ural longleaf ecosystem is remarkably diffi- vegetation types to guide future conservation cult for several reasons (Walker and Silletti and restoration efforts. 3. Ecological Classification of Longleaf Pine Woodlands 53 Physiography and Ecoregion is derived from marine sediments of Miocene age or younger. This region tends to Ecoregions of the Longleaf be extremely flat with significant topography Ecosystem restricted to banks and bluffs of major rivers. Local topographic relief is consistently under The relatively gentle to imperceptible topog- 20 meters. These Atlantic coastal flatlands are raphy that characterizes much of the range perhaps best visualized as a series of old bar- of longleaf belies the considerable geographic rier dunes and shorelines. Soils of the barrier variation and complexity in its environment dunes and shorelines per se tend to be ex- and biota. One approach to understanding tremely sandy and dry due to rapid percola- and managing this subtle complexity is to tion of water, whereas the soils of the once break the longleaf system into ecoregions that embayed regions tend to be seasonally satu- are relatively consistent in their climate, soils, rated as a result of their low relief and finer soil and physiography. Comparison of vegetation texture, which make for poor drainage (DuBar . within and between ecoregions provides a et al 1974; Daniels et al. 1984; Soller and Mills framework within which variation in the long- 1991). Old marine terraces are also promi- leaf ecosystem can be understood. nent and tend to be flat with fine-textured, Several alternative ecoregion systems have poorly drained soils. In almost all cases, the been published (e.g., Omernik 1987; Bailey soils tend to be highly phosphorus deficient. 1995; Brown et al. 1998; Ricketts et al. 1999). Coarse, siliceous sands have formed dune sys- However, the system that seems to match tems on the northeast sides of all major rivers best the natural variation of longleaf vegeta- draining into the Atlantic as well as on the tion is that of the U.S. Environmental Protec- northeast sides of Carolina bay depression wet- tion Agency (EPA; see Omernik 1987, and uplands, in both cases the sands having been dates). EPA ecoregions are hierarchical with blown out of the river valleys and bay depres- 4 levels. Although I have found level-4 ecore- sions prior to the last glacial advance. gions to correlate well with compositional vari- The inland half of the Atlantic Coastal Plain ation in longleaf ecosystems, there are far too Ecoregion is dominated by landforms that are many types to provide a useful context for generally older and the topography more hilly examining large-scale, range-wide patterns. I and complex than on the outer coastal plain. here describe six ecoregions largely based on The overall aspect is of low, rolling hills, of- the nine EPA level-3 ecoregions that span the ten with loamy rather than sandy soils. A natural range of longleaf pine, plus I treat sepa- distinctive region of clay hills occurs in South rately one level-4 segregate, the Fall-line Sand- Carolina (Myers et al. 1986). The pronounced hills of the Carolinas and Georgia (Fig. 1). topographic relief of the rolling inner coastal plain results in generally well drained soils with the consequence that seasonally wet sites The Atlantic Coastal Plain are less common than on the outer Coastal Plain, mostly of local occurrence, and associ- Ecoregion ated with outcrops of impermeable soil hori- The Atlantic Coastal Plain Ecoregion includes zons. the coastal flatlands of the EPA Middle At- lantic Coastal Plain region and that portion The Fall-line Sandhills of the EPA Southeastern Plains region occur- Ecoregion ring from Virginia southwest to and including the Altamaha Grit region of Georgia, stop- The Fall-line Sandhills Ecoregion is identical ping at the Flint River as an arbitrary division to the same-named level-4 segregate of the between the Atlantic and Gulf Coastal Plains. EPA Southeastern Plains ecoregion and forms The outer portion of the Atlantic Coastal Plain its inland fringe from central North Carolina 54 II. Ecology

FIGURE 1. Six longleaf pine vegetation ecoregions, largely derived from EPA Ecoregions (see text). to easternmost Alabama. This is a region of distinct from the rest of Southeastern Plains, primarily coarse, Cretaceous-age sediments, which tend to be dominated by fine-textured often with a veneer of Miocene-age eolian soils rather than coarse sands. dune sands. The sandhill sediments appear to have eroded from the ancient uplands that to- The Southern Coastal Plain day form the Piedmont. Intermixed with the fall-line sands are layers of clay that impede Ecoregion drainage and result in seepage wetlands where The Southern Coastal Plain Ecoregion includes the clay layers appear near or at the surface. most of the EPA Southern Coastal Plain Ecore- The sandhills are also characterized by scat- gion including the range of longleaf in Florida tered subtle depressions, perhaps similar in ori- south of the Cody Scarp, the lower coastal plain gin to the Carolina bay depression wetlands of Georgia, and a narrow fringe along the out- of the flatlands south and east of the sandhills ermost coastal plain of southern South Car- (James 2000). These sandhill depressions often olina. Soils of this ecoregion, like those of the contain finer-textured soils than the adjacent Atlantic Coastal Plain, are derived almost ex- sandy landscape, perhaps deposited by winds. clusively from marine sediments of Miocene The area is both ecologically and edaphically age or younger, and the region tends to be 3. Ecological Classification of Longleaf Pine Woodlands 55 extremely flat with significant topography re- River and is included in the EPA Western stricted to banks and bluffs of major rivers. Gulf Coastal Plain and South Central Plains The Southern Coastal Plain differs from the ecoregions. The more southerly portion of this Atlantic and Gulf Coastal Plains in the gen- ecoregion is similar to its southeastern coun- eral absence of fine-textured soils, nearly the terparts in being strikingly flat with extensive whole region being characterized by a veneer areas of poorly drained soils. The more in- of marine sands. In addition, the influence of land portions are characterized by rolling hills the underlying limestone of the Florida penin- and even occasional outcrops of the underly- sula not infrequently is expressed in the form ing sandstone. The Pleistocene terraces of this of higher soil phosphorus content than gener- region are dominated by silty and even clayey ally encountered elsewhere in the sandy soils soils, some exhibiting strong vertic tendencies. of the longleaf region (Peet, Carr and Gramling 2006). The Piedmont and Montane Uplands Ecoregion The Eastern Gulf Coastal Plain The Piedmont and Montane Uplands Ecore- Ecoregion gion includes the eastern fringe of the igneous The Eastern Gulf Coastal Plain Ecoregion in- EPA Piedmont ecoregion in the Carolinas as cludes the EPA Southeastern Plains ecoregion well as the Piedmont and adjacent EPA South- from the Flint River in Georgia southward to western Appalachians and Blue Ridge eco- the Cody Scarp including the Tallahassee Red regions of northwestern Georgia and adjacent Hills of north Florida, and west to include the eastern Alabama. This ecoregion is the most EPA Loess Plains ecoregion of Louisiana and atypical longleaf ecoregion in both topography Mississippi. Also included is the narrow fringe and substrate. Here are found mature land- of EPA Southern Coastal Plain ecoregion west scapes with well-developed drainage networks of the Florida–Alabama border. As with the in- and complex topography which, except for ner Atlantic Coastal Plain, landforms here are the Ridge and Valley region, are character- generally older and the topography more hilly ized by kaolinitic clay soils largely derived from and complex than on most of the Coastal Plain. igneous rocks. Generally, longleaf sites of this The overall aspect is of low, rolling hills, often region tend to be well drained, though seep- with loamy rather than sandy soils. A distinc- age areas with longleaf vegetation do occur tive region of clay hills occurs in Alabama and occasionally. Mississippi (Hodgkins 1965; Hodgkins et al. 1979), which extend into southern Georgia (Harper 1930). The soils are generally well A Representative Longleaf drained soils with wet sites primarily confined to the coastal fringe. The region near the Mis- Landscape sissippi River is recognized by EPA as a separate Over much of the range of longleaf pine, lo- ecoregion owing to its distinctive thick cap of cal variation in vegetation can be interpreted loess, blown out of the river during the late in terms of two primary gradients: soil mois- glacial period. The loess cap gives the region ture and soil texture. This typical pattern serves distinctive, fine-textured, highly fertile soils as a general model for factors that affect lon- that set it off from all other longleaf landscapes. gleaf pine vegetation (Fig. 2). Although variation in soil nutrients can also be important, The West Gulf Coastal Plain for a given moisture and texture regime the Ecoregion soils in a region are relatively predictable, as is the overall character of the associated vegeta- The West Gulf Coastal Plain Ecoregion in- tion. Only when one leaves the Coastal Plain cludes all longleaf lands west of the Mississippi for the Piedmont and Montane longleaf types 56 II. Ecology

Dry Ultisols – No natural vegetation remains; primarily agriculture

Well-drained Wet Ultisols – Ultisols – Savannas Silty uplands Percent Silt in Soil A Horizon Entisols – Spodosols – Entisols –

20 40 60Subxeric 80 100 Flatwoods Xeric sand sandy uplands barrens and sandy uplands 0 Super-Xeric Xeric Sub Xeric Mesic Wet Mesic Hydric Soil Moisture

FIGURE 2. A model landscape of Coastal Plain longleaf pine vegetation showing dominant vegetation types in relation to soil silt content and soil moisture.

does this general, two-gradient model break generally Ultisols and support what I refer to down, being replaced by a model where soil here as pine savanna vegetation types. clay and incident solar radiation are key vari- Species diversity varies in a consistent pat- ables and soil moisture is less significant. tern across the gradient model (Table 1). By Dominant soil orders vary in a manner all measures, the barrens of sandy, xeric sites consistent with the two-dimensional gradi- are least diverse. At the scale of a 1000 m2 ent model of variation in longleaf vegetation. plot, diversity increases with increasing soil silt Sandy, dry sites are dominated by Entisols. The content, but with no conspicuous difference most well-drained, coarse sands are extremely between wet and well-drained sites. Examples dry, intrinsically low in nutrients, and support of exceptionally species-rich plots with about open, sand-barren vegetation, whereas nearly 170 species per 1000 m2 have been reported all of the less extreme sites support a well- from well-drained soils of the Mississippi loess developed ground layer dominated by grasses. plains (W. Platt personal communication) and Soils of poorly drained, sandy sites are primar- from silty seeps in the Tallahassee Red Hills ily Spodosols, and on the wettest sites Histisols (Peet, Carr and Gramling 2006). However, at can occur. The Spodosols, which are highly scales between 10 m2 and 0.01 m2, the savanna infertile, particularly with respect to nitrogen vegetation types of wet, silty soils are the most and phosphorus, support a vegetation type diverse of the longleaf ecosystem. The average often called flatwoods. The portion of the gra- plot on moist, silty soils has on the order of dient diagram with well-drained, silty soils is 20 species per square meter, and where soils largely associated with Ultisols. Unfortunately, are particularly fine and fire occurs nearly an- most such sites have been converted to agri- nually, species richness can exceed 40 species culture. Poorly drained, silty soils are also per square meter. 3. Ecological Classification of Longleaf Pine Woodlands 57

TABLE 1. Species richness as a function of plot size for representative longleaf community types of southeast North Carolina and of Florida.a Area m2 0.01 0.1 1 10 100 1000 Xeric sand barrens & uplands North Carolina 0.5 1.3 3.2 6.6 12.9 22.5 Florida 1.1 3.2 9.5 20.9 42.6 74.1 Subxeric sandy uplands North Carolina 0.8 2.7 5.5 10.3 19.1 34.8 Florida 1.1 3.9 11.2 24.4 48.1 84.0 Silty & clayey uplands North Carolina 2.7 7.5 15.7 27.1 51.7 81.4 Florida 2.1 7.1 17.1 32.8 63.7 107.5 Flatwoods North Carolina 2.3 6.0 11.2 18.7 33.2 54.6 Florida 1.8 5.2 11.7 21.4 40.0 71.2 Savannas North Carolina 4.4 11.3 22.4 36.0 61.1 94.4 Florida 3.1 9.2 18.7 30.4 54.5 89.8 a (Values are based on 180 vegetation plots from southeastern North Carolina and 281 from Florida, each typically 1000 m2 in area and containing eight subplots each of 0.01, 0.1, 1, and 10 m2, and four subplots of 100 m2. Averages were calculated for association types, and the associations within a community type were averaged within a major community type—types were averaged rather than plots to avoid weighting types by numbers of plots.

Major patterns of variation in vegetation spp., and Rhynchospora spp. being nearly ubiq- composition also correspond well with the uitous (though Ctenium is absent from the general gradient model. All longleaf vegetation savannas of Texas, dropping out at the Sabine except that of extremely xeric barrens sites has River). a well-developed grass-sedge layer. The best- Oaks (e.g., Quercus laevis, Q. incana, Q. gem- known grass of the longleaf woodlands is wire- inata) frequently occur as co-dominants on grass (Aristida stricta and A. beyrichiana1). From xeric and subxeric sandy soils, with scrubby northern South Carolina north to roughly evergreen oak diversity (e.g., Q. chapmanii, Q. the Pamilico Sound, Aristida stricta domi- myrtifolia) increasing with proximity to penin- nates. On either side of the range of Aris- sular Florida. The oaks decline in importance tida stricta, there is no wiregrass, but instead with increasing soil moisture or silt content, Schizachyrium scoparium dominates. South of with two exceptions: live oak (Q. virginiana) the Santee River in South Carolina, wire- occupies somewhat more mesic sites than most grass again occurs, here in the form of Aris- of its congeners, and runner oaks (Q. minima, Q. tida beyrichiana. Wiregrass extends southward pumila) can be prominent on somewhat fine- through most of the Florida Peninsula and textured Ultisols and on moderately drained westward along the Gulf Coast into the south- Spodosols. Shrubs of the heath family are well eastern corner of Mississippi. Wiregrass is developed in flatwoods and in subxeric sand- generally the dominant grass type of Spo- hills types, but decrease significantly in im- dosols. Wiregrass also occurs on Ultisols, but portance with increasing silt content. In con- the bluestems (Andropogon spp., Schizachyrium trast, legumes are almost entirely confined to spp.) tend to be at least equally important. the fine-textured soils, where they can be ex- On wet, fine-textured soils numerous grasses ceptionally diverse and abundant (Gano 1917; and sedges share dominance with Ctenium aro- Wells and Shunk 1931; Taggart 1994; Hainds maticum, Muhlenbergia expansa, Dichanthelium et al. 1999; James 2000). The mechanism 58 II. Ecology behind the distribution of legumes remains Geographic Data Committee (FGDC), and with undocumented, though I have found experi- advice on standards from the Vegetation Panel mentally in both North Carolina and Missis- of the Ecological Society of America. Most of sippi that legume abundance on wet silty soils the current content of this classification is the can be significantly increased by application of result of efforts by NatureServe staff or the staff phosphorus. Plants of the lily and orchid fam- of state Natural Heritage Programs. Details can ilies are well known to be showy, diverse, and be examined at the NatureServe Explorer web- abundant on the wet longleaf sites over both site (http://www.natureserve.org/explorer/; Ultisols and Spodosols, but to drop off with de- see also Anderson et al. 1998). The associa- creasing soil moisture (Walker and Peet 1984; tions recognized in the National Classification Peet and Allard 1993). Savannas and wet flat- are derived from multiple sources and range woods are also the area of greatest concen- from types based on careful, quantitative tration of the region’s rich assemblage of in- analysis of multiple plots to types based only sectivorous plants. Palms are largely plants of on old literature descriptions or author field the flatwoods; from the southeastern corner notes. of South Carolina southward, saw palmetto I also compiled the results of a series of four, (Serenoa repens) can be an aspect dominant of plot-based studies conducted by my own re- the ground layer, and other palms appear in search group (Duncan et al. 1994; Peet et al. particular flatwoods types (e.g., Sabal palmetto, 1994; Kjellmark et al. 1998; Peet, Carr and Sabal etonia). Gramling 2006). These studies covered the range of variation my colleagues and I could identify in extant stands of longleaf and re- Toward a Community lated vegetation types of North Carolina, South Classification Carolina, Georgia, and Florida. The studies were based on approximately 900 inventory The primary purpose of this chapter is to plots of typically 1000 m2 each and collected present and describe a comprehensive classi- following the protocol of Peet et al. (1998). The fication of longleaf pine vegetation based on a plot data are available for further analysis in compilation and synthesis of prior work, both VegBank (http://vegbank.org). Plots were se- my own and that of others. Such a classifica- lected only from stands that appeared to be in tion can serve as a framework for understand- relatively natural condition with a sustained ing ecological variation, planning and assess- history of low-intensity fire. Sites severely fire ing restoration, and developing conservation suppressed, impacted by pine-straw ranking or strategies. I chose to define the focus of this military training, or otherwise degraded were study as fire-maintained longleaf pine vegeta- avoided. tion, plus related vegetation from within the For each of the four studies we used ag- natural range of longleaf where other pines glomerative cluster analysis methods to iden- dominate, or pines are typically present but too tify groups of relatively similar plots and sparse to generally be described as among the examined the results of each in an effort to dominants. identify groups that were both ecologically in- The most mature work on classification terpretable and consistent across multiple clas- of longleaf vegetation is represented by the sification methods. For our final cluster anal- longleaf vegetation types treated in the U.S. ysis we generally used the flexible beta group National Vegetation Classification. As a first linkage clustering method (McCune and Grace step I compiled all qualifying vegetation types 2003) with Sorenson’s metric to quantify eco- (associations) within the U.S. National Vegeta- logical distances among plots. In each case we tion Classification as of 2004. This classification compared our results with the types in the is maintained by NatureServe on behalf of then current version of the National Vegetation and in compliance with the standards of the Classification. This led to proposals for changes Vegetation Subcommittee of the U.S. Federal in the classification documented below. I also 3. Ecological Classification of Longleaf Pine Woodlands 59 reviewed available published articles and book the longleaf ecosystem. Vegetation on these chapters that have summarized longleaf veg- sites consists of scattered overstory longleaf etation for particular places or geographic re- with an oak understory and often little else. gions. These are described in the relevant lo- The sand is typically bright white and is con- cations later in this section. spicuous for lack of ground cover. Grasses are To create a single classification for this re- often sparse and the open ground layer nor- view, I compiled the recognized longleaf types mally includes low shrubs such as Gaylussa- into six ecological groups related to physiog- cia dumosa and often a discontinuous mat of nomy and soil type: (1) xeric sand barrens Selaginella and fruticose lichens (e.g., Cladonia and uplands, (2) subxeric sandy uplands, (3) and Cladina) (Fig. 3). silty uplands, (4) clayey and rocky uplands, (5) Pine barren vegetation types generally have flatwoods, and (6) savannas and seeps. These lower fire-frequency than the other longleaf largely conform to the types recognized in our types owing to insufficient fuel production to general gradient model above (Fig. 2), with carry frequent fire. Nonetheless, the vegeta- the exception that clayey and rocky uplands tion is prone to degradation from fire suppres- are recognized primarily from the Piedmont sion in that in the absence of fire scrub oaks and Mountains. I then sorted the associations such as Quercus laevis assume a dense crown within a group into the six ecoregions de- cover, shading out much of the sparse herb and scribed above (Fig. 1): (1) Atlantic Coastal grass layer and leaving a more fire resistant, Plain, (2) Fall-line Sandhills, (3) Southern desertlike Quercus laevis community (1.2.2 — Coastal Plain, (4) Eastern Gulf Coastal Plain, see Table 2). On the Western Gulf Coast, be- (5) Western Gulf Coastal Plain, and (6) Pied- yond the range of Quercus laevis, the same phe- mont and Montane Uplands. Where our plot- nomenon occurs but with Quercus incana and based studies suggest a need for revision of the Q. margarettiae assuming dominance (1.5.2). National Classification, these changes are implemented in the compilation of types. In the following sections I refer to the 135 Atlantic Coastal Plain vegetation associations recognized in Table 2 Pine barrens and extreme xeric woodlands of by three-digit code where the first digit refers the Atlantic Coastal Plain are best developed to the six ecological groups, the second refers on eolian sands on the northeast sides of ma- to the six ecoregions, and the third is a se- jor rivers (e.g., Altamaha, Cape Fear, Peedee, quence number. Relationships between types Savannah; see Bozeman 1971) and northeast I recognize and types recognized in the 2004 sides of Carolina bays (Wells and Shunk 1931), National Classification are indicated in Table 2 but also occur in other areas with extensive using a standard set of relationship symbols. > sand deposits. On these sites one encounters and < indicate “includes” and “included in,” an understory typically dominated by Quercus respectively; no symbol indicates the commu- laevis, a layer of low shrubs such as Gaylus- nity concepts are similar, and ( ) indicates the sacia dumosa and Vaccinium stamineum, and a name has been revised. >< indicates that the very sparse herb layer of xerophytes such as two community concepts overlap, but each has Stipulicida setacea, Cnidoscolus stimulosus, Rhyn- unique components. Aff. indicates a weak or chospora megalocarpa, Minuartia caroliniana, undefined affinity or relationship. Euphorbia ipecacuanhae, Polygonella polygama, Bulbostylis ciliatifolia, and Selaginella acanthonota (1.1.1). Composition shifts with proximity to the coast where Quercus geminata and Q. hemis- Xeric Sand Barrens and phaerica share dominance with Q. laevis (1.1.3, Uplands 1.1.4); inland on slightly silty soils, Q. margarettiae and Q, incana sometimes share dominance Deep, coarse, well-drained sands can be found with Q. laevis. Grasses are typically unimpor- scattered across the coastal plain portions of tant and wiregrass (Aristida stricta north of 60 II. Ecology < (CEGL003586) < CEGL003590 < (CEGL003590) < (CEGL003590) >< (CEGL003583) > (CEGL003587) > (CEGL004491) > (CEGL008569) CEGL003602 < CEGL003591 CEGL003592 CEGL007513 CEGL003588 4 CEGL003577 3 6 CEGL003589 2 2 9 CEGL003584 10 11 Plots NVC type 2 20 17 a C S C GANCSC 7 States NC SC FL GA SC GA FL? GA SC 2 CEGL004492 LA TX NC SC NC SC GA 2 SC FL NC,SC MS AL FL GA MS 11 NC SC Woodland LA TX Woodland GA var . Woodland Woodland Woodland americanus /N ( baccata, frondosa ) Woodland NC VA Woodland Aristida beyrichiana ) michauxii/Aristida beyrichiana Aristida beyrichiana secundum frondosa – Bulbostylis warei related ecosystems of the Southeastern United States. patens – / Schizachyrium scoparium Woodland scoparium , Woodland arboreum/ Cnidoscolus texanus–Stylisma setacea chrysophylla floridana tenellum/Aristida stricta dumosa/Schizachyrium scoparium dumosa–Gaylussacia Woodland ) Woodland spp. Woodland Woodland Woodland pattersonii Woodland var. ( Rhynchospora megalocarpa, Selaginella acanthonota americanus/Aristida stricta Woodland Woodland Woodland Aristida stricta pickeringii Pinus palustris–Pinus taeda/Quercus geminata–Quercus hemisphaerica–Osmanthus Pinus palustris/Quercus laevis–Quercus incana/Gaylussacia Pinus palustris/Quercus laevis–Quercus geminata/Vaccinium Pinus palustris/Quercus laevis/Aristida purpurascens–Stipulicida Pinus palustris/Quercus laevis–Quercus myrtifolia/Chrysoma pauciflosculosa/ Pinus palustris/Quercus laevis–Quercus incana/Gaylussacia Pinus palustris/Quercus laevis–Q. incana/Aristida beyrichiana–Sorghastrum Pinus palustris/Quercus laevis/Aristida beyrichiana–Tephrosia Pinus palustris/Quercus laevis/Aristida stricta/Cladonia Pinus palustris/Quercus laevis/Gaylussacia dumosa/ ( Schizachyrium Pinus palustris/Quercus laevis/Aristida stricta–Baptisia cinerea–Stylisma Pinus palustris/Quercus laevis–Quercus incana–Quercus margarettiae/Licania Pinus palustris/Quercus laevis/Chrysoma pauciflosculosa Pinus palustris–Pinus (echinata, taeda)–Quercus (incana, margarettiae) Pinus palustris/Quercus laevis–Quercus (incana, margarettiae)/Gaylussacia Pinus palustris/Quercus laevis/Serenoa repens/Aristida condensata Pinus palustris/Quercus incana–Quercus margarettiae/Vaccinium Pinus palustris/Quercus laevis/Licania michauxii/Euphorbia Vegetation associations of fire-maintained longleaf pine and 2. 1.1.4 1.1.2 1.1.3 1.1.1 1.1.5 2.1.2 1.3.1 1.3.2 1.2.2 1.2.3 1.2.1 1.1.6 1.2.4 1.5.2 2.1.1 1.4.2 1.5.1 1.4.1 Atlantic Coastal Plain Southern Coastal Plain Fall–line Sandhills Atlantic Coastal Plain Western Gulf Coastal Plain Eastern Gulf Coastal Plain ABLE T Community Xeric Sand Barrens and Uplands Subxeric Sandy Uplands 3. Ecological Classification of Longleaf Pine Woodlands 61 ( cont. ) > (CEGL003593) < CEGL003591 < CEGL003586 CEGL007842 CEGL007844 < (CEGL007714) >< (CEGL003583) CEGL008491 CEGL004957 CEGL003601 6 3 6 6 4 CEGL004489 4 CEGL004488 4 CEGL004490 4 CEGL004487 7 6 8 aff CEGL007714 4 aff CEGL003601 4 6 18 28 (CEGL008586) 13 18 SC SC, GA GA SC NC AL? FL 14 GA SC GA GA SC GA SC GA NC FL GA? MS NC,SC FL AL GA? MS? NC,SC FL FL FL FL LA MS? AL, GA, MS ( croceus, Woodland SC,GA Woodland Woodland Woodland Woodland Woodland Woodland Woodland Woodland Woodland scoparium Woodland atrorubens georgiana var. junceus–Nolina georgiana georgiana stricta–Astragalus michauxii reniformis stoloniferum–Cyperus – Aristida stricta beyrichiana perfoliata var. Woodland Woodland Woodland var. scoparium aquilinum – Polygonella gracilis Woodland crassifolium Woodland scoparium walteri Woodland pubescens/Schizachyrium plukenetii laevis)/ Schizachyrium scoparium beyrichiana–Nolina pubescens/Aristida beyrichiana–Nolina michauxii/Aristida beyrichiana–Dyschoriste beyrichiana –Pterocaulon virgatum beyrichiana–Helianthus stamineum/Aristida virginiana Woodland alifanus Woodland Woodland Woodland pseudocaudatum–Schizachyrium scoparium–Tephrosia virginiana var. pseudocaudatum–Schizachyrium scoparium–Symphyotrichum Woodland oblongata Woodland odoratissimus retrorsus ) Woodland stoloniferum–Ionactis linariifolius Pinus palustris/Quercus laevis–(Quercus incana)/Toxicodendron Pinus palustris/Quercus laevis–Quercus incana/Aristida beyrichiana–Baptisia Pinus palustris/Quercus incana–Quercus marilandica/Aristida Pinus palustris/Quercus marilandica–Quercus laevis/Aristida Pinus palustris/Quercus laevis/Gaylussacia dumosa/Aristida Pinus palustris/Quercus laevis/Serenoa repens–Vaccinium Pinus palustris/Clethra alnifolia–Quercus pumila/Pteridium aquilinum Pinus palustris/Quercus hemisphaerica/Gaylussacia dumosa/Pteridium Pinus palustris/Gaylussacia dumosa–Vaccinium tenellum/Vaccinium Pinus palustris/Quercus incana–Quercus stellata/Aristida beyrichiana–Sporobolus Pinus palustris/Quercus laevis/Aristida stricta–Tephrosia Pinus palustris/Quercus laevis–Quercus margarettiae/Licania Pinus palustris–Pinus (echinata, taeda)/Quercus (marilandica, Pinus palustris/Aristida stricta–Coreopsis major–Rhexia Pinus palustris/Quercus laevis–Quercus incana/Toxicodendron Pinus palustris/Quercus minima–Serenoa repens/Aristida beyrichiana–Carphephorus Pinus palustris/Quercus (incana, margarettiae)/Aristida beyrichiana–Rhynchosia Pinus palustris/Quercus geminata/Serenoa repens/Schizachyrium Pinus palustris/Quercus chapmanii–Serenoa repens/Aristida Pinus palustris/Quercus geminata/Serenoa repens–Conradina canescens Pinus palustris/Quercus geminata/Aristida condensata–Cyperus Pinus palustris/Licania michauxii/Aristida beyrichiana–Schizachyrium Pinus palustris/Quercus incana/Sporobolus clandestinus Pinus palustris/Quercus laevis/Serenoa repens–Clinopodium coccineum 2.2.3 2.2.1 2.2.2 2.1.9 2.1.8 2.1.7 2.1.6 2.1.5 2.1.4 2.1.3 2.2.4 2.4.4 2.4.5 2.2.5 2.2.6 2.3.2 2.3.1 2.3.5 2.3.4 2.3.3 2.3.6 2.4.1 2.4.2 2.4.3 Fall-line Sandhills Southern Coastal Plain Eastern Gulf Coastal Plain 62 II. Ecology > (CEGL004485) > (CEGL007749) < (CEGL003573) < (CEGL003569) CEGL003580 < (CEGL003569) >< (CEGL003664) CEGL003575 CEGL008572 CEGL008571 CEGL004945 CEGL008452 9 5 8 (CEGL007738) 9 7 Plots NVC type 20 CEGL004496 GA FL NC, SC 21 (CEGL003578) LA TX States AL? FL 21 NC, SC? 9 GA SC SC LA? TX AL, FL?, GA SC,GANC 17 NC NC,SC 14 (CEGL003595) AL LA MS NC, SC? 16 LA TX SC NC, SC? 8 AL GA Woodland GA SC 16 Woodland Woodland var . pilosa stoloniferum–Liatris var. Woodland var. aquilinum Aristida Woodland / pilosa – Woodland Woodland / Woodland Schizachyrium scoparium / scoparium Woodland – Schizachyrium scoparium nutans Woodland / Schizachyrium scoparium Woodland atlanticum Woodland pubescens pumila stricta ephrosia virginiana argyranthemus elegans–Opuntia integrifolium pumila rotundifolium texana scoparium beyrichiana–Pteridium Loam–hill Woodland Woodland Woodland Woodland Woodland Woodland ) var . humifusa Woodland humifusa Dyschoriste oblongifolia gracilis Aristida stricta–Tephrosia virginiana stricta–Pycnanthemum flexuosum pseudocaudatum–Schizachyrium scoparium Woodland Pinus palustris/Quercus incana/Schizachyrium scoparium–Croton Pinus palustris/Quercus incana/Schizachyrium scoparium–Liatris Pinus palustris/Quercus falcata/Cornus florida/Aristida beyrichiana Pinus palustris/Schizachyrium scoparium var. stoloniferum–Sorghastrum Pinus palustris/Schizachyrium scoparium/Verbesina aristata Pinus palustris/Vaccinium myrsinites/Aristida beyrichiana–Schizachyrium Pinus palustris/Schizachyrium scoparium–(Aristida beyrichiana)–T Pinus palustris–Pinus taeda/Carya pallida–Cornus florida/Aristida Pinus palustris/Aristida stricta–Panicum virgatum–Eupatorium Pinus palustris/Quercus laevis–Quercus marilandica/Toxicodendron Pinus palustris/Quercus marilandica/Aristida stricta–Parthenium Pinus palustris/Quercus marilandica/Clethra alnifolia–Rhododendron Pinus palustris/Quercus falcata/Cornus florida/Schizachyrium Pinus palustris/Quercus marilandica/ Vaccinium tenellum–Quercus Pinus palustris/Gaylussacia dumosa/Aristida stricta–Ionactis linariifolius–Liatris Pinus palustris–Pinus (echinata, taeda)–Quercus falcata–Carya Pinus palustris/Quercus pumila–Vaccinium tenellum/Aristida beyrichiana Pinus palustris/Quercus stellata/Vaccinium tenellum/Aristida Pinus palustris/Clethra alnifolia–Gaylussacia frondosa–Quercus Pinus palustris/Vaccinium tenellum/Aristida stricta–Desmodium tenuifolium ( Continued 2. 2.5.1 2.5.2 3.4.2 3.4.1 3.4.4 3.4.3 3.2.4 3.2.5 3.2.3 3.2.1 3.2.2 3.1.7 3.4.5 3.1.4 3.1.6 2.5.3 3.1.1 3.1.2 3.1.3 3.1.5 Western Gulf Coastal Plain Atlantic Coastal Plain Eastern Gulf Coastal Plain Fall-line Sandhills ABLE Community Silty Uplands T 3. Ecological Classification of Longleaf Pine Woodlands 63 ( cont. ) CEGL003572 CEGL003581 CEGL003597 > CEGL003599 > (CEGL007767) CEGL004783 CEGL003596 CEGL003598 CEGL004774 CEGL003571 CEGL004060 CEGL003600 CEGL003579 CEGL008580 CEGL004955 CEGL004432 5 1 CEGL003608 C LA TX LA TX? LA AL? MS AL GANC SC? 1 CEGL008437 GA LA TX MS LA TX GA AL LA LA MS? SC, NC?NC SC 4 12 LA TX TX GA Woodland Woodland – Ruellia humilis leucophaea – var. tenerum–Rhexia alifanus Woodland –N gracile Woodland laciniatum spp. nuttallii bracteata stricta virginica Woodland scoparium Woodland cuthbertii–Penstemon ernonia angustifolia var . alismifolia Woodland marilandica–(Quercus accinium pallidum georgiana Woodland nuttallii/Cladonia Woodland var . pseudocaudatum Forest Woodland Woodland Altamaha Grit Herbaceous Vegetation Woodland Woodland Packera obovata Woodland Herbaceous Vegetation prinus)/Vaccinium pallidum dissectus Woodland Vaccinium tenellum/Pteridium aquilinum Serpentine Woodland Pinus palustris/Schizachyrium scoparium–Coreopsis tripteris–Baptisia Pinus palustris–Pinus (echinata, taeda)/Schizachyrium tenerum–V Pinus palustris/Schizachyrium scoparium–Liatris pycnostachya Pinus palustris/Schizachyrium scoparium–Rudbeckia grandiflora Pinus palustris/Kalmia latifolia–Vaccinium arboreum Pinus palustris/Schizachyrium scoparium–Schizachyrium tenerum–Silphium Pinus palustris/Quercus marilandica/Schizachyrium tenerum–Bigelowia Pinus palustris/Quercus marilandica/Vaccinium crassifolium/Aristida Pinus palustris/Quercus marilandica/Schizachyrium scoparium–Schizachyrium Pinus palustris/Quercus marilandica/Ilex vomitoria/Schizachyrium Pinus palustris/Quercus marilandica/Panicum virgatum Pinus palustris–Pinus echinata–(Pinus virginiana)/Quercus Pinus palustris–(Pinus echinata)/Oxydendrum arboreum/V (Pinus palustris)/Bigelowia nuttallii–Talinum teretifolium–Allium Pinus palustris/Quercus marilandica/Schizachyrium scoparium–Silphium Quercus prinus–Quercus palustris Pinus palustris–Pinus echinata/Schizachyrium scoparium–Manfreda Pinus palustris–Pinus echinata/Quercus coccinea–Quercus 3.4.7 3.4.8 3.5.1 3.5.2 4.1.1 3.5.3 3.5.4 4.1.2 3.4.6 4.5.2 4.5.3 4.5.4 ( Pinus palustris )/ Schizachyrium scoparium–Bigelowia 4.6.1 4.6.2 4.1.3 4.4.1 Insufficient data to document; no extant stands 4.5.1 4.6.3 4.6.4 4.6.5 Western Gulf Coastal Plain Atlantic Coastal Plain and Fall-line Sandhills Piedmont and Montane Uplands Eastern Gulf Coastal Plain Western Gulf Coastal Plain Rocky and Clayey Uplands 64 II. Ecology < (CEGL004658) > (CEGL003808) >< (CEGL003653) >< (CEGL003653) > CEGL004967 > (CEGL004969) < (CEGL004791) > (CEGL003797) CEGL003650 < CEGL004658 CEGL003658 CEGL003647 CEGL004680 CEGL003656 CEGL003662 2 6 5 7 8 (CEGL007750) 1 CEGL004113 9 aff CEGL003795 1 19 (CEGL004236) 1712 CEGL003661 CEGL003649 L F L A GA? LA MS SC GA FL GAFL 8 GA CEGL004486 7 FL FL GA? NC SC FL LA MS FL GA AL FL 11 FL States Plots NVC type NC NC NC VA AL FL MS NC SC 28 (CEGL003648) AL FL FL FL Woodland FL Woodland Shrubland Shrubland Woodland Woodland –G racemosa Woodland Woodland floridana /A uniflora Herbaceous Vegetation Aristida beyrichiana Woodland – Carphephorus odoratissimus carolina) vomitoria repens – Woodland – Elephantopus elatus elliottii expansa racemiflora Woodland crassifolium–Aristida stricta Woodland beyrichiana myrtifolia–Quercus beyrichiana frondosa–(Kalmia Woodland Woodland elliottii)/Sporobolus pinetorum–Oclemena reticulata Woodland var. Woodland elliottii)/Gaylussacia nana–Serenoa repens–Vaccinium var. Herbaceous Vegetation Woodland elliottii/Ilex vomitoria–Serenoa repens–Morella cerifera elliottii/Quercus chapmanii–Quercus geminata–Serenoa densa/Quercus minima/Panicum abscissum elliottii–Pinus palustris/Lyonia ferruginea–Serenoa repens/ elliottii/Lyonia lucida–Serenoa repens/Aristida beyrichiana–Befaria var. var. var. var. var. var . elliottii/Serenoa repens–Ilex glabra–Morella cerifera–Ilex ) Rhynchospora megalocarpa Woodland geminata–Lyonia ferruginea Pinus palustris–Pinus serotina/Quercus chapmanii–Quercus Quercus (chapmanii, myrtifolia)–Serenoa repens/Aristida beyrichiana–Chapmannia Pinus elliottii Pinus elliottii Panicum abscissum Pinus elliottii Pinus elliottii Lyonia fruticosa–Serenoa repens–Quercus minima/Aristida Serenoa repens/Aristida beyrichiana–Scleria muehlenbergii–Eragrostis Pinus palustris–Pinus serotina/Pleea tenuifolia–Aristida stricta Pinus palustris/Leiophyllum buxifolium/Aristida stricta Pinus palustris–(Pinus serotina)/Ilex glabra–Gaylussacia Pinus palustris/Serenoa repens–Vaccinium myrsinites/Aristida Pinus palustris–(Pinus elliottii Sporobolus curtissii ( Sporobolus curtissii ) Woodland Pinus (palustris, elliottii myrsinites/Sporobolus (floridanus, curtissii) Pinus palustris/Serenoa repens–Quercus pumila/Aristida beyrichiana–Balduina Pinus elliottii Pinus palustris–(Pinus elliottii var. elliottii)/Ilex coriacea–Cyrilla Pinus palustris–Pinus serotina/Gaylussacia dumosa/Vaccinium Pinus palustris–Pinus serotina/Ctenium aromaticum–Muhlenbergia Pinus elliottii ( Continued 2. 5.3.9 5.3.5 5.3.6 5.3.7 5.3.8 5.1.3 5.1.4 5.1.5 5.3.1 5.3.2 5.3.3 5.3.4 5.4.2 5.4.3 5.1.1 5.1.2 5.4.1 5.3.13 5.3.11 5.3.12 5.3.10 Southern Coastal Plain Atlantic Coastal Plain Eastern Gulf Coastal Plain ABLE Community Flatwoods T 3. Ecological Classification of Longleaf Pine Woodlands 65 ( cont. ) < (CEGL004958) < (CEGL003653) > (CEGL004086) > (CEGL004501) > (CEGL003795) > (CEGL004152) > (CEGL004154) CEGL004790 < (CEGL004105) > (CEGL004153) > (CEGL003673) CEGL003659 > (CEGL003645) > (CEGL004155) < (CEGL004105) CEGL004956 CEGL004687 4 6 3 3 5 1 CEGL004498 3 6 4 CEGL004499 2 CEGL004500 7 CEGL004502 2 CEGL004085 1 aff CEGL004790 5 2 A G L NC SC 16 FL SC NC 5 CEGL004495 GA SC FL FL FL NC,SC 7 NC SC SC AL FL 13 SC FL GA SC 2 CEGL004497 SC NC NC GA NC SCSC 5 CEGL003660 LA MS? FL LA MS FL Woodland Woodland Woodland Woodland Herbaceous Herbaceous Woodland Woodland ensifolium ( Sporobolus floridanus ) Woodland Woodland tracyi Woodland Herbaceous Vegetation var. flava Herbaceous vegetation –F Herbaceous Vegetation – Oxypolis filiformis Woodland radula Herbaceous Vegetation – Calamovilfa brevipilis – Rhynchospora latifolia – Eriocaulon decangulare dichotomum pseudoliatris virgatum expansa expansa erectifolium Woodland ephrosia spicata floridanus muehlenbergii teretifolius–Carex striata aromaticum radula –Galactia erecta –Woodland ssp . tecta–Liquidambar styraciflua/Andropogon glomeratus–Sarracenia var . elliottii/Styrax americanus/Sporobolus floridanus var . elliottii/Ctenium aromaticum–Aristida beyrichiana– Woodland Woodland Herbaceous Vegetation var . elliottii–Sabal palmetto/Morella cerifera/Panicum var . elliottii/Aristida beyrichiana–Eriocaulon decangulare–Lachnanthes var . elliottii/Ilex glabra/Andropogon glaucopsis–Panicum rigidulum Woodland Herbaceous Vegetation Woodland Vegetation minor Woodland integrifolium Vegetation macrophylla Panicum hemitomon caroliana Pinus palustris/Sporobolus pinetorum–Schizachyrium scoparium–Eryngium Aristida beyrichiana–Amphicarpum muehlenbergianum–Scleria Aristida beyrichiana–Rhynchospora oligantha–Scleria muehlenbergii–Sabatia Pinus palustris/Ilex glabra/Aristida beyrichiana–Helianthus Pinus palustris/Arundinaria gigantea Pinus serotina–Pinus palustris/Osmunda cinnamomea/Dichanthelium Pinus palustris–Pinus serotina/Ctenium aromaticum–Muhlenbergia Pinus palustris–Pinus elliottii Pinus palustris–Pinus elliottii Pinus palustris–Pinus serotina/Ctenium aromaticum–Scleria pauciflora–Sarracenia Pinus palustris–Pinus serotina/Magnolia virginiana/Sporobolus Pinus palustris–Pinus serotina/Aristida palustris–Sarracenia flava Pinus palustris–Pinus serotina/Sporobolus pinetorum–Ctenium Pinus palustris–Pinus (serotina, taeda)/Sporobolus curtissii Pinus palustris–Pinus serotina/Ctenium aromaticum–Muhlenbergia Amphicarpum muehlenbergianum–Aristida palustris–Sporobolus Panicum tenerum–Aristida beyrichiana–Schizachyrium rhizomatum–Rhynchospora Pinus elliottii Aristida beyrichiana–Ctenium aromaticum–Andropogon arctatus–Rhynchospora Pinus elliottii Pinus elliottii Rhynchospora oligantha–Sarracenia (alata, psittacina)–Carphephorus chapmanii Pinus palustris/Schizachyrium scoparium–Muhlenbergia expansa–Helianthus Quercus pumila–Rhododendron canescens/Ctenium aromaticum–T Dichanthelium wrightianum–Panicum verrucosum–Dichanthelium 6.1.1 6.3.8 6.3.4 6.3.3 6.1.8 6.2.1 6.2.2 6.3.1 6.3.2 6.1.7 6.1.6 6.1.4 6.1.5 6.1.3 6.1.2 6.3.9 6.3.5 6.4.1 6.3.7 6.3.6 6.4.2 6.4.3 6.4.4 6.4.5 6.3.10 Atlantic Coastal Plain Fall–line Sandhills Southern Coastal Plain Eastern Gulf Coastal Plain Savannas and Seeps 66 II. Ecology CEGL004175 CEGL007802 CEGL003654 CEGL003646 CEGL004792 other), () (roughly LA TX LA TX States Plots NVC type LA TX? LA TX? FL GA?LA? MS 2 NC SC? 2 CEGL003663 Where the type was derived or validated in one of or circumscription of the association differs from the by NatureServe (see NatureServe 2005), (2) Duncan the USNVC are indicated by NatureServe codes. Where the Woodland (overlaps, but each type includes occurrences not in the Herbaceous Piedmont Woodland Woodland >< Woodland croceum Woodland virgatum –( Helenium drummondii ) asteroides (includes), nivea > code is shown in parentheses. Where the concept laxum–Panicum et al. (1998), and (5) Peet, Carr and Gramling (2006). spp .–(Ctenium aromaticum) (included in), < National Vegetation Classification (USNVC) as maintained as or undefined affinity or similarity). vegetation plots is indicated. Synonymy relationships with ssp . tecta/Ctenium aromaticum–Pycnanthemum flexuosum var . elliottii)/Ilex vomitoria/– Agalinis filicaulis ) Vegetation Woodland Pinus palustris–Pinus taeda–Pinus serotina/Chasmanthium Pinus palustris/Sporobolus silveanus–Muhlenbergia capillaris–Chaetopappa Pinus palustris/Eryngium integrifolium–Rhynchospora Sarracenia alata–Rhynchospora gracilenta–Rudbeckia scabrifolia–Schoenolirion Pinus palustris–(Pinus elliottii Pinus palustris/Rhynchospora elliottii–Lobelia flaccidifolia–Platanthera Pinus serotina/Arundinaria gigantea ( Continued 2. 6.6.1 6.5.4 6.5.2 6.5.3 6.4.7 6.5.1 6.4.6 Piedmont and Montane Uplands Western Gulf Coastal Plain Associations shown were collected from (1) the U.S. ABLE Community name used differs from the Name in the USNVC, the NatureServe the last four references, the number of supporting and Peet (1994), (3) Peet and Duncan (1994), (4) Kjellmark USNVC type, the relationship is indicated symbolically T a equivalent, though with a different name), aff. (weak 3. Ecological Classification of Longleaf Pine Woodlands 67

FIGURE 3. Xeric sand barrens and uplands, Atlantic Coastal Plain. Pinus palustris and Quercus laevis barren on the rim of a Carolina bay. The characteristic sparse ground layer consists primarily of small mats of Selaginella acanthonota and Minuartia caroliniana. Salters Lake, Bladen County, NC. central South Carolina, A. beyrichiana to the Fall-line Sandhills south) is limited to the favorable microsites. There is a conspicuous latitudinal gradient in The extreme xeric sites of the Fall-line Sand- composition; the few examples remaining in hills resemble those of the Atlantic Coastal northeastern North Carolina and southeastern Plain in the co-dominance of Pinus palustris Virginia (1.1.2) have a conspicuously low di- and Quercus laevis, as well as the occurrence versity of xerophytic forbs and woody plants of extreme xerophytic herbs such as Stipuli- compared with examples from the south- cida setacea, Cnidoscolus stimulosus, and Euphor- ern extreme of the ecoregion. In particu- bia ipecacuanhae. Grasses remain sparse but are lar, Chrysoma pauciflosculosa and Ceratiola eri- somewhat more continuous than on the ex- coides occur occasionally northward into South treme Coastal Plain sites and with the same lat- Carolina and Quercus chapmanii and Q. myrtifo- itudinal gradient in dominance: Aristida stricta lia occupy extreme sites on dune systems as- in the north (1.2.1), Aristida beyrichiana in sociated with rivers in Georgia (1.1.5). Licania the south (1.2.3), and Schizachyrium scoparium michauxii occurs on a few sites north of the throughout but particularly conspicuous be- Savannah River where it defines the north- tween the ranges of the two wiregrasses in ern boundary of a longleaf, mixed oak type of South Carolina. As on the true Coastal Plain, xeric sands with the xerophytes more typical of some extreme sites in the Georgia and South Georgia where additional southern taxa such Carolina sandhills support a shrub layer with as Nolina georgiana and Serenoa repens occur specialist xerophytes such as Chrysoma pauci- (1.1.6). flosculosa and Ceratiola ericoides (1.2.4). 68 II. Ecology

TABLE 3. Scientific names in text and associated common names.a Scientific name Common name Agalinis filicaulis Jackson false foxglove Aletris colicroot Allium cuthbertii striped garlic Amphicarpum muehlenbergianum Muhlenberg maidencane Andropogon arctatus pinewoods bluestem Andropogon glaucopsis purple bluestem Andropogon glomeratus bushy bluestem Andropogon gyrans Elliott’s bluestem Andropogon gyrans var. stenophyllus Elliott’s bluestem Andropogon mohrii Mohr’s bluestem Andropogon virginicus broomsedge bluestem Anthaenantia rufa purple silkyscale Aristida beyrichiana Beyrich threeawn (Southern wiregrass) Aristida condensata piedmont threeawn Aristida mohrii Mohr’s threeawn Aristida palustris longleaf threeawn Aristida purpurascens arrowfeather threeawn Aristida stricta pineland threeawn (Carolina wiregrass) Arnoglossum floridanum Florida cacalia Arundinaria gigantea ssp. tecta switchcane Astragalus michauxii sandhills milkvetch Balduina angustifolia coastalplain honeycombhead Balduina uniflora oneflower honeycombhead Baptisia alba white wild indigo Baptisia bracteata var. leucophaea longbract wild indigo Baptisia cinerea grayhairy wild indigo Baptisia perfoliata catbells Befaria racemosa tarflower Berlandiera subacaulis Florida greeneyes Bigelowia nuttallii Nuttall’s rayless goldenrod Bulbostylis ciliatifolia capillary hairsedge Bulbostylis warei Ware’s hairsedge Calamovilfa brevipilis pine barren sandreed Calopogon grasspink Carex lutea sulphur sedge Carex striata Walter’s sedge Carphephorus corymbosus coastalplain chaffhead Carphephorus odoratissimus vanillaleaf Carphephorus paniculatus hairy chaffhead Carphephorus pseudoliatris bristleleaf chaffhead Carya pallida sand hickory Carya texana black hickory Chaetopappa asteroides Arkansas leastdaisy Chapmannia floridana Florida alicia Chasmanthium laxum slender woodoats Chrysoma pauciflosculosa woody goldenrod Chrysopsis mariana Maryland goldenaster Cladina reindeer lichen Cladonia cup lichen Cleistes rosebud orchid Clethra alnifolia coastal sweetpepperbush Clinopodium coccineum scarlet calamint Clinopodium georgianum Georgia calamint Cnidoscolus stimulosus finger rot (spurge-nettle) Cnidoscolus texanus Texas bullnettle 3. Ecological Classification of Longleaf Pine Woodlands 69

TABLE 3. (Continued ) Scientific name Common name Conradina canescens false rosemary Coreopsis major greater tickseed Coreopsis tripteris tall tickseed Cornus florida flowering dogwood Croton argyranthemus healing croton Ctenium aromaticum toothache grass Cyperus croceus Baldwin’s flatsedge Cyperus plukenetii Plukenet’s flatsedge Cyperus retrorsus pine barren flatsedge Cyrilla racemiflora swamp titi Desmodium tenuifolium slimleaf ticktrefoil Dichanthelium dichotomum var. ensifolium cypress panicgrass Dichanthelium erectifolium erectleaf panicgrass Dichanthelium leucothrix rough panicgrass Dichanthelium wrightianum Wright’s rosette grass Dionaea muscipula Venus flytrap Diospyros virginiana common persimmon Drosera sundew Dyschoriste oblongifolia oblongleaf snakeherb Echinacea sanguinea sanguin purple coneflower Elephantopus elatus tall elephantsfoot Eragrostis elliottii field lovegrass Eriocaulon decangulare tenangle pipewort Eriocaulon decangulare var. decangulare tenangle pipewort Eriogonum tomentosum dogtongue buckwheat Eryngium integrifolium blueflower eryngo Eupatorium mohrii Mohr’s thoroughwort Eupatorium rotundifolium roundleaf thoroughwort Euphorbia floridana Greater Florida spurge Euphorbia ipecacuanhae American ipecac Euthamia tenuifolia var. tenuifolia slender goldentop Galactia erecta erect milkpea Gaylussacia baccata black huckleberry Gaylussacia dumosa dwarf huckleberry Gaylussacia frondosa blue huckleberry Gaylussacia nana Confederate huckleberry Gaylussacia tomentosa hairytwig huckleberry Helenium drummondii fringed sneezeweed Helianthus atrorubens purpledisk sunflower Helianthus radula rayless sunflower Hypericum fasciculatum peelbark St. Johnswort Hypericum lloydii sandhill St. Johnswort Ilex coriacea large gallberry Ilex glabra inkberry Ilex vomitoria yaupon Ionactis linariifolius flaxleaf whitetop aster Kalmia carolina Carolina laurel Kalmia hirsuta hairy laurel Kalmia latifolia mountain laurel Lachnanthes caroliana Carolina redroot Leiophyllum buxifolium sandmyrtle Liatris elegans pinkscale blazing star Liatris gracilis slender blazing star Liatris pilosa var. pilosa shaggy blazing star Liatris pycnostachya prairie blazing star Licania michauxii gopher apple 70 II. Ecology

TABLE 3. (Continued ) Scientific name Common name Lilium lily Liquidambar styraciflua sweetgum Lobelia flaccidifolia foldear lobelia Ludwigia linifolia southeastern primrose-willow Lyonia ferruginea rusty staggerbush Lyonia fruticosa coastalplain staggerbush Lyonia lucida fetterbush lyonia Magnolia grandiflora southern magnolia Magnolia virginiana sweetbay Manfreda virginica false aloe Minuartia caroliniana pinebarren stitchwort Morella cerifera wax myrtle Muhlenbergia capillaris hairawn muhly Muhlenbergia expansa cutover muhly Nolina georgiana Georgia beargrass Nyssa sylvatica blackgum Oclemena reticulata pinebarren whitetop aster Opuntia humifusa var. humifusa devil’s-tongue (Eastern pricklypear) Osmanthus americanus var. americanus devilwood (wild olive) Osmunda cinnamomea cinnamon fern Oxydendrum arboreum sourwood Oxypolis filiformis water cowbane Packera obovata roundleaf ragwort Panicum abscissum cutthroat grass Panicum hemitomon maidencane Panicum rigidulum redtop panicgrass Panicum tenerum bluejoint panicgrass Panicum verrucosum warty panicgrass Panicum virgatum switchgrass Parnassia caroliniana Carolina grass of Parnassus Parthenium integrifolium wild quinine Penstemon dissectus dissected beardtongue Pinguicula butterwort Pinus echinata shortleaf pine Pinus elliottii var. densa Florida slash pine Pinus elliottii var. elliottii Honduras pine (slash pine) Pinus palustris longleaf pine Pinus serotina pond pine Pinus taeda loblolly pine Pinus virginiana Virginia pine Pityopsis aspera pineland silkgrass Platanthera fringed orchid Platanthera nivea snowy orchid Pleea tenuifolia rush featherling Pogonia pogonia Polygonella gracilis tall jointweed Polygonella polygama October flower Pteridium aquilinum var. pseudocaudatum western brackenfern (tailed bracken) Pterocaulon virgatum wand blackroot Pycnanthemum flexuosum Appalachian mountainmint Pyxidanthera barbulata flowering pixiemoss Quercus chapmanii Chapman oak Quercus coccinea scarlet oak Quercus falcata southern red oak Quercus geminata sand live oak Quercus georgiana Georgia oak 3. Ecological Classification of Longleaf Pine Woodlands 71

TABLE 3. (Continued ) Scientific name Common name Quercus hemisphaerica Darlington oak (sand laurel oak) Quercus incana bluejack oak Quercus inopina sandhill oak Quercus laevis turkey oak Quercus margarettiae runner oak (sand post oak) Quercus marilandica blackjack oak Quercus minima dwarf live oak Quercus myrtifolia myrtle oak Quercus palustris pin oak Quercus prinus chestnut oak Quercus pumila running oak Quercus stellata post oak Rhexia alifanus savannah meadowbeauty Rhododendron atlanticum dwarf azalea Rhododendron canescens mountain azalea Rhynchosia cytisoides royal snoutbean Rhynchosia reniformis dollarleaf Rhynchospora beaksedge Rhynchospora chapmanii Chapman’s beaksedge Rhynchospora elliottii Elliott’s beaksedge Rhynchospora gracilenta slender beaksedge Rhynchospora latifolia sandswamp whitetop Rhynchospora megalocarpa sandyfield beaksedge Rhynchospora oligantha featherbristle beaksedge Rhynchospora tracyi Tracy’s beaksedge Rudbeckia grandiflora var. alismifolia rough coneflower Rudbeckia scabrifolia roughleaf coneflower Ruellia humilis fringeleaf wild petunia Sabal palmetto cabbage palmetto Sabatia macrophylla largeleaf rose gentian Sarracenia pitcherplant Sarracenia alata yellow trumpets Sarracenia flava yellow pitcherplant Sarracenia minor hooded pitcherplant Sarracenia psittacina parrot pitcherplant Schizachyrium rhizomatum Florida little bluestem Schizachyrium scoparium little bluestem Schizachyrium scoparium var. stoloniferum creeping bluestem Schizachyrium tenerum slender little bluestem Schoenolirion croceum yellow sunnybell Scleria muehlenbergii Muehlenberg’s nutrush Scleria pauciflora fewflower nutrush Selaginella acanthonota spiny spikemoss Serenoa repens saw palmetto Silphium compositum kidneyleaf rosinweed Silphium gracile slender rosinweed Silphium laciniatum compassplant Sorghastrum nutans Indiangrass Sorghastrum secundum lopsided Indiangrass Spiranthes ladies’-tresses Sporobolus clandestinus rough dropseed Sporobolus curtissii Curtis’ dropseed Sporobolus floridanus Florida dropseed Sporobolus junceus pineywoods dropseed Sporobolus pinetorum Carolina dropseed Sporobolus silveanus Silveus’ dropseed 72 II. Ecology

TABLE 3. (Continued ) Scientific name Common name Sporobolus teretifolius wireleaf dropseed Stipulicida setacea pineland scalypink Stylisma patens coastalplain dawnflower Stylisma pickeringii var. pattersonii Patterson’s dawnflower Styrax americanus var. pulverulentus downy American snowbell Styrax americanus American snowbell Symphyotrichum adnatum scaleleaf aster Symphyotrichum walteri Walter’s aster Talinum teretifolium quill fameflower Tephrosia chrysophylla scurf hoarypea Tephrosia mohrii pineland hoarypea Tephrosia spicata spiked hoarypea Tephrosia virginiana Virginia tephrosia Thalictrum cooleyi Cooley’s meadow-rue Tofieldia tofieldia Toxicodendron pubescens Atlantic poison oak Utricularia bladderwort Vaccinium arboreum farkleberry Vaccinium crassifolium creeping blueberry Vaccinium darrowii Darrow’s blueberry Vaccinium fuscatum black highbush blueberry Vaccinium myrsinites shiny blueberry Vaccinium pallidum Blue Ridge blueberry Vaccinium stamineum deerberry Vaccinium tenellum small black blueberry Verbesina aristata coastalplain crownbeard Verbesina chapmanii Chapman’s crownbeard Vernonia angustifolia tall ironweed Zigadenus deathcamas

a Scientiﬁc and common names follow USDA, NRCS (2005), except for Sporobo- lus, which follows Peterson et al. (2003), Muhlenbergia, which follows Peterson (2003), and Styrax, which follows Gonsoulin (1974). Alternate common names in common usage are suggested in parentheses.

Southern Coastal Plain floridanum, Desmodium floridanum, and Berlandiera subacaulis in north-central penin- Although Quercus geminata is confined to the sular Florida (1.3.1) and Bulbostylis warei, maritime climate of the extreme coastal fringe Tephrosia chrysophylla, Balduina angustifolia, in the Atlantic Coastal Plain Ecoregion, it oc- and Carphephorus corymbosus on sites in the curs throughout peninsular Florida, probably central highlands (1.3.2). Where fire has been because of the more moderate climate. Here suppressed for some years, Ceratiola ericoides extreme xeric pine barrens are co-dominated tends to invade, reducing herb layer cover, and by Quercus geminata, Q. laevis, and the ubiqui- shifting the fire regime toward less frequent, tous longleaf pine. Quercus incana can also be more catastrophic events. important, but decreases in importance south of northern Florida. The sparse grass layer is diverse with Aristida beyrichiana being joined Eastern Gulf Coastal Plain by southern specialties such as Schizachyrium scoparium var. stoloniferum, and Sorghastrum se- Vegetation of the Eastern Gulf Coastal Plain cundum. Several species act as strong indica- is reminiscent of that of the Atlantic Coastal tors of Florida barrens including Arnoglossum Plain, particularly the coastal fringe of western 3. Ecological Classification of Longleaf Pine Woodlands 73

Florida where the subcanopy is primarily dom- dissipates via percolation or evaporation. Den- inated by Quercus laevis and the understory has sity and height of understory tree vegetation a significant component of Aristida beyrichiana. depends heavily on past fire and land-use his- However, Aristida beyrichiana is largely absent tory with fire suppression leading to increased from the northern half of the western Panhan- oak density, though upland oak species are dle, the range limit (and edge of the ecore- common throughout. These species typically gion) crossing Eglin Air Force Base (Rodgers include Quercus laevis, Q. incana, Q. margaret- and Provencher 1999) and eventually reach- tiae and in the more southern areas Q. gemi- ing the coast in eastern Mississippi (Peet and nata. Unlike in the xeric barrens, the ground Allard 1993; Peet 1993). As in the southern layer is a nearly continuous sward of grass. portion of the Atlantic Coastal Plain, there is a Wiregrass (Aristida stricta, A. beyrichiana)isthe high frequency of Eriogonum tomentosum, Lica- dominant grass within its range, and elsewhere nia michauxii, Andropogon virginicus, and Pityop- Schizachyrium scoparium dominates (Figs. 4, 5). sis aspera, but also present are such distinctive taxa as Rhynchosia cytisoides, Tephrosia mohrii, and Aristida mohrii (1.4.1). Westward the im- Atlantic Coastal Plain pact of the Mississippi River drainage is ex- Subxeric longleaf woodlands of the Atlantic pressed in the increasing content of silt in sedi- Coast Plain are generally of two main types, ments relative to sand such that extreme xeric one occupying low sandhills and the other sites are much less common west of Florida. on flatter, moister terrain transitional to A relatively distinctive form does occur in the flatwoods—here called dry flatwoods. The vicinity of Camp Shelby, Mississippi Quercus sandhill systems generally have an under- laevis dominates the understory and Serenoa story layer of Quercus laevis and Q. incana. repens is an important component, but, being In the north, Aristida stricta dominates the beyond the range of wiregrass, the grass layer is ground layer (2.1.1); in the wiregrass gap of dominated by Aristida condensata and to a lesser central South Carolina Schizachyrium scopar- extent by Andropogon ternarius and Sorghastrum ium dominates (2.1.2); and south of the gap secundum (1.4.2). from southern South Carolina southward Aris- tida beyrichiana dominates with a scattering of Western Gulf Coastal Plain Sporobolus junceus (2.1.3). The dry flatwoods types also differ latitudinally. In the north Aris- Extreme xeric sites of the Western Gulf Coastal tida stricta is again dominant, and Vaccinium Plain are distinctive because they are beyond crassifolium can be very important as it is in the ranges of both wiregrass and Quercus lae- true flatwoods of the region (2.1.4). Coast- vis (see Bridges and Orzell 1989; Harcombe ward in the wiregrass gap there is a maritime et al. 1993). The dominant oaks here are Quer- form with Quercus hemisphaerica (2.1.5) and cus incana and Q. margarettiae, and the herb a more inland form with Quercus pumila im- layer contains a number of western specialties portant (2.1.6). From southeasternmost South closely related to eastern taxa, two clear ex- Carolina south along the eastern portion of amples being Cnidoscolus texanus and Stylisma the Atlantic Coastal Plain are dry flatwoods pickeringii var. pattersonii (1.5.1). With fire sup- where Aristida beyrichiana is the dominant pression, diversity declines and oak increases grass and Serenoa repens is conspicuous (2.1.7). in importance (1.5.2). Throughout the dry flatwoods, as in the true flatwoods, Pteridium aquilinum var. pseudocaudatum is abundant. Finally, two more types Subxeric Sandy Uplands characterize the uplands of the inner Atlantic Coastal Plain, one with a dense sward of Longleaf pine landscapes with topographic re- Aristida beyrichiana and scattered diagnostic lief of several meters and deep, sandy soils are herbs such as Helianthus atrorubens (2.1.8), and consistently droughty as precipitation rapidly the other characteristic of the Altamaha Grit 74 II. Ecology

FIGURE 4. Subxeric sandy uplands, Southern Coastal Plain. Pinus palustris, Quercus laevis woodland on a ﬂuvial dune. The ground layer of Aristida beyrichiana and Sporobolus junceus is continuous but relatively sparse. Fort Stewart, Bryan County, GA. region with Nolina georgiana as a common uncommon associates Vaccinium staminium and derstory species (2.1.9). Toxicodendron pubescens (2.2.3). From northern South Carolina northward Aristida stricta dominates. These sites, like those in the south, can Fall-line Sandhills be divided into a dry type with Tephrosia vir- Subxeric longleaf woodlands of the Fall-line giniana as an indicator species (2.2.4), and a Sandhills typically have an understory layer mesic type with Rhexia alifanus and Coreopsis dominated by Quercus laevis, which on slightly major as indicators (2.2.5). One other type re- silty sites (“yellow sands”) is joined by Quercus stricted to the northern sandhills is that of dry incana. Where clay layers approach the soil sur- depressions in the xeric upland sands in which face, Q. marilandica is often abundant. As in the silt has collected. The increased silt ameliorates Atlantic Coastal Plain Ecoregion, vegetation the normally harsh soil chemistry and mois- types can be arranged latitudinally by dom- ture status. A typical indicator of this type is inant grasses. The southern types are domi- Astragalus michauxii (2.2.6). nated by Aristida beyrichiana. In this zone two types are recognizable, one of relatively dry Southern Coastal Plain sites with Baptisia perfoliata as an indicator (2.2.1), and one of more mesic sites with Nolina Subxeric pinelands of northern peninsular georgiana as an indicator (2.2.2). Northward in Florida and the eastern Panhandle typically the wiregrass gap region of South Carolina, support an understory of Quercus incana and Schizachyrium scoparium dominates and has as Q. margarettiae and a diverse ground layer 3. Ecological Classiﬁcation of Longleaf Pine Woodlands 75

FIGURE 5. Subxeric sandy uplands, Eastern Gulf Coastal Plain. Old-growth longleaf pine over dry flatwood with Serenoa repens and Quercus minima. Patterson Natural Area. Eglin Air Force Base, Okaloosa County, FL. dominated by Aristida beyrichiana mixed with phase is characterized by fewer oaks and more Schizachyrium scoparium, Andropogon gyrans, Schizachyrium scoparium var. stoloniferum and Sorghastrum secundum, and Sporobolus junceus Sorghastrum nutans (2.3.5). Small sand ridges (2.3.1). Westward in the Apalachicola region isolated in relatively mesic habitats and sur- this is replaced by subxeric woodlands with rounded by hammock vegetation can be very the understory dominated by Quercus laevis different with the sparse graminoid layer char- and with considerable Q. minima (2.3.2), likely acterized by Aristida condensata and Cyperus owing to lower soil phosphorus content in plukenetii (2.3.6). the Panhandle than in peninsular Florida (see Peet, Carr and Gamling 2006). Further westward on the well-drained low terraces of the Eastern Gulf Coastal Plain western Panhandle Quercus geminata domi- Subxeric sandy uplands are relatively un- nates over Serenoa repens and Aristida beyrichi- common on the Gulf Coastal Plain owing ana (2.3.3). Throughout the Southern Coastal to the generally fine-textured sediments Plain are low sandhills somewhat intermedi- encountered with increasing proximity to ate to flatwoods with Quercus geminata as a the Mississippi River. The subxeric sandhill subcanopy dominant. One phase in west cen- vegetation of the upper Panhandle (2.4.1) tral Florida and the adjacent Big Bend area continues west sporadically into the eastern is characterized by abundant Quercus chap- portion of the Gulf Coastal Plain to be replaced manii and other scrub oaks, Sorghastrum secun- in Mississippi and eastern Louisiana by a some- dum and Pterocaulon virgatum (2.3.4), whereas what depauperate version wherein Aristida a somewhat overlapping but more northern beyrichiana is replaced by Sporobolus clandestinus 76 II. Ecology as the dominant grass (2.4.2). A few exception- been abundant on such sites, but inventory ally sandy areas in southern Mississippi sup- data in early surveys (e.g., Hale 1883; Ashe port vegetation with a mixed scrub oak layer 1897) suggest longleaf woodlands to have been and a diversity of shrubs including Clinopodium the predominant type. The few remnants re- coccineum (2.4.3). Occasional sandy areas also maining (e.g., Fig. 6) suggest a ground layer occur on the inner coastal plain. On the with high herb diversity, particularly with re- Tifton Upland and Dougherty Plain, Aristida spect to legumes and composites. Through- beyrichiana dominates the groundlayer often out, longleaf pine is the canopy dominant, with Licania michauxii, being joined by typical and subcanopy oaks are generally unimpor- dry sandy site understory specialists such as tant. Pteridium aquilinum var. pseudocaudatum Quercus laevis, and Q. margarettiae (2.4.4). On and Schizachyrium scoparium are ubiquitous in the inner Coastal Plain west of the range of the ground layer. Silty uplands are essentially wiregrass and near the range limit of longleaf, absent from the Southern Coastal Plain ecore- Schizachyrium dominates the grass layer and gion. overall diversity tends to be low. Here Pinus palustris tends to share dominance with Pinus echinata and P. taeda (2.4.5). Atlantic Coastal Plain Longleaf vegetation of the upland silty sites Western Gulf Coastal Plain of the Atlantic slope can be conveniently sorted along two axes: a soil moisture gra- On uplands and high terraces over deep sand dient divided into mesic and subxeric, and a on the Western Gulf Coast, the subxeric lon- geographic axis. The four geographic regions gleaf woodlands, being outside the range of Q. include the range of southern wiregrass laevis and Aristida beyrichiana, have an under- (Aristida beyrichiana) south of the Santee River story generally dominated by Quercus incana system of South Carolina and across Georgia and a ground layer of Schizachyrium scoparium (3.1.1, 3.1.2), the wiregrass gap in central with Andropogon gerardii, A. ternarius, Panicum South Carolina dominated by Schizachyrium virgatum, and Sporobolus junceus. These sites scoparium (3.1.3, 3.1.4), the outer coastal support dryland species such as Croton argy- plain within the range of Carolina wiregrass ranthemus, Tragia spp., and Pityopsis graminifolia (Aristida stricta) starting in northern South (2.5.1). More xeric, fine-sandy stream terraces Carolina (3.1.5, 3.1.6), and the inner coastal are characterized by Liatris elegans and Opuntia plain within the range of Carolina wiregrass humifusa var. humifusa (2.5.2). A final subxeric (3.1.7, cf. 3.2.2). Although longleaf vegetation type is confined to sandy ridge tops in rolling on dry, silty uplands of the upper coastal plain landscapes where dominance is shared among is essentially gone, the few roadside scraps Pinus palustris, P. echinata, and P. taeda, with that persist suggest it to have been similar to hardwood species such as Carya texana and the silty, dry uplands of the Fall-line Sandhills. Quercus falcata scattered throughout (2.5.3). The understory contains occasional oaks (Quercus incana, Q. marilandica, Q. stellata), and the shrub layer generally contains Vaccinium Silty Uplands tenellum and Quercus pumila.

Longleaf vegetation on silty upland soils is Fall-line Sandhills scarce in the modern landscape as most such sites were converted to agriculture long be- Fine-textured soils are something of an fore 1900. Ultisol soils are abundant across the anomaly in the Fall-line Sandhills. Some areas Coastal Plain, hinting at the one-time domi- of silty soil are found associated with terraces nance of an ecosystem that has essentially van- of small streams where silts have been trans- ished. Some authors such as Phillips (1994) ported into the area by water and redistributed have suggested that hardwoods might have by wind. Perhaps the best-known examples 3. Ecological Classiﬁcation of Longleaf Pine Woodlands 77

FIGURE 6. Silty uplands, Atlantic Coastal Plain. This rare remnant of fire-maintained silty soils supports a mixture of oaks (Quercus marilandica, Q. stellata) and longleaf pine over a grass layer of Schizachyrium scoparium and Andropogon spp. Jasper County, SC. of silty soils occur in shallow depressions as Parthenium integrifolium and Silphium com- where fine-textured soils have either washed positum (3.2.2). On mesic sites shrubs like Ilex or blown in from the neighboring landscape glabra, grasses such as Panicum virgatum, and and become trapped. Such depressions are an abundance of taller forbs like Eupatorium relatively frequent within the range of Aris- rotundifolium give the vegetation a lush aspect tida stricta and are hypothesized to represent (3.2.3). A different variant of silty vegetation blow-outs similar to those that initially formed can be found near the transition to Piedmont the Carolina bays of the Coastal Plain (James where hardwoods like Carya pallida, Quercus 2000). The vegetation of these sandhill depres- stellata, and Cornus florida give the aspect of sions tends to be exceptionally species-rich and a transition from longleaf woodland to open their abundance of legumes has earned them hardwood woodland (3.2.5). the colloquial name among botanists of “bean dips” (James 2000). Vegetation of silty sandhill soils can be arranged along a moisture gradient Eastern Gulf Coastal Plain with the dry end characterized by species like The Eastern Gulf Coastal Plain is a region Toxicodendron pubescens, Tephrosia virginiana and of predominantly silty soils. Unfortunately, small trees like Quercus margarettiae, Q. mari- quantitative studies of this vegetation have landica, Q. incana, and Diospyros virginiana (3.2.1 been confined to a few studies of specific local- in the Carolinas, 3.2.4 in Georgia). Intermedi- ities (e.g., The Jones Center near Albany, GA, ate sites have greater dominance by legumes, a few plantations in the Tallahassee Red Hills), Pteridium, and forbs with prairie affinities such and no comprehensive, regionwide study has 78 II. Ecology

FIGURE 7. Silty uplands, Eastern Gulf Coastal Plain. Widely spaced old-growth longleaf pine over Aristida beyrichiana and Schizachyrium spp. Wade Tract. Thomas County, GA. yet been undertaken. As a consequence, the Andropogon gyrans, A. ternarius, and Danthonia current classification recognizes regional vari- sericea (3.4.4). West of the range of wiregrass, ants associated with different geomorphic sur- distinctive silt-soil longleaf types have been de- faces, but little more. Doubtless more study scribed from the coastal flatlands and rolling would allow greater resolution with respect to hills (3.4.5), the Mississippi loam hills (3.4.6), moisture regime. Within its geographic range, and the loess soils immediately east of the Aristida beyrichiana typically co-dominates with Mississippi River (3.4.7, 3.4.8). The higher fer- Schizachyrium scoparium and various species of tility of these sites leads to a greater abundance Andropogon. Variants of this vegetation have of hardwoods (e.g., Quercus marilandica, Q. in- been recognized associated with the Tallahas- cana, Q. margarettiae, Q. stellata, Carya glabra) see Red Hills, Mariana Lowlands, and Tifton and a particularly diverse forb layer. The lon- Plain of southwestern Georgia and adjacent gleaf vegetation of the loess soils of eastern Florida (3.4.1; Fig. 7), the Dougherty Plain Louisiana and adjacent Mississippi supports (3.4.2), and the coastal plain south and west the highest known 1000 m2 species richness of the Mariana Lowlands (3.4.3). All of these for upland habitats in the United States with types support tall grasses and a highly di- values reaching around 170 species (W. J. Platt verse assemblage of forb species. North of the personal communication). range of Aristida beyrichiana in western Georgia and the adjacent inner coastal plain of Al- abama (e.g., southern Ft. Benning, Tuskegee Western Gulf Coastal Plain and Talladega-Oakmulgee National Forests), Silty soils predominate on the Western Gulf Schizachyrium scoparium dominates along with Coastal Plain and support diverse longleaf 3. Ecological Classification of Longleaf Pine Woodlands 79 woodlands somewhat intermediate between deposits or where rocky hills and scarps asso- the longleaf woodlands of the Eastern Gulf ciated with ironstone caps punctuate the oth- Coastal Plain and the prairies of the east- erwise relatively gentle topography. Vegeta- ern Great Plains. Schizachyrium scoparium is tion intermediate to the more typical subxeric the ubiquitous dominant, in places mixed longleaf woodlands occur where a thin veneer with S. tenerum. Longleaf vegetation of the of sandy soil overlies the ironstones. These sites high Pleistocene terraces (3.5.1) is extremely support an abundance of such classic long- species rich, contains prairie taxa like Lia- leaf species as Vaccinium crassifolium and Aris- tris pycnostachya, and alternates with the sa- tida stricta, as well as species more strongly vanna vegetation of slightly lower sites (6.5.2). indicative of the clayey site conditions, such A somewhat wetter variant co-dominated by as Leiophyllum buxifolium, Pyxidanthera barbu- Schizachyrium scoparium and S. tenerum occupies lata, and Hypericum lloydii, (4.1.2; Fig. 8). In the the tops of pimple mounds on the lower Pleis- Altamaha Grit region of Georgia, rocky glade tocene terraces (3.5.2). More xeric silt-soil lon- vegetation can be found over similar isolated gleaf vegetation can be found on the dissected occurrences of indurated sandstones although topography of ridge tops farther inland, similar with such distinctive rockland species as Tal- floristically to the terrace type, but with such inum teretifolium, Allium cuthbertii, and Penste- drier-site prairie taxa as Rudbeckia grandiflora mon dissectus (4.1.3). var. alismiflora and Echinacea sanguinea (3.5.3). A rare vegetation type known only from Rapi- des Parish Louisiana is a glade type forming on Eastern Gulf Coastal Plain silty soils over clay or siltstone with interdigita- The innermost Coastal Plain of Alabama and tion of wet and dry microsites and a dominance adjacent Georgia contains significant areas of of Schizachyrium scoparium, S. tenerum, Muhlen- clay hills that in the original landscape sup- bergia expansa, and Bigelowia nuttallii (3.5.4). ported longleaf-dominated vegetation. There is little if any vegetation of this type left to study, though brief descriptions can be found Clayey and Rocky Uplands in early descriptive works (4.4.1; e.g., Mohr 1901; Harper 1943). This vegetation super- ficially appears to have resembled that of Although stereotypic longleaf vegetation oc- clay and rocky soils of the Atlantic Coastal curs on soft sediments of the southeastern Plain (4.1.1), particularly in the occurrence of Coastal Plain, longleaf pine occurs on other Kalmia latifolia, xeric oaks like Q. marilandica substrates such as ironstone and sandstone and Q. laevis, and the scarcity of herbs. South- hills of the inner Coastal Plain and Fall-line ern Mississippi and Alabama contain scattered Sandhills, as well as on clay soils derived from clay hills somewhat similar in character but the bedrock of the Piedmont or southwestern- lacking Kalmia and supporting a richer under- most Appalachian Mountains. story.

Atlantic Coastal Plain and Western Gulf Coastal Plain Fall-line Sandhills Outcroppings of calcareous sandstones (4.5.1) Scattered along the Fall-line Sandhills are and high-calcium, shrink-swell clay soils patches of longleaf woodland with a relatively (4.5.2, 4.5.3) on Tertiary terraces of the West- dense shrub layer of mountain laurel (Kalmia ern Gulf Coastal Plain support a typically latifolia) and other woody species reminiscent stunted woodland that contrasts with sur- of the Piedmont (e.g., Vaccinium arboreum), rounding Pinus palustris woodlands on deeper but with a sparse to absent herbaceous layer soils. Quercus marilandica and Q. stellata share (4.1.1). These communities can occur on ei- dominance on the shrink-swell clays here as ther eroded hills formed in marine kaolinite elsewhere on the Coastal Plain and Piedmont. 80 II. Ecology

FIGURE 8. Rocky uplands, Fall-line Sandhills. Soils are shallow sands over impermeable ironstones formed over a thick clay layer. Diversity is low with a canopy of Pinus palustris over a subcanopy of Quercus marilandica and a ground layer of Aristida stricta. Fort Bragg, Hoke County, NC.

On drier sites Andropogon gyrans, A. ternar- site conversion to agriculture. Consider that in ius, and Schizachyrium scoparium are the dom- 1860 Chatham County on the North Carolina inant grasses, whereas on moister clays the Piedmont ranked fifth among North Carolina dominant grasses are Panicum virgatum, Pan- counties in amount of standing longleaf timber icum anceps var. rhizomatum, Schizachyrium sco- (Hale 1883), whereas today fewer than a dozen parium var. divergens, and Sporobolus junceus. mature longleaf trees and perhaps only a sin- Also present are occasional herbaceous glades gle clump of wiregrass persist in the county. where sandstone is present at the surface Some indication of the original composition (4.5.4). These sandstone glades support pri- and diversity of this vegetation can be found in marily Schizachyrium scoparium and Bigelowia Mohr’s (1901) and Harper’s (1943) summaries nuttallii, though numerous forbs occur with of the forests of Alabama, and the treatments less abundance. of the Georgia highlands by Harper (1905) and Andrews (1917). Remaining montane and Piedmont longleaf Piedmont and Montane Uplands woodland vegetation types generally occur on Except for central Alabama, longleaf wood- exposed ridges and south-facing slopes. All lands were probably always relatively uncom- these sites support relatively consistent, un- mon in the Piedmont and mountain land- remarkable vegetation as the flora is charac- scapes. On these upland sites little remains of teristic of hardwood-dominated xeric, acidic the original longleaf vegetation owing to fire sites. Pinus echinata often occurs as a co- suppression combined with timber harvest and dominant along with such deciduous forest 3. Ecological Classification of Longleaf Pine Woodlands 81 taxa as Oxydendron arboreum, Quercus prinus, west of these boundaries recent marine sed- and Q. marilandica (4.6.1, 4.6.2). On the few iments are significantly more silty, and true remaining sites from the Uwharrie National Spodosols are uncommon to absent. Forest, NC, to Fort McClennan, AL, one finds widespread taxa such as Nyssa sylvatica, Pterid- ium aquilinum var. pseudocaudatum, Tephrosia Atlantic Coastal Plain virginiana, Pinus virginiana, Vaccinium tenellum, Flatwoods are largely a phenomenon of and Schizachyrium scoparium. On slightly less ex- the Southern Coastal Plain, which extends posed sites Quercus prinus and Pinus palustris can into southeastern South Carolina along the co-dominate (4.6.3) (Maceina et al. 2000). coastal fringe where it can be recognized by A few peculiar, local longleaf types occur dominance of such characteristic species as over unusual substrates. Burke Mountain on Serenoa repens, Vaccinium myrsinites, and Aristida the Georgia Piedmont supports the only lon- beyrichiana. True flatwoods are largely absent gleaf stand on serpentine soil (unusual taxa in- from the central coast of South Carolina where clude Clinopodium georgianum and Baptisia alba; soils tend to be too silty for Spodosol develop- 4.6.4). On the quartzite ridges of Pine Moun- ment, but do occur on the outer coastal plain tain, GA, is another distinctive type where be- from the Santee River of South Carolina north low the canopy of longleaf and shortleaf is a to the Neuse River in North Carolina. In this subcanopy dominated by Quercus coccinea and region flatwoods are of relatively low diversity Q. georgiana (4.6.5). and are often co-dominated by Pinus palustris and P. serotina with a ground layer dominated by Aristida stricta and Gaylussacia dumosa (5.1.1; Flatwoods Fig. 9). Somewhat wetter sites tend to be dominated by Ctenium aromaticum and Muh- Although “flatwoods” vegetation has been de- lenbergia expansa, and often have Pinus serotina scribed in all major treatments of the vege- in the canopy (5.1.2). Among taxa that dis- tation of the southeastern United States, the tinguish these northern flatwoods are Dionaea only long-term consistency in use of the term muscipula (Venus flytrap) and Vaccinium cras- is its application to fire-maintained pine wood- sifolium. Aspect dominants of the groundlayer land over flat, Coastal Plain landscapes. Some that distinguish unusual community variants of the confusion derives from the central place include the lily Pleea tenuifolia (disjunct to of flatwoods on the landscape. Flatwoods are the Apalachicola region of Florida; 5.1.3) and found between the more extreme vegetation the low shrub Leiophyllum buxifolium (better types of xeric sandhill, scrub, baygall, prairie, known from the New Jersey pine barrens and and savanna. When these physiognomically Blue Ridge rock outcrops; 5.1.4). Although and topographically more extreme types are longleaf woodlands are now largely extirpated given narrow definitions, as in Abrahamson north of the Neuse River, a few persistent sites and Hartnett (1990) and Harper (1914), flat- in northeastern North Carolina and south- woods become the broader concept referred to eastern Virginia have affinities with flatwoods by Christensen (2000) as “intractable.” vegetation and perhaps should be so classified I reserve use of the term “flatwoods” for (5.1.5), though their floristic composition is fire-maintained pinelands of low, flat terrain largely distinctive in the absence of species where marine sands were deposited during characteristic to the south, rather than species Pleistocene incursions and where the resultant occurrences (see Frost and Musselman 1987). soils are primarily poorly drained Spodosols, a convention also followed by Stout and Mar- Southern Coastal Plain ion (1993). Vegetation consistent with this definition is found along the Atlantic coast north Flatwoods vegetation is best developed in the into southern North Carolina, and west along Southern Coastal Plain ecoregion. The canopy the Gulf coast to eastern Mississippi. North and is generally dominated by Pinus palustris, but 82 II. Ecology

FIGURE 9. Flatwoods. Northern variant of longleaf flatwoods missing Serenoa repens, and Vaccinium myrsinites of the Southern Coastal Plain, but with Aristida stricta and Vaccinium crassifolium. Atlantic Coastal Plain. Croatan National Forest, Carteret County, NC. in much of Florida and southeast Georgia, P. Ctenium floridanum, in addition to Aristida elliottii var. elliottii replaces longleaf completely beyrichiana. Forbs characteristics of the East on the wettest sites (Clewell 1971; Gano 1917; include Symphyotrichum walteri, Carphephorus Monk 1968), and in the deep south of the paniculatus, Euthamia tenuifolia var. tenuifo- Florida peninsula Pinus elliottii var. densa can lia, and Eupatorium mohrii (5.3.1). In addi- dominate (e.g., 5.3.9). The understory layer tion to the typical Sporobolus curtissii flatwood is dominated throughout by such widespread type, there are types with Sporobolus pineto- taxa as Aristida beyrichiana, Vaccinium myrsinites, rum characteristic of the finer-texture soils of Serenoa repens, Pteridium aquilinum var. pseudo- northeast Georgia (5.3.2) and with Sporobo- caudatum, Quercus pumila, and Q. minima; their lus floridanus characteristic of the wettest sites co-dominance makes flatwoods vegetation rel- (5.3.3). In contrast, flatwoods of the east- atively easy to recognize. Other distinctive ern Panhandle have the widespread Vaccinium woody taxa of the southern flatwoods include myrsinites joined by its cousin V. darrowii, and Kalmia hirsuta, Lyonia fruticosa and Lyonia fer- commonly contain such herbaceous taxa as ruginea, and Gaylussacia tomentosa. Balduina uniflora, Carphephorus odoratissimus, Flatwoods vegetation shows substantial geo- Symphyotrichum adnatum, and Chrysopsis mari- graphic variation along a longitudinal gradient ana (5.3.4; Fig. 10), which elsewhere in the re- from northeastern Florida and adjacent Geor- gion are largely absent from flatwoods. Moister gia across to the Florida Panhandle. The east- sites are dominated by Pinus elliottii var. elliot- ern flatwoods are often co-dominated by such tii and sort according to soil chemistry with distinctive grasses as Sporobolus curtissii and near-coastal sites having high phosphorus and 3. Ecological Classification of Longleaf Pine Woodlands 83

FIGURE 10. Flatwoods. Southern Coastal Plain. Classic longleaf pine flatwoods with abundant Serenoa repens, Pteridium aquilinum var. pseudocaudatum, and Vaccinium myrsinites and V. darrowii. Apalachicola Na- tional Forest, Wakulla County, FL. calcium characterized by presence of Lyonia fer- type is that dominated by Pinus elliottii var. ruginea and Kalmia hirsuta (5.3.5), while the densa and Panicum abscissum, which appears in- more sterile and acidic sites have abundant Ly- termediate in composition between the typical onia lucida and Befaria racemosa (5.3.6). flatwoods and the dry prairies, but with an un- The soils of the drier scrubby flatwoods and derstory dominance of P. abscissum (5.3.9). In dry prairies of the central and more south- the extreme, we find sites lacking trees and ap- ern portions of peninsular Florida are some- pearing as a P. abscissum prairie (5.3.10). what sandier than those of the typical flat- Oaks, except for the running oaks (Quer- woods and the vegetation tends to be more cus pumila, Q. minima), are generally absent open. This vegetation supports only scattered from the true flatwoods. However, drier sites stems of Pinus palustris and P.elliottii var. elliottii, in central Florida and the Gulf Coast region of though several characteristic flatwood species the Southern Coastal Plain (5.3.11) contain a occur as dominants including Serenoa repens, mix of typical upland and “scrub oak” species Vaccinium myrsinites, and Aristida beyrichiana. as shrub and understory layer dominants, in- Quercus pumila is largely absent. Two of these cluding Quercus geminata, Q laevis, Q. chapmanii, types of open vegetation are commonly recog- Q. myrtifolia, and Q. inopina. Other low shrubs nized, one where Lyonia fruticosa and Quercus with high constancy include Ilex glabra, Quercus minima share dominance with Serenoa (5.3.7), minima, Gaylussacia dumosa, and Morella pumila. and one where there is a near prairie of Aristida In addition, two types transitional to flat- beyrichiana with Serenoa the only conspicuous woods, recognized from the barrier islands and woody species (5.3.8). An unusual vegetation extreme coastal fringe, contain abundant oak. 84 II. Ecology

Along the barrier islands of the Gulf coast one cent years usage of the term has generally con- finds typical flatwood vegetation but with an verged on open pine woodlands of season- abundance of oaks like Quercus chapmanii and ally saturated, fine-textured soils (e.g., Peet Q. geminata (5.3.12). On somewhat drier sites, and Allard 1993). The wet soil conditions of especially in northeastern Florida and adjacent savannas often lead to the tree canopy being Georgia where the much reduced fire regime very open with composition dominated by any allows development of a nearly impenetrable one of Pinus palustris, P. serotina, or P. elliottii understory thicket, evergreen shrubs like Quer- var. elliottii, or a combination of these. This cus chapmanii, Q. myrtifolia, Q. geminata, and vegetation is generally rich in species across a Lyonia ferruginea form a scrublike variant of broad range of spatial scales (10−3–104 m2) and flatwood (5.3.13). characterized by an abundance of showy forbs. For example, such sites often contain a wealth of orchids (e.g., Calopogon, Cleistes, Platanthera, Eastern Gulf Coastal Plain Pogonia, Spiranthes), insectivorous plants (e.g., West of Florida, sandy soils are largely replaced Drosera, Dionaea, Pinguicula, Sarracenia, Utricu- by silts, and longleaf pine flatwoods are gener- laria), and lilies (e.g., Aletris, Lilium, Tofieldia, ally confined to a narrow coastal fringe. One Zigadenus), to say nothing of numerous grasses, distinctive form of flatwoods of the Eastern sedges, and composites. Legumes are con- Gulf Coastal Plain and adjacent coastal fringe spicuously scarce in moist savannas, a phe- of the Florida Panhandle is that of barrier is- nomenon noted by Gano (1917), Wells and land interdunal swales and flats strongly domi- Shunk (1931), and Taggart (1990, 1994). The nated by Serenoa repens and with an abundance floristic novelty and diversity of pine savannas of such maritime shrubs as Morella cerifera and have led to this vegetation being perhaps the Ilex vomitoria (5.4.1; cf. Huffman et al. 2004). best known of the original longleaf commu- Somewhat higher and drier barrier island and nity types (e.g., Kologiski 1977; Folkerts 1982; maritime fringe sites support a distinctive un- Walker and Peet 1983; Norquist 1984; Taggart derstory of oak including Quercus virginiana, 1994). Q. geminata, and Q. hemisphaerica, along with The pine savannas exhibit significant varia- Magnolia grandiflora (5.4.2). The isolation of tion in composition driven by subtle changes in these islands keeps fire frequency low and al- hydrologic regime, soil texture, and soil chem- lows shrub density to be high. On the main- istry, more so than any other vegetation type land, flatwoods from the Apalachicola to the of the fire-maintained southeastern pinelands Pascagoula River tend to be attenuated ver- (see Fig. 11). In addition to environmentally sions of those of the eastern Panhandle (e.g., driven, local variation, geographic turnover is 5.3.5). West of the Pascagoula River soils are pronounced, and the insular distribution of silty and flatwoods are mostly absent. The clos- wet pinelands has led to chance migration est approximation in Mississippi is found in events generating significant stochastic varia- sandy wet longleaf areas of the De Soto Na- tion in composition among sites. As a conse- tional Forest (5.4.3), where this western ex- quence, despite the diversity of savanna types, treme of flatwoods has few characteristic flat- the range of variation in each recognized type woods taxa other than Serenoa repens and Ilex can be significantly greater than in other veg- glabra. etation types recognized in this chapter. On the outer Coastal Plain where the landscape is extremely flat, individual savannas Savannas, Seeps, and can cover considerable area. Moving inland, Prairies savanna areas are constrained to smaller areas, often narrow swales within the gen- The term “savanna” has many meanings, even tly rolling topography. In the inner Coastal within the context of the pinelands of the Plain, Fall-line Sandhill, and Piedmont regions southeastern United States. However, in re- where the landscape is far less flat and much 3. Ecological Classification of Longleaf Pine Woodlands 85

FIGURE 11. Subxeric sandy upland-to-savanna transition. Southern Coastal Plain. Sparse Pinus palustris over Aristida beyrichiana and Schizachyrium scoparium grading down slope into moist savanna. Apalachicola National Forest, Liberty County, FL. more topographically complex, savannalike and the local endemic Sporobolus pinetorum vegetation is confined to small seepage areas (6.1.1; Fig. 12). On somewhat wetter sites Aris- associated with impermeable clays or near- tida stricta drops out to be replaced by greater surface bedrock. dominance of Ctenium aromaticum and Muhlen- bergia expansa, though with an admixture of various Rhynchospora, Andropogon, and other Atlantic Coastal Plain graminoids (6.1.2). Numerous orchids (often Savannas are best developed on the out- as many as a half dozen species in a 1000 m2 ermost Coastal Plain where extensive areas plot) occur within these vegetation types, and of flat, fine-textured soils occur, represent- insectivorous plants are especially well devel- ing terraces in regions of predominantly fine- oped on the wetter sites (e.g., 6.1.2) where it textured sediments or one-time embayments is not uncommon for a single site to have two behind now-vanished barrier island systems. or three species of Sarracenia, two or three Pin- Although these savannas are generally dom- guicula, two Drosera, and Dionaea. inated by grasses, the dominant species shift Dominance shifts as one moves south to- significantly with soil moisture and texture. ward the more loamy landscapes of the South In those southeastern North Carolina sa- Carolina low country. Aristida stricta drops out vannas on seasonally saturated soils that are near the Pee Dee River system in northeast sufficiently well-drained that surface water is South Carolina to be replaced by stronger dom- short-lasting, silty sites are dominated by the inance of Schizachyrium scoparium, whereas grasses Aristida stricta, Schizachyrium scoparium, Sporobolus pinetorum penetrates southward to 86 II. Ecology

FIGURE 12. Savanna. Atlantic Coastal Plain. Species-rich longleaf savanna in an area with an average richness of 35 species/m2. Green Swamp Preserve, Brunswick County, NC. the latitude of Charleston, where it is limited to endemic Sporobolus, S. teretifolius, tends to dom- just a few sites (appearing again on a few sites inate (6.1.6). Inland from the coastal flatlands, in northeast Georgia). In contrast, Sporobolus savannas are mostly confined to gentle seep- curtissii, a grass primarily dominating flatwoods age slopes influenced by groundwater (6.1.7). of the Southern Coastal Plain where it is abun- Where soils are more fertile, composition shifts dant northward to Savannah, is absent from toward greater abundance of Arundinaria and the Atlantic Coastal Plain region except for Andropogon with Liquidambar often present in clayey savannas near Charleston (6.1.3). Very the understory (6.1.8). On the more extreme wet savannas of the loamy South Carolina low examples in the original landscape, such sites country sometime have the grass layer domi- supported canebrakes with nearly complete nated by Aristida palustris (6.1.4). dominance by Arundinaria, though such vege- Subtle changes in soil chemistry can have tation has all been lost to either agriculture or a significant impact on savanna composition. fire suppression. Where soils are wet but somewhat influenced by a marl substrate, composition shifts toward stronger dominance by a combination of Fall-line Sandhills Sporobolus pinetorum and Ctenium (6.1.5). These The Fall-line Sandhills, like the inner coastal sites have a suite of narrowly distributed sa- plain rolling hills, generally do not have the ex- vanna calciphiles such as Carex lutea, Thalic- tensive flatlands with impeded drainage nec- trum cooleyi, and Parnassia caroliniana. On those essary to support true savanna. However, im- rare sites with extremely wet marl-clay soils, permeable clay layers and indurated spodic species richness drops and another regionally horizons occur in these regions and, where 3. Ecological Classification of Longleaf Pine Woodlands 87

FIGURE 13. Seep. Fall-line Sandhills. Species-rich seep dominated by Ctenium aromaticum, Muhlenbergia expansa, and Andropogon spp. with abundant Sarracenia lutea. Fort Bragg, Hope County, NC. they approach the surface, seepage sometimes region. Here can be found a diverse array occurs. These seeps are usually similar to true of savannalike vegetation, again with com- coastal plain savannas in their species compo- positional sorting by moisture levels. At the sition (see Wells and Shunk 1931), but lower drier end of the gradient, mesic pineland in diversity (6.2.1, 6.2.2; Fig. 13). taxa such as Helianthus radula and Galactia erecta occur (6.3.3). Where the soils become chronically wet Rhynchospora diversity greatly Southern Coastal Plain increases and Verbesina chapmanii is often im- The Southern Coastal Plain is largely a land portant (6.3.4). On the wettest sites Eriocaulon of sandy soils with the fine-textured soils decangulare, Lachnanthes caroliana, Pleea tenuifo- characteristic of savannas occurring on either lia, Sarracenia psittacina, and S. flava are conspic- side. In northeast Georgia and southeastern uous (6.3.5). In peninsular Florida savannalike South Carolina the wet sites on fine-textured vegetation can occur despite sandy soils un- soils are dominated by Sporobolus floridanus, der special circumstances. Particularly wet soils often with a scattering of Morella cerifera and can be dominated by grasses such as Panicum the distinctive dwarf shrub Styrax americanus rigidulum, Panicum hemitomon, and Andropogon var. pulverulentus (6.3.1; Fig. 14), while on glaucopsis (6.3.6). Where the soil is particularly slightly drier sites Aristida beyrichiana shares high in calcium so that the pH is about 6, a veg- dominance with Ctenium aromaticum (6.3.2). To etation type sometimes referred to as “sweet reach the next area of extensive savanna it is flats” forms with Sabal palmetto and Morella necessary to jump over the Florida peninsula cerifera sharing dominance with Pinus elliottii to treeless flats of silty soils in the Apalachicola var. elliottii (6.3.7). 88 II. Ecology

FIGURE 14. Savanna, Southern Coastal Plain. Savanna with saturated soils that limit the growth of trees to a few scattered longleaf and slash pines. Sporobolus ﬂoridanus dominates a ground layer punctuated by scattered Morella cerifera and Styrax americanus var. pulverulentus. Jasper County, SC.

Wet prairie vegetation, largely devoid of Southern Coastal Plain owing to the predom- trees is relatively frequent on and largely inantly siltier soils of the ecoregion. Vege- unique to the Florida peninsula and adjacent tation in these savannas bears considerable eastern Panhandle. Soils of these wet grass- floristic similarity to the savannas of the At- lands are too sandy to match the definition lantic Coastal Plain as seen in southeastern of savanna I use here, consistently contain- North Carolina. Wet savannas similar in char- ing over 95% sand in both the A and B hori- acter to 6.1.2 occupy particularly wet sites. zons. Considerable floristic variation occurs East of Mobile Bay these sites are still domi- in these types with variation in soil mois- nated by Aristida beyrichiana but with consid- ture regime, soil chemistry, and geographic erable Ctenium and Muhlenbergia expansa, and position. Nonetheless, Aristida beyrichiana, A. the hillside seepage areas of the more rolling palustris, Amphicarpum muehlenbergianum, and higher terraces of the Florida Panhandle are Sporobolus floridanus are relatively widespread largely dominated by Aristida beyrichiana with dominants suggesting close affinities with lon- co-dominance of Andropogon arctatus, Ctenium gleaf pine savannas (6.3.8–6.3.10). aromaticum, and Dichanthelium leucothrix (6.4.1; Fig. 15), whereas west of Mobile Bay Aris- tida beyrichiana is spotty, dropping out com- Eastern Gulf Coastal Plain pletely at the Pascagoula River. In place of True savanna vegetation is better developed Aristida beyrichiana one see increased domi- on the Eastern Gulf Coastal Plain and adja- nance of Muhlenbergia expansa and various An- cent Panhandle than the main body of the dropogon (e.g., mohrii, gyrans var. stenophyllus) 3. Ecological Classification of Longleaf Pine Woodlands 89

FIGURE 15. Savanna. Gulf Coastal Plain. Moist coastal savannas can support a wealth of orchids and insectivorous plants. Abundant pitcher plants (Sarracenia alata) and sundews (Drosera tracyi) are visible in the foreground. Sand Hill Crane National Wildlife Refuge, Jackson County, MS.

(6.4.2). Somewhat drier savannas (similar to flected in the increased abundance of Arun- 6.1.1) also occur in southwestern Mississippi dinaria (6.4.6). Seeps intermediate between and southeastern Louisiana, these dominated those of the Florida Panhandle and the outer by Schizachyrium scoparium and Muhlenbergia ex- coastal plain of Mississippi can be found in pansa (6.4.3). the vicinity of the De Soto National Forest. Higher on the Gulf Coastal Plain, above These sites are dominated by grasses such as the coastal flatlands, true savannas are mostly Muhlenbergia expansa, Schizachyrium scoparium, absent but one finds similar species occur- S. tenerum, and Anthaenantia rufa (6.4.7). ring on seepage slopes. The seeps of the silty Tallahassee red hills are particularly striking with the somewhat drier sites dominated by Western Gulf Coastal Plain Quercus pumila and Rhododendron canescens with Wet savanna vegetation was at one time Ctenium aromaticum (6.4.4), and the very wet widely distributed on the Western Gulf Coastal sites dominated by unusual Dichantheliums and Plain. In contrast to savannas east of the Missis- Panicum verrucosum (6.4.5). These latter sites sippi bottomlands, these savannas occupy soils have species richness values that are among that are relatively calcareous and often con- the highest found in longleaf communities tain a significant amount of shrink-swell clays. with as many as 170 species per 1000 m2. Nonetheless, there are strong floristic affinities The overall silt dominance of this landscape between this savanna vegetation and the more can lead to somewhat more fertile soils, re- eastern savannas of fine-textured soils as far 90 II. Ecology away as the Atlantic Coastal Plain (e.g., 6.1.2, flora of the fire-maintained pinelands. Much 6.1.7). These savannas can usefully be divided fieldwork remains before we can claim to have into the lower to middle terraces with soils pre- carefully documented the compositional vari- dominantly Glossaqualfs (6.5.1) and the some- ation of the remaining longleaf pine vegeta- what higher and better-drained savannas of tion. However, drawing on work embedded the upper terraces where the soils are pri- in the U.S. National Vegetation Classification, marily Paleudults (6.5.2). Both savanna types my own preliminary analysis of approximately are dominated by Ctenium (east of the Sabine 900 vegetation plots scattered over the east- River), Muhlenbergia expansa, Schizachyrium sco- ern two-thirds of the range of longleaf, and the parium, and multiple species of Rhynchospora. work of many other authors, I here tentatively Among the rolling hills of the older land sur- accepted 135 longleaf vegetation associations. faces of the West Gulf Coast, seepage bogs with Although this may seem like a high number, Sarracenia alata and Rhynchospora gracilenta oc- I expect this number to increase substantially cur as small patches, though rather few are with increased collection and analysis of plot left for study (6.5.3; see Bridges and Orzell data. This represents what is likely the mini- 1989). A rare savanna type is sometimes found mum number of units for classifying longleaf on saline Pleistocene terraces where beneath vegetation for conservation and for providing the longleaf pine the dominant grasses are targets for ecological restoration. Sporobolus silveanus and Muhlenbergia capillaris. The remarkable diversity of the greater The flora is relatively rich, but the taxa have longleaf ecosystem is being lost rapidly, stronger affinities with the Midwestern prairies both through active habitat destruction and than the Eastern Gulf Coastal Plain (6.5.4). through neglect. If even a fraction of the diversity of the longleaf ecosystem is to be pre- Piedmont and Montane Uplands served, action must be taken quickly to both protect and manage the best remaining ex- Piedmont and montane longleaf sites are gen- amples of each of the longleaf community erally well drained and lacking in seepage veg- types. Of particular importance in any such etation. A few exceptions can be found in endeavor is recognizing the considerable vari- the Uwharrie National Forest of North Car- ation in longleaf vegetation that occurs with olina where small depressions and flat, seepy simple geographic distance and subtle environ- small-stream flats occur. On these sites the mental variation. Conservation and preserva- mixed pine canopy has an groundlayer domi- tion of the longleaf ecosystem cannot simply nated by the grasses Chasmanthium laxum and focus on a small number of high-quality pre- Panicum virgatum. Shrubs are well developed serves, as these will inevitably capture only a in these sites, including such typically coastal modest fraction of the natural variation. We plain taxa as Lyonia mariana, Gaylussacia fron- need to devise a reserve system that includes dosa, Vaccinium fuscatum, and Ilex glabra (6.6.1). preserves and restored sites that span the geo- Additional variants should be expected in east- graphic and environmental range of the origi- ern Alabama. nal longleaf ecosystem.

Concluding Remarks Acknowledgments

Although the once extensive southeastern lon- Fieldwork on which this chapter was based gleaf pine woodlands may appear to the casual was supported by grants from the U.S. Forest observer as a homogeneous expanse of lon- Service and the Florida Fish and Wildlife Con- gleaf pine, grass, and scrub oak, this is a gross servation Commission. This work would not oversimplification. The Southeastern Coastal have been possible without the extensive work Plain is exceptionally rich in endemic species, on the National Vegetation Classification pro- and much of this endemism is manifest in the vided by the employees of NatureServe, The 3. Ecological Classification of Longleaf Pine Woodlands 91

Nature Conservancy, and the various state Nat- Vegetation, Second Edition, eds. M.G. Barbour and ural Heritage programs of the region. Portions W. D. Billings, pp. 397–448. London: Cambridge of this chapter are based on collaborations with University Press. other individuals, including in particular Susan Clewell, A.F. 1971. The vegetation of the Carr, Richard Duncan, Eric Kjellmark, Patrick Apalachicola National Forest: An ecological McMillan, Michael Schafale, Alan Weakley, perspective. Contract 38–2249, Final Report. US Forest Service, Atlanta, GA. and Tom Wentworth. To all of these people Daniels, R.B., Kleiss, H.J., Buol, S.W., Byrd, H.J., and programs, and many others, I am deeply and Phillips, J.A. 1984. Soil systems in North indebted. Carolina. North Carolina Agricultural Research Service, Bulletin 467, Raleigh. Endnote DuBar, J.R., Johnson, H.S., Jr., Thom, B.G., and Hatchell, W.O. 1974. Neogene stratigra- 1. Botanical nomenclature follows USDA, NRCS phy and morphology, south flank of the Cape (2005); except for Sporobolus and Muhlenbergia, Fear arch, North and South Carolina. In Post- which follow Peterson et al. (2003) and Peter- Miocene Stratigraphy, Central and Southern Atlantic son (2003), and Styrax, which follows Gonsoulin Coastal Plain, eds. R.Q. Oaks, Jr., and J.R. (1974); common names are shown in Table 3. DuBar, pp. 139–173. Logan: Utah State University Press. References Duncan, R.P., Peet, R.K., Wentworth, T.R., Schafale, M.P., and Weakley, A.S. 1994. Vegetation of the Abrahamson, W.G., and Hartnett, D.C. 1990. Pine fire-dependent pinelands of the Fall-line Sand- flatwoods and dry prairies. In Ecosystems of Florida, hills of Southeastern North America. Final Re- eds. R.L. Myers and J.J. Ewel, pp. 103–149. port. USDA Forest Service, Southern Forest Ex- Orlando: The University of Central Florida Press. periment Station. Anderson, M., Bourgeron, P., Beyer, M.T., Craw- Folkerts, G.W. 1982. The Gulf Coast pitcher plant ford, R., Engelking, L., Faber-Langendoen, D., bogs. Am Sci 70:260–267. Gallyoun, M., Goodin, K., Grossman, D.H., Lan- Frost, C.C. 1993. Four centuries of changing land- daal, S., Metzler, K., Patterson, K.D., Pyne, M., scape patterns in the longleaf pine ecosystem. Proc Reid, M., Sneddon, L., and Weakley, A.S. 1998. Tall Timbers Fire Ecol Conf 18:17–44. International Classification of Ecological Communities: Frost, C.C. 2000. Studies in landscape fire ecology Terrestrial Vegetation of the United States. Volume II. and presettlement vegetation of the southeast- The National Vegetation Classification System: List of ern United States. Ph.D. dissertation, University Types. Arlington, VA: The Nature Conservancy. of North Carolina, Chapel Hill. Andrews, E.F. 1917. Agency of fire in propagation Frost, C.C., and Musselman, L.J. 1987. History and of longleaf pines. Bot Gaz 64:497–508. vegetation of the Blackwater Ecological Preserve. Ashe, W.W. 1897. Forests of North Carolina. NC Geol Castanea 52:16–46. Surv Bull 6:139–224. Frost, C.C., Walker, J., and Peet, R.K. 1986. Fire- Bailey, R.G. 1995. Description of the Ecoregions of the dependent savannas and prairies of the South- United States: Second Edition. Miscellaneous Publi- east: Original extent, preservation status and cations No. 1391. Washington, DC: USDA Forest management problems. In Wilderness and Natural Service. Areas in the Eastern United States, eds. D.L. Kulhavy Bozeman, J.R. 1971. A sociologic and geographic and R.N. Conner, pp. 348–357. Nacogdoches, TX: study of the sand ridge vegetation of the coastal Center for Applied Studies. plain of Georgia. Ph.D. dissertation, University of Gano, L. 1917. A study of physiographic ecology in North Carolina, Chapel Hill. northern Florida. Bot Gaz 63:337–372. Bridges, E.L., and Orzell, S.L. 1989. Longleaf pine Gonsoulin, G.J. 1974. A revision of Styrax (Styra- communities of the West Gulf Coastal Plain. Nat caceae) in North America, Central America and Areas J 9:246–263. the Caribbean. Sida 5:191–258. Brown, D., Reichenbacher, F., and Franson, S. 1998. Hale, P.M. 1883. The woods and timbers of North A Classification of North American Biotic Communities. Carolina. P.M. Hale Publisher, Raleigh (including Salt Lake City: University of Utah Press. a reprint of M.A. Curtis. 1860. Geological and Christensen, N.L. 2000. Vegetation of the South- Natural History Survey of North Carolina. Part eastern Coastal Plain. In North American Terrestrial III—Botany). 92 II. Ecology

Hainds, M.J., Mitchell, R.J., Palik, B.J., Boring, L.R., Maceina, E.C., Kush, J.S., and Meldahl, R.S. 2000. and Gjerstad, D.H. 1999. Distribution of native Vegetational survey of a montane longleaf pine legumes (Leguminoseae) in frequently burned community at Fort McClellan, Alabama. Castanea longleaf pine (Pinaceae)–wiregrass (Poaceae) 65:147–154. ecosystems. Am J Bot 86:1606–1614. McCune, B., and Grace, J.B. 2003. Analysis of Ecolog- Harcombe, P.A., Glitzenstein, J.S., Knox, R.G., ical Communities. Gleneden Beach, OR: MjM Soft- Orzell, S.L., and Bridges, E.L. 1993. Vegetation of ware Design. the longleaf pine region of the West Gulf Coast. Mohr, C. 1901. Plant life of Alabama. Contrib US Natl Proc Tall Timbers Fire Ecol Conf 18:83–104. Herb 6. Harper, R.M. 1905. Some noteworthy stations for Monk, C.D. 1968. Successional and environmen- Pinus palustris. Torreya 5:55–60. tal relationships of the forest vegetation of north- Harper, R.M. 1906. A phytogeographical sketch of central Florida. Am Midl Nat 79:441–457. the Altamaha Grit Region of the coastal plain of Myers, N., Mittermeier, R.A., Mittermeier, C.G., Georgia. NY Acad Sci 7:1–415. Da Fosseca, G.A.B., and Kent, J. 2000. Biodiver- Harper, R.M. 1914. Geology and vegetation of North sity hotspots for conservation priorities. Nature Florida. Florida Geological Survey. Sixth Annual 403:853–858. Report 163–451. Myers, R.K., Zahner, R., and Jones, S.M. 1986. For- Harper, R.M. 1930. The natural resources of Geor- est habitat regions of South Carolina from LAND- gia. School of Commerce, Bureau of Business Re- SAT imagery. Clemson University, Department of search, Study No. 2. Bulletin of the University of Forestry Research Series No. 42. Georgia 30(3):105. NatureServe. 2005. NatureServe Explorer version Harper, R.M. 1943. Forests of Alabama. Alabama 4.4. 7 April 2005. (http://www.natureserve.org/ Geological Survey Monograph 10. Alabama Geo- explorer/). logical Survey, University of Alabama. University, Norquist, H.C. 1984. A comparative study of the AL. soils and vegetation of savannas in Mississippi. Hodgkins, E.J. 1965. Southeastern forest habitat re- Masters thesis, Mississippi State University, Mis- gions based on physiography. Agricultural Exper- sissippi State. iment Station, Auburn University, Forestry De- Omernik, J.M. 1987. Ecoregions of the contermi- partment Series, No. 2. Auburn, AL. nous United States. Map (scale 1:7,500,000). Ann Hodgkins, E.J., Golden, M.S., and Miller, W.F. Assoc Am Geogr 77(1):118–125 (for updates con- 1979. Forest habitat regions and types on a sult http://www.epa.gov/wed/pages/ecoregions/ photomorphic-physiographic basis: A guide to ecoregions.htm). forest site classification in Alabama–Mississippi. Peet, R.K. 1993. A taxonomic study of Aristida stricta Southern Coop Series 210, Alabama Agricultural and A. beyrichiana. Rhodora 95:25–37. Experiment Station, Auburn. Peet, R.K., and Allard, D.J. 1993. Longleaf pine veg- Huffman, J.M., Platt, W.J., Grissino-Mayer, H., and etation of the Southern Atlantic and Eastern Gulf Boyce, C.J. 2004. Fire history of a barrier island Coast regions: A preliminary classification. Proc slash pine (Pinus elliottii) savanna. Nat Areas J Tall Timbers Fire Ecol Conf 18:45–81. 24:258–268. Peet, R.K., Carr, S. and Gramling, J. 2006. Fire- James, M.M. 2000. Legumes in loamy soil commu- adapted pineland vegetation of northern and cen- nities of the Carolina sandhills: Their natural dis- tral Florida: A framework for inventory, manage- tributions and performance of seeds and seedlings ment, and restoration. Florida Fish and Wildlife along complex ecological gradients. Master’s the- Commission. In press. sis, Curriculum in Ecology, University of North Peet, R.K., Duncan, R.P., Wentworth, T.R., Schafale, Carolina, Chapel Hill. M.P., and Weakley, A.S. 1994. Vegetation of the Kjellmark, E.W., McMillan, P.D., and Peet, R.K. fire-dependent pinelands of the North Carolina 1998. Longleaf pine vegetation of the South Car- Coastal Plain. Final Report. USDA Forest Service, olina and Georgia Coastal Plain: A preliminary Southern Forest Experiment Station. classification. Final report, USDA Forest Service, Peet, R.K., Wentworth, T.R., and White, P.S. 1998. Southern Forest Experiment Station. A flexible, multipurpose method for recording Kologiski, R.L. 1977. The phytosociology of the vegetation composition and structure. Castanea Green Swamp, North Carolina. North Carolina 63:262–274. Agricultural Experiment Station, Technical Bul- Peterson, P.M. 2003. Muhlenbergia Schreb. In letin 250. Raleigh. Flora of North America Volume 25: Magnoliophyta: 3. Ecological Classification of Longleaf Pine Woodlands 93

Commelinidae (in part): Poaceae, part 2, eds. M.E. Sorrie, B.A., and Weakley, A.S. 2006. Developing a Barkworth, K.M. Capels, S. Long, and M.B. blueprint for conservation of the endangered lon- Piep, pp. 145–200. New York: Oxford University gleaf pine ecosystem based on centers of Coastal Press. Plain plant endemism. Applied Vegetation Science Peterson, P.M., Hatch, S.L., and Weakley, A.S. 2003. 9(1) (in press). Sporobolus R. Br. In Flora of North America Volume Stout, I.J., and Marion, W.R.1993. Pine flatwoods 25: Magnoliophyta: Commelinidae (in part): Poaceae, and xeric pine forests of the Southern (lower) Part 2, eds. M.E. Barkworth, K.M. Capels, S. Long, Coastal Plain. In Biodiversity of the Southeastern and M.B. Piep, pp. 115–139. New York: Oxford United States: Lowland Terrestrial Communities, eds. University Press. W.H. Martin, S.G. Boyce, and A.C. Echternacht, Phillips, J.D. 1994. Forgotten hardwood forests of pp. 373–446. New York: John Wiley & Sons. the Coastal-plain. Geogr Rev 84:162–171. Taggart, J.B. 1990. Inventory, classification, and Platt, W.J. 1999. Southeastern pine savannas. In Sa- preservation of coastal plain savannas in the vannas, Barrens, and Rock Outcrop Plant Communities Carolinas. Ph.D. dissertation, University of North of North America, eds. R.C. Anderson, J.S. Fralish, Carolina, Chapel Hill. and J.M. Baskin, pp. 21–51. London: Cambridge Taggart, J.B. 1994. Ordination as an aid in determin- University Press. ing priorities for plant community protection. Biol Ricketts, T.H., Dinertstein, E., Loucks, C.J., Eich- Conserv 68:135–141. buam, W., DellaSella, D., Kavanagh, K., Hedao, P., USDA, NRCS. 2005. The PLANTS Database, Version Hurley, P.T., Carney, K.M., Abell, R., and Walters, 3.5 (http://plants.usda.gov). National Plant Data S. 1999. Terrestrial Ecoregions of North America: Center, Baton Rouge, LA 70874–4490 USA. A Conservation Assessment. World Wildlife Fund, Walker, J., and Peet, R.K. 1984. Composition United States and Canada. Washington, DC: Island and species diversity of pine–wiregrass savannas Press. of the Green Swamp, North Carolina. Vegetatio Rodgers, H.L., and Provencher, L. 1999. Analysis 55:163–179. of longleaf pine sandhill vegetation in northwest Ware, S., Frost, C., and Doerr, P.D. 1993. South- Florida. Castanea 64:138–162. ern mixed hardwood forest: The former longleaf Soller, D.R., and Mills, H.H. 1991. Surficial ge- pine forest. In Biodiversity of the Southeastern United ology and geomorphology. In The Geology of States: Lowland Terrestrial Communities, eds. W.H. the Carolinas, eds. J.W. Horton and V.A. Zullo, Martin, S.G. Boyce, and A.C. Echternacht, pp. pp. 290–308. Knoxville: University of Tennessee 447–493. New York: John Wiley & Sons. Press. Wells, B.W., and Shunk, I.V. 1931. The vegetation Sorrie, B.A., and Weakley, A.S. 2001. Coastal and habitat factors of the coarser sands of the Plain vascular plant endemics: Phytogeographic North Carolina coastal plain: An ecological study. patterns. Castanea 66:50–82. Ecol Monogr 1:466–520. Chapter 4 Longleaf Pine Regeneration Ecology and Methods

Dale G. Brockway, Kenneth W. Outcalt, and William D. Boyer

Introduction lighting recent work that calls for application of silviculture techniques that more closely Regenerating longleaf pine (Pinus palustris)is mimic natural disturbance regimes. key to its long-term sustainable production of forest resources and its perpetuation as the dominant tree species in a variety of important Ecological Relationships ecosystems ranging from xeric to mesic to hydric site conditions. Early regeneration prob- Indispensable Nature of lems and the subsequent efforts to overcome Regeneration these are significant features of the continuing longleaf pine saga. This chapter discusses Successful reproduction is essential to perpet- recent restoration relevant to longleaf pine uate any population of organisms. Indeed, if an regeneration, disturbance dynamics including existing generation is unable to produce a suc- fire as an ecological process and describes ceeding generation, then the existing genera- the uniqueness of longleaf pine’s regeneration tion can appropriately be considered an eco- environment. Fundamental information con- logical and evolutionary “dead end” for that cerning reproductive biology (including genet- population and perhaps the entire species. De- ics, flowering, pollination, fertilization, cone spite the extended longevity of many tree production, and seed dispersal) and seedling species, some approaching 500 years, all in- development (including germination, shoot dividual organisms eventually die. If none of growth, rooting, sprouting, competition, ini- the offspring survive to the age of reproductive tiation of height growth, effects of fire, and maturity, then the entire species will eventu- seedling morality) is then presented. Various ally perish. If the species is a dominant organ- aspects of natural regeneration and artifi- ism, then entire ecosystems will be degraded cial regeneration are discussed and the even- or lost, potentially threatening the survival of aged (i.e., clearcutting, seed-tree and shelter- associated plant and animal species. wood) and uneven-aged (i.e., group selection While species extinction and ecosystem loss and single-tree selection) forest reproduction may seem like rare events in the shorter term methods are introduced. We conclude by high- of human experience, from the longer-term

r Dale G. Brockway and William D. Boyer Southern Research Station, USDA Forest Service, Auburn, Alabama r 36849. Kenneth W. Outcalt Southern Research Station, USDA Forest Service, Athens, Georgia 30602.

95 96 II. Ecology

FIGURE 1. Naturally regenerated even-aged second-growth longleaf pine forest on mesic uplands. Photo courtesy of the Forestry Images Organization. perspective of geologic time, such events have the eighteenth and nineteenth centuries was not been uncommon. The survival of species generally modest, as populations grew and log- and sustainability of ecosystems cannot al- ging activity expanded during the late nine- ways be assumed, even under the best cir- teenth and early twentieth centuries, most na- cumstances. And when new species and/or tive longleaf pine forests were harvested. The cultures encounter native ecosystems, new land was often converted to agricultural, resi- pressures can stress the indigenous organisms dential and urban uses or planted with planta- and threaten ecological sustainability. Such has tions of other easier-to-establish, faster grow- been the case in the southern United States, ing trees such as slash pine (Pinus elliottii) where the Age of Discovery and the Industrial and loblolly pine (Pinus taeda) (Croker 1987, Revolution brought substantial change to na- Outcalt 2000). Although many second-growth tive longleaf pine forests (Frost this volume). longleaf pine forests naturally regenerated fol- Effective means of regenerating longleaf pine lowing this initial harvest (Fig. 1), recovery are important for continuation of this species was impaired by irregular seed production, and the long-term sustainability of longleaf with good seed years occurring at intervals pine forest ecosystems. of five or more years (Boyer 1990a). Where longleaf pine seedlings did survive logging, they were often consumed by feral hogs (Sus Declining Trend scrofa), causing many areas of potential lon- Longleaf pine ecosystems were once among gleaf pine forest to be lost (Schwarz 1907, the most extensive in North America, occu- Croker 1987, Simberloff 1993, McGuire 2001). pying about 37 million ha prior to European As the southern landscape became increas- settlement (Frost 1993, Landers et al. 1995). ingly domesticated, the modified structure While the initial impact of immigrants during of expansive agricultural areas and linear 4. Longleaf Pine Regeneration Ecology and Methods 97 transportation corridors fragmented previ- as stream bottoms, wetlands, and seeps (Hilton ously contiguous habitat, thereby impeding 1999). The natural variability of these ecosys- the movement of natural surface fires across tems makes them excellent habitat for a variety these lands (Walker 1999). During the twen- of game animals and numerous nongame and tieth century, organized programs of fire sup- rare wildlife species (Kantola and Humphrey pression and policies of fire exclusion from the 1990; Engstrom 1993; Guyer and Bailey 1993; forest further interrupted natural fire regimes Crofton 2001; Engstrom et al. 2001; Brockway (Croker 1987). Since the absence of frequent and Lewis 2003; Means this volume). surface fires impedes the natural regeneration The complex natural patterns and processes of longleaf pine and allows invasion of long- unique to longleaf pine forests create ex- leaf pine sites by hardwoods and more ag- traordinarily high levels of biological diver- gressive southern pines, interruption of nat- sity in these ecosystems, with the great num- ural fire regimes is believed to be the most ber of plant species per unit area qualifying ecologically significant cause for its continu- these as among the most species-rich terres- ing decline (Wright and Bailey 1982, Landers trial ecosystems outside the tropics. As many et al. 1990, Pyne 1997, Gilliam and Platt 1999). as 140 vascular plant species have been ob- Longleaf pine forests have undergone a steady served in a 1000 m2 area and equally impres- decrease to 8 million ha in 1935 (Wahlenberg sive counts of more than 40 species per m2 1946), 2 million ha by 1975, 1.5 million ha in have been recorded (Peet and Allard 1993), 1985 (Kelly and Bechtold 1990) and less than a large number of which are restricted to 1.2 million ha currently (Outcalt and Sheffield or found principally in longleaf pine habi- 1996). Occupying less than 3% of their original tats. Habitat reduction resulting from decline range (Ware et al. 1993), longleaf pine ecosys- of longleaf pine ecosystems has caused the tems are now recognized as being at high risk increased rarity of 191 vascular plant taxa (Noss et al. 1995, Kush 2002). Unfortunately, (Hardin and White 1989, Walker 1993) and area reductions continue for stands in every several vertebrate species. Concern over loss diameter class below 41 cm (Kelly and Bech- of this unique ecosystem (Means and Grow told, 1990), an indication that most remaining 1985; Noss et al. 1995) has led to many efforts longleaf pine forests are aging without replace- focused on effectively restoring longleaf pine ment. ecosystems (Walker and Boyer 1993; Walker 1995; Kush 1996, 1998, 1999, 2001; Johnson and Gjerstad 1998, 1999; Brockway et al. 1998; Ecological Restoration Seamon 1998; Outcalt et al. 1999; Brockway Extending along the Gulf and Atlantic Coastal and Outcalt 2000; Provencher et al. 2001a,b; Plains from Texas to Virginia and inland to Mulligan et al. 2002). Since longleaf pine still the Piedmont and mountains in Alabama and occurs in isolated fragments over most of its Georgia, longleaf pine forests, woodlands, and natural range, it is reasonable to conclude savannas may occupy a wide variety of sites, that restoration of these ecosystems is possible ranging from wet poorly drained flatwoods (Landers et al. 1995). Effective methods for re- to mesic uplands, xeric sandhills, and rocky generating longleaf pine will no doubt play a mountain ridges (Boyer 1990a; Stout and key role in ecological restoration efforts. Marion 1993). Distinguished by a generally open, parklike stand structure (Schwarz 1907; Disturbance Dynamics and Fire Wahlenberg 1946), naturally regenerated lon- as an Ecological Process gleaf pine forests are typically an uneven-aged mosaic of even-aged patches distributed across Longleaf pine ecosystems exist in an envi- the landscape, which vary in size, structure, ronment influenced by large-scale catastrophic composition, and density (Platt and Rathbun disturbance, such as damaging tropical storms. 1993; Brockway and Outcalt 1998) and con- Lightning is an important agent in individ- tain numerous embedded special habitats such ual tree morality and creation of small-scale 98 II. Ecology disturbance in longleaf pine forests (Komarek Regeneration Environment 1968; Taylor 1974). The structure, pattern, and diversity of longleaf pine ecosystems are main- Difficulties encountered during early attempts tained by a combination of site factors and to regenerate longleaf pine impeded its recov- periodic disturbance events, including light- ery and contributed to its historical decline ning strikes, tree mortality, and animal inter- (Croker 1987). Erratic seed production, poor actions at local scales and tropical storms, soils, seedling survival, and slow early growth of and hydrologic regimes at broader scales. Dis- seedlings discouraged forestland managers turbances across site gradients provide large from investing in longleaf pine. Management living trees, snags, coarse woody debris, for- policies based on these initial observations est canopy gaps, and hardwood thickets that further contributed to the decrease of longleaf support numerous plant and animal species pine forests, as harvested stands were deliber- adapted to these disturbance-prone, yet largely ately converted to other southern pine species stable ecosystems. rather than being regenerated with longleaf Longleaf pine is closely associated with pine. Fortunately, later research illuminated wiregrass (Aristida spp.) in the eastern part the ecological mechanisms and identified and bluestem grasses (Andropogon spp. and silvicultural methods for effectively regener- Schizachyrium spp.) in the western portion of ating longleaf pine by natural and artificial its range. The understories of longleaf pine means (Boyer and White 1990; Barnett et al. forests are typically dominated by herbaceous 1990; Kush 2002) and the earlier policies of plants because these bunchgrasses facilitate the forest type conversion have now been largely ignition and spread of frequent surface fires reversed. (Landers 1991). In these ecosystems, longleaf The unique structural and process dynam- pine and bunchgrasses function together as ics characteristic of longleaf pine forests pro- keystone species that facilitate but are resis- vide both challenges to and opportunities for tant to fire (Platt et al. 1988; Noss 1989). They applying science and adapting technology to also exhibit substantial longevity and demon- efficiently obtain regeneration. Foremost, all strate nutrient and water retention to a de- longleaf pine forests are obligatorily pyro- gree that reinforces their site dominance and phytic ecosystems. Therefore, all regenera- minimizes change in the plant community fol- tion techniques employed must be compati- lowing disturbance (Landers et al. 1995). As a ble with periodic surface fires. Longleaf pine key ecological process and disturbance agent, forests are disturbance-prone and naturally re- the benefits of periodic fire include (1) main- generate in a variety of configurations rang- taining the physiognomic character of longleaf ing from relatively small circular or elliptical pine ecosystems through excluding invasive canopy gaps and attenuated strings to larger plants that are ill-adapted to fire, (2) preparing areas of partially blown down or almost com- a seedbed favorable for the establishment of pletely blown down overstory trees (Croker longleaf pine seedlings, (3) reducing the den- and Boyer 1975; Palik and Pedersen 1996; sity of understory vegetation thus providing Brockway and Outcalt 1998). As a tree species microsites for a variety of herbaceous plants, that is intolerant of competition, whether for (4) releasing nutrients immobilized in accu- light, moisture, or nutrients (Boyer 1990a), mulated phytomass for recycling to the in- its seedlings become established and flourish fertile soil and subsequently more rapid up- as opportunists responding to resource avail- take by plants, (5) improving forage for graz- ability. Although longleaf pine displays many ing, (6) enhancing wildlife habitat, (7) control- traits consistent with an intolerant, early seral ling harmful insects and pathogens, and (8) species, its seedlings often persist in the for- reducing fuel levels and wildfire hazard (Mc- est understory for prolonged periods similar Kee 1982; Wade and Lewis 1987; Boyer 1990b; to those of more tolerant, late-seral species. Wade and Lundsford 1990; Dickmann 1993; Suppressed longleaf pine seedlings will not re- Brennan and Hermann 1994; Brockway and spond with improved growth until released Lewis 1997). from competition with overstory trees and 4. Longleaf Pine Regeneration Ecology and Methods 99 long-suppressed individuals are, through time, is no phenological barrier to natural cross- increasingly prone to mortality. Whether a spe- ing (Boyer 1990a). Hybridization between lon- cific longleaf pine seedling survives to eventu- gleaf pine and slash pine is far less likely be- ally become a dominant member of the forest cause of differences between their dormancy canopy may depend as much on stochastic and heat requirements for flowering (Boyer events (i.e., drought severity and duration, 1981); however, artificial crossing can be easily fire intensity and season, disturbance that re- achieved. Hybrids between longleaf pine and sults in timely release from competition) as shortleaf pine (Pinus echinata) have not been it does on the competitive vigor (i.e., genetic observed in nature but have been artificially attributes) of the individual. Whether by ar- produced (Snyder et al. 1977). tificial or natural methods, by even-aged or With a pattern of genetic variation similar uneven-aged silvicultural techniques, not only to that in other southern pines, longleaf pine is efficiently regenerating longleaf pine feasi- is suitable for genetic improvement. However, ble, it is also imperative for achieving ecological the effort expended on this species is easily restoration and ecosystem sustainability goals. dwarfed by the immense resources devoted to loblolly pine and slash pine improvement programs (Schmidtling 1999). Traditional tree improvement approaches, which select “plus” Reproductive Biology trees in the forest based on size and form, have not proven to be suitable for longleaf pine. Genetics Variation in the “grass” stage of longleaf pine Longleaf pine is a tree species of consider- makes it impossible to determine the true age able genetic diversity, with variation among of a tree and thus its true growth potential. individuals typically greater than that among Therefore, tree improvement programs for stands or geographically different seed sources longleaf pine have shifted their emphasis to a (Snyder et al. 1977; Lynch 1980). Although ge- progeny test approach, with the duration of the netic variation among populations is thought grass stage and resistance to brown-spot fun- to be a result of the diversity of environments gus (Mycosphaerella dearnessii) the most impor- in which longleaf pine occurs throughout its tant inherited traits of interest. When focusing native range (Boyer 1990a), measures of ge- efforts on accelerating the early height growth netic diversity appear unrelated to climate of longleaf pine, one of the greatest dangers variables (Schmidtling and Hipkins 1998). is the possibility of incorporating loblolly pine While the pattern of genetic variation for long- genes that will result in a hybrid that begins leaf pine is similar to that of other south- growing earlier but has poor form and in- ern pines (Schmidtling 1999), its unique pat- creased susceptibility to fusiform rust fungus tern of allozyme variation is indicative of a (Cronartium quercuum f. sp. fusiforme). There very different history during the recent Ice is no ecotypic differentiation in longleaf pine Age (Schmidtling et al. 2000). Unlike other based on site conditions and no important dif- pines, longleaf pine appears to have migrated ference in survival or growth between east- eastward across the southeastern United States ern and western populations, as occurs with from a single refuge in southern Texas and/or loblolly pine (Schmidtling 1999). northeastern Mexico after the Pleistocene (Schmidtling and Hipkins 1998). The progressive decrease in allozyme diversity from Flowering, Pollination, and western to eastern longleaf pine populations Fertilization represents a loss in genetic variability from stochastic events during migration. Longleaf pine is monoecious, with male strobili Longleaf pine may form a natural hybrid (catkins) predominating in the lower crown with loblolly pine, referred to as “Sondereg- and female strobili (conelets) occurring most ger pine” (Pinus sondereggeri). Since flowering frequently in the upper crown of the same tree of these two species frequently overlaps, there (Schopmeyer 1974). Development of catkins 100 II. Ecology

wet spring and early summer followed by a dry period in late summer (Shoulders 1967). Be- cause of the differential conditions that favor each sex, large crops of male and female flowers do not necessarily coincide (Boyer 1990a). Heavy annual losses of longleaf pine conelets can usually be expected, with observed losses ranging from 65% to 100% (Boyer 1974a; McLemore 1977; White et al. 1977). Insects, weather extremes, and insufficient pollen appear to be the primary causes for these losses, which primarily occur during spring pollination or the following summer. Peak pollen shed and conelet receptivity typically vary from late February in the southern portion of the native range to early April in more northern areas (Boyer 1990a). Although pollen shedding and receptivity coincide on individual trees, there appears to be very little synchrony among trees in a longleaf pine forest, with some trees being consistently early and others being consistently late and overall dates highly influenced by air temperatures before and during the flowering period. Pollen FIGURE 2. Elongated catkins, subtending a terminal shedding typically occurs during a period of 5 bud, shown after their pollen has been shed. Photo to 21 days, with an average of 13 days (Boyer courtesy of the Forestry Images Organization. 1981). Year-to-year variation in the time of pollen shedding appears to be related to accumulation of degree-day heat sums, with all (Fig. 2) and conelets (Fig. 3) is initiated dur- temperatures greater than 10o C after January ing the growing season before buds emerge, 1 promoting development of the male strobili with catkins beginning to form during July (Boyer and Woods 1973; Boyer 1973, 1978). and conelets being formed during a short pe- While pollination takes place in late winter riod in August (Boyer 1990a). First appearing or spring, fertilization does not occur until the at the base of vegetative buds between mid- following spring. Conelets grow rapidly after November and early December, purple catkins fertilization, increasing in length from 2.5 cm remain dormant for several weeks before re- to 18 cm by May or June (Boyer 1990a). Cones suming development between late December reach maturity between mid-September and and early February (Boyer 1981). Conelet buds mid-October of their second year and range in appear in January or February and conelets, length from 10 cm to 25 cm. Although cone upon emerging from the bud, are red until color changes from green to brown as they pollinated, after which they fade to yellowish ripen, cones may be ripe before changing color green. The development rates of both catkins (Schopmeyer 1974). and conelets are almost entirely dependent on ambient temperature (Boyer 1990a). The number of flowers produced appears re- Cone Production and Seed lated to weather conditions during the year Dispersal of initiation. Catkin production is favored by abundant rainfall throughout the growing sea- Longleaf pine cone crops are highly variable son, while conelet production is promoted by a from year to year (Fig. 4), with 1860 cones/ha 4. Longleaf Pine Regeneration Ecology and Methods 101

FIGURE 3. Conelets, located peripherally to terminal buds, are most often observed in the upper crown. Photo courtesy of the Forestry Images Organization.

(e.g., 30 cones/tree and 62 seed trees/ha) nor- eter at breast height produce on average 65 mally required for successful natural regen- cones/year compared with 15 cones/year from eration (Boyer 1996). While cone production trees in the 25–33 cm diameter at breast height may be influenced by the density of airborne size class (Boyer 1990a). The number of vi- pollen during flowering, low cone crop fre- able seeds per cone varies with the seed crop quencies appear to be more a result of flower for a specific year, ranging from 50 seeds/cone losses rather than a failure to produce flow- in good years to 35 seeds/cone during average ers (Boyer 1974a, 1987a). Since 1986, cone years to 15 seeds/cone in poor years (Croker crops on coastal plain sites from Louisiana to 1973). North Carolina have increased to an average of Peak seed production is observed in lon- 36 cones/tree from an earlier average of only gleaf pine forests having stand densities be- 14 (Boyer 1998). This increase in cone pro- tween 6.9 and 9.2 m2/ha, when principally duction appears to be a result of both an in- comprised of dominant and codominant trees crease in flower production and an increase of cone-bearing size (Boyer 1979). Such stands in the fraction of flowers surviving to become produce seed crops adequate for natural re- mature cones. Cone production of individual generation once every 4 to 5 years on aver- trees is influenced foremost by genetics and age (Croker and Boyer 1975). When forests secondarily by tree size, crown class, stand of substantially greater density are thinned to density, and site quality. The greatest cone this level, increased cone production resulting production occurs on dominant, open-grown from decreased intraspecific competition does longleaf pines having large crowns (Croker not occur for three growing seasons (Croker and Boyer 1975). Trees of 38–48 cm diam- 1952). Release following conelet initiation 102 II. Ecology

150

120

Cones per Tree 60

0 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 Year

FIGURE 4. Example of high temporal variability for longleaf pine seed production: Cone crops at the Escambia Experimental Forest, southern Alabama, 1957–2003. benefits the crop only by reducing moisture (Croker and Boyer 1975). Although germina- stress during dry periods (Boyer 1990a). Seeds tion and establishment require that seeds con- are dispersed by wind during a 2- to 3-week pe- tact mineral soil, the large seed and wing can riod between late October and late November, impede penetration through dense grass or ac- depending on weather conditions. Longleaf cumulated litter on the forest floor. Root sys- pine seeds are the largest of the southern pines tems of premature germinants fail to reach and their dispersal distance is limited, with mineral soil and readily die from desiccation. 71% of sound seeds falling within 20 meters The risk of regeneration failure can be signifi- of the base of a parent tree (Croker and Boyer cantly reduced by using mechanical treatment 1975). and/or prescribed fire to prepare a suitable seedbed before seedfall. Germination begins with emergence of the radicle and an almost simultaneous elongation of the cotyledons Seedling Development (Allen 1958). Hypocotyl elongation begins soon after radicle emergence but is limited. Germination and Shoot Growth Growth of the cotyledons lifts the seedcoat Longleaf pine seeds typically germinate (Fig. 5) from the ground. Limited hypocotyl growth within a week of contacting the ground (Boyer causes the cotyledons of new germinants to 1990a). While rapid germination may be an remain at or near the ground line. Newly ger- adaptation to reduce the risk of exposure to minated seedlings are relatively inconspicuous seed predators, newly germinated seedlings with their small primary needles. Although are also vulnerable to mortality from animals, secondary needles appear within 2 months, pathogens, and adverse weather conditions the epicotyl reaches a length of only 0.38 cm 4. Longleaf Pine Regeneration Ecology and Methods 103

FIGURE 5. Newly germinated longleaf pine seedlings emerging from thin forest litter (foreground) and among cones fallen from the parent trees. Photo courtesy of the Forestry Images Organization. and, during the next 8 weeks, does not elon- typical rates are 0.8 cm/day during the ﬁrst gate as in other pines. This stemless condi- 60 days (Allen 1958). Taproot development tion is one of the unique characteristics of lon- is inversely related to moisture conditions, gleaf pine, commonly referred to as the “grass with greater root elongation occurring in drier stage” (Fig. 6). Depending on ambient condi- soils. The general root structure of natural tions, longleaf pine seedlings may remain in seedlings is unaffected by soil conditions, but this stage for 2 to as long as 15 years, during these seedlings have slower root growth than which they are most susceptible to their ma- those measured in greenhouse tests. The root jor disease, brown-spot needle blight (Boyer system of a typical 1-year-old seedling con- 1990a). sists of a taproot 60–70 cm long and a number of strong laterals 50–60 cm long in the upper soil layers with numerous attached feeder tips Roots and Sprouts (Wahlenberg 1946). Laterals extend outward, While in the grass stage, longleaf pine seedlings and though they often change direction due to devote much of their energy to root produc- obstructions, they remain at a uniform depth tion. Following germination, the radicle forms from origin to tip (Heyward 1933). On wet- a taproot that develops very rapidly. Growth ter ﬂatwoods sites, root systems have the same rates may be as high as 50 cm in 15 days follow- general architecture, but root distribution is re- ing germination in sandy soils under green- stricted mostly to surface horizons, above the house conditions (Wahlenberg 1946). More spodic layer that occurs in soils typical of these 104 II. Ecology

FIGURE 6. Longleaf pine seedling in the grass stage. Photo courtesy of the Forestry Images Organization. areas. Even in well-drained sands, 90% of all smaller seedlings are rather weak and often lateral roots are found in the upper 30 cm of die following subsequent fires. the soil. Longleaf pine seedlings can sprout from the root collar if the top is killed. Under natural Competition conditions, sprouts often result following top Longleaf pine seedlings are intolerant of kill from fire. This sprouting ability diminishes interspecific and intraspecific competition rapidly once seedlings emerge from the grass (Harrington, this volume). Seedling growth stage and begin stem elongation. As many as rates improve as the distance from adult pines 40% of the grass-stage seedlings that were increases, with the suppressive effect from cut off at the groundline have been found stands of overstory trees adjacent to clearcut to have living sprouts 1 year later (Farrar strips being greater than that from single over- 1975). Only 14% of the seedlings, that had story trees (Boyer 1963). The relationship of initiated height growth but were less than declining seedling growth rates with increasing 1.37 m tall, produced sprouts when cut at amounts of overstory competition (i.e., over- the groundline. None of the trees more than story basal area) follows that of a general expo- 1.37 m tall produced sprouts after cutting. nential decay curve. The root collar diameter Three years after cutting, 30% of the grass- for 4-year-old seedlings has been reported to stage seedlings still had surviving sprouts, but sharply decrease from 1.2 cm to 0.7 cm as over- these were growing very slowly. Although story basal area increased from 0 to 6.9 m2/ha. fewer of the seedlings that had begun height More recent work has substantiated the sup- growth sprouted, those sprouts were much pressive effect of overstory adults upon long- larger 3 years after treatment. Sprouts from leaf pine seedlings on redhills sites (Grace and 4. Longleaf Pine Regeneration Ecology and Methods 105

Platt 1995) and flatwoods sites (Gagnon et al. significantly related to N, soil water, and their 2003, 2004) and the exponential relationship interaction in a greenhouse study (Jose et al. between seedling growth and overstory basal 2003). Light was important only if water was area (Palik et al. 1997). not limiting. Therefore, competition affecting Overstory hardwood trees are even greater longleaf pine seedling growth follows the fun- competition for longleaf pine seedlings. While damental principle of limiting factors, with 60% of the seedlings in the 0–3 m zone near growth most impaired by whichever essential adult longleaf pines reached sufficient root col- factor is most limited in a specific environment. lar diameter by age 9 to begin height growth (Smith 1962), only 24% of the pine seedlings Initiation of Height Growth near oak trees had initiated height growth. Re- Although varying considerably, longleaf pine moval of overstory oaks significantly improved seedlings usually begin emerging from the height growth of longleaf seedlings (Walker grass stage when their root-collar diameter 1954). Understory vegetation, including other reaches about 2.5 cm (Boyer 1990a). Infec- longleaf pine seedlings and herbaceous plants, tion by brown-spot needle blight can substan- also competes with longleaf pine seedlings. tially delay grass stage emergence. The height Single season growth increases of 18% were growth of seedlings also depends on the in- measured in seedling densities of 247,100/ha tensity of competition in the ambient environ- and increases of 64% were observed at ment (Ramsey et al. 2003). Therefore, treat- densities of 2471/ha (Pessin 1938). When ments that reduce competition will increase herbaceous competition was removed, growth the proportion of seedlings that emerge from increased by 25 and 456% for the same respec- the grass stage (Haywood 2000). Differences in tive seedling densities. Thus, intraspecific com- levels of intraspecific and interspecific compe- petition in the understory appears much less tition and individual genetic control of growth important than competition from other species rates and resistance to brown-spot infection rein this layer. sult in considerable variation in emergence of Ascertaining the underlying causes of longleaf pine seedlings from the grass stage. seedling growth reduction has proven chal- This high within-stand variation is actually lenging. Root competition was believed to beneficial as it leads to the early establish- be much more important than light because ment of dominant seedlings, preventing stand growth depression extends well beyond the stagnation even when seedlings occur at very height of adjacent overstory trees (Walker and high densities. The key to height growth ini- Davis 1956; Croker and Boyer 1975). This in- tiation is accumulation of carbohydrate re- hibitory effect is also greater on sites with serves in the root sufficient to support rapid poor soils, another indication of belowground upward expansion of the stem. This interval competition. The importance of belowground is known as the bolting stage in which rapid competition was more recently substantiated growth of the needle-covered stem occurs with by research on xeric sandhills sites in Florida, limited lateral branch development, produc- where total daily light influx did not differ sig- ing the characteristic “bottlebrush” structure nificantly with distance from adult trees, but (Fig. 7). Under favorable conditions, longleaf the fine root biomass of overstory trees did pine seedlings can grow 30–90 cm in a single decline with distance (Brockway and Outcalt season (Wahlenberg 1946). 1998). Others working on more fertile soils with higher densities of woody plants have shown that light does increase with distance Effects of Fire from gap edge (Palik et al. 1997; McGuire et al. Longleaf pine evolved in ecosystems subject 2001; Gagnon et al. 2003). Palik et al. (1997) to frequent low-intensity surface fires and also documented a positive relationship be- has developed many adaptations for enhanc- tween available N and seedling growth. Lon- ing survival and growth in this environment gleaf pine seedling growth was found to be (Landers 1991). Although newly germinated 106 II. Ecology

FIGURE 7. Longleaf pine seedlings in the bolting stage. Photo courtesy of the Forestry Images Orga- FIGURE 8. This larger grass-stage longleaf pine sur- nization. vived a recent fire; note the thick root collar and partially consumed needles that adequately protected its terminal bud. Photo courtesy of the Forestry Im- and small grass-stage seedlings are suscepti- ages Organization. ble to fire-caused mortality, once they attain a root-collar diameter of 1.3 cm, they become quite fire-resistant (Boyer 1974b). Such larger caused mortality. During this stage, the termi- grass-stage seedlings have thicker bark and a nal bud is situated directly in the flaming zone, large tuft of needles that protect the central where it may be exposed to maximum tem- meristem (Fig. 8) and, as noted previously, peratures. These seedlings are especially vul- even if the top is killed, they will often sprout. nerable during early season bud break when However, since many natural seedlings do not the candles are rapidly expanding and there attain this size until they are 2 or 3 years old, is very little protective needle mass. Once they mortality from fires can be quite high. A sin- have bolted above the flame zone (about 1 m), gle growing-season fire, at age 2, can result in seedlings are again quite resistant to fires of as much as 80% of the total seedling mortal- moderate intensity. Although fire does not kill ity for the 3-year period following germina- many longleaf pines once they reach this small tion. Fire can also be beneficial to longleaf pine sapling stage, burning can still cause reductions seedlings by consuming brown-spot-infected in growth (Boyer 1987b). needles. Fire stimulates height growth (Grelen 1983) by reducing competition and releasing Seedling Mortality nutrients immobilized in organic matter. When initiating height growth, seedlings Seedlings are susceptible to losses from once again become more susceptible to fire- frost heaving, flooding, disease, logging, fire, 4. Longleaf Pine Regeneration Ecology and Methods 107

mortality during the first 2 years, where losses up to 90% may often occur (Grace and Platt 1995; Provencher et al. 2001b). The probability of seedling mortality from fire is greatest in the zone around adult trees, because greater needle litter accumulations there often result in hotter fires. Higher intraspecific competition in this zone also results in slower seedling growth, thus predisposing them to higher rates of fire-related mortality. Before the first fire occurs, seedlings growing close to adult pines rarely attain the 1.3 cm root-collar diameter needed for survival. Drought can cause significant losses of first- year seedlings, especially on sandhills sites where 50% mortality is common during the spring drought period. Many of the seedlings that perish during dry periods are growing on locations where the forest floor is somewhat thicker. Although bare mineral soil is ideal for longleaf pine germination and growth, sufficient moisture is all that is needed to initiate the germination process. Therefore, many of the seedlings that germinate on thicker forest floor material in the fall, fail to survive the spring drought because their root system FIGURE 9. Close-up view of longleaf pine needles has not penetrated sufficiently into the min- infected with brown-spot fungus; note highly var- eral soil to obtain adequate moisture. Litterfall iegated color tones and distinctive dark spots on is of course not distributed evenly, but rather the needles. Photo courtesy of the Forestry Images is highest near adult trees. The greater for- Organization. est floor thickness, increased root competition, and reduced rainfall from crown interception drought, and animals. As expected, losses are all combine to substantially lower the prob- greatest during the first growing season. Frost ability of longleaf pine seedling survival near heaving is a threat only near the northern limit adult overstory trees. This typical spring mor- of the natural range on finer-textured soils. tality significantly contributes to development Flooding can be a problem on wet coastal plain of the often-observed zone of seedling exclu- sites where prolonged standing water can kill sion near adult pines (Brockway and Outcalt newly germinated seedlings (Croker and Boyer 1998). 1975). Brown-spot needle blight (Fig. 9) may Both wild and domestic animals can deci- at times become so severe that it destroys all mate stands of longleaf pine seedlings. Locally the foliage and directly kills the seedling. More high populations of rabbits can cause signif- often, the seedling becomes so weakened that icant damage to newly germinated seedlings it dies during the next fire. While depending by clipping off the needles (Croker and Boyer on the harvest method, season, and volume of 1975). Local losses can also result from pocket trees removed, losses from logging generally gophers feeding on the starch-laden roots of average about 50% (Boyer 1990a). To mini- grass-stage seedlings. While cattle will also mize logging losses, harvesting should be done graze on longleaf pine seedlings, most damage when seedlings are in the grass stage. As noted and mortality results from trampling where above, fire is a significant cause of seedling sites are allowed to suffer from overstocking 108 II. Ecology and excessive use. However, the most destruc- Regeneration Requirements tive of all animals is the wild hog, a nonnative species introduced from Europe. The roots When the fundamental requirements can be of grass-stage seedlings, rich in starch, are a met, natural regeneration is a practical and favorite food of wild hogs. A single hog can inexpensive management option for existing uproot and consume a great many seedlings longleaf pine forests (Boyer 1993a). The fore- in a single day. The destructive power of the most requirements for successfully regenerat- vast herds of wild hogs was noted in the ing longleaf pine by natural means are an ad- early 1900s (Schwarz 1907). An early fenc- equate seed source and a receptive seedbed ing study demonstrated their effect on longleaf (Boyer and White 1990). A sufficient num- pine seedling densities, with 14,826 stems/ha ber of sexually mature parent trees, having occurring within the protective fence and a desirable characteristics and being well dis- mere 20 stems/ha surviving outside (Mattoon tributed throughout the area, typically serves 1922). In some locations, wild hogs still pose to meet this need (Dennington and Farrar a problem for regeneration. However, remov- 1991). Since longleaf pine seed crops are infre- ing hogs for only 2 years resulted in successful quent, averaging one every 5 to 7 years, annual regeneration of longleaf pine in the South Car- monitoring of conelet and cone production is olina Coastal Plain (Lipscomb 1989). necessary to appropriately forecast years when adequate seed will be available (Boyer 1996). For initiating even-aged stands, the size, number, and distribution of seed-bearing trees must Natural Regeneration be such that a minimum of 1860 and preferably 2470 or more cones/ha will be provided Regeneration Problems (Boyer and White 1990). Peak seed production may be encouraged in future years by reduc- Perhaps the greatest impediment to wider ac- 2 ceptance of longleaf pine, as a tree worthy ing stand density to between 6.9 and 9.2 m /ha of forest management investment, is its rep- (Boyer 1979). utation as a difficult species to successfully One of the most pervasive problems imped- regenerate, whether by natural or artificial ing the natural regeneration of longleaf pine means. Schwarz (1907) was among the first to is high-density understory and midstory lay- identify poor regeneration as a major threat to ers dominated by woody plants. If preestablish- the future of longleaf pine. Insufficient num- ment competition control is not achieved, lon- bers of seed trees, infrequent seed crops, cone gleaf pine regeneration success will be highly infestation by insects, limited dispersal dis- doubtful (Boyer and White 1990). Therefore, tance of heavy seeds, seed predation, seed competing hardwood trees and shrubs must perching in litter above mineral soil, vulner- be effectively controlled before seed disper- ability of early germinating seed to tempera- sal. The presence of competing woody plants ture extremes, brown-spot fungus infection, is not only deleterious to seedlings as they slow early seedling growth, untimely fires, are becoming established, but may also inter- and fire exclusion favoring competing species fere with dispersed seeds as they attempt to are among the primary reasons for regenera- reach mineral soil. Although prescribed fire tion failure of longleaf pine (Wahlenberg 1946; conducted well in advance of seedfall may Croker and Boyer 1975; Boyer 1979). Fortu- be sufficient to control these woody plants, nately, the extensive research conducted on herbicide application or mechanical treatment longleaf pine regeneration during the second may also be necessary to obtain adequate con- half of the twentieth century has now re- trol under certain circumstances (Croker and duced these problems to considerations that Boyer 1975). can be effectively addressed through appropri- Longleaf pine seeds need to be in contact ate management strategies and practices (Kush with mineral soil for successful germination 2002). and establishment (Boyer and White 1990). 4. Longleaf Pine Regeneration Ecology and Methods 109

Since seed supplies are typically limited, a well- seedlings. Seedling growth beneath adults re- prepared mineral seedbed is necessary to pro- mains inhibited until nearby overstory trees vide optimum opportunity for regeneration are removed (Boyer and White 1990). Peri- success. A suitably prepared seedbed can be odic prescribed fire should adequately curtail created by a prescribed fire conducted within understory plants, with herbicide application 1 year prior to seedfall. Such a fire will remove reserved for use under specific circumstances. enough forest litter to double the rate of lon- Control of the brown-spot fungus, the most gleaf pine seedling establishment compared to serious pathogen of longleaf pine, is essential unburned sites. While only 14% of longleaf for successful natural regeneration (Boyer and pine seeds became established as seedlings on White 1990; Boyer 1990a; Dennington and unburned areas, 36% became established on Farrar 1991). Timely application of prescribed winter-burned areas and 24% became estab- fire, which consumes the infected needles, lished on fall-burned areas (Croker 1975). The is the most effective remedy for this needle lower establishment level on fall-burned areas blight. In newly established seedlings, pre- is believed to result from lighter subsequent lit- scribed fire should be deferred until an ade- ter accumulation causing higher rates of seed quate number of the seedlings reach a fire- predation by birds. resistant size. Unfortunately, seedlings beneath Successful regeneration requires an ade- adult trees grow slowly and delaying fire could quate number of established seedlings well- allow brown-spot needle blight to spread in a distributed throughout the area (Boyer and stand, further inhibiting seedling growth. This White 1990). Although exact criteria must be risk can be abated by maintaining relatively left up to individual land managers with spe- low basal areas where seedlings are less likely cific management objectives, a reasonable goal to be suppressed by overstory trees. when using the shelterwood method might be Young seedlings must also be protected from about 14,800 established 1-year-old or older hogs, which can rapidly destroy seedlings by seedlings per hectare. This standard allows for consuming them while in the grass stage about 50% mortality among seedlings, when (Boyer 1990a; Boyer and White 1990). Cattle the stand is later logged, and anticipates high grazing will also reduce the fine fuel needed to rates of mortality from other sources, espe- effectively carry fire in these ecosystems and cially brown-spot fungus, during the early may directly damage seedlings if conducted at years following establishment. It is reasonable high intensities (Dennington and Farrar 1991). to expect that at least 1235 well-distributed Regenerating longleaf pine by natural means seedlings per hectare could result that are requires careful attention to details relevant about1minheight and therefore free from to advanced planning, monitoring site condi- brown-spot infection and relatively safe from tions, prescription development, and timing of damage by fire (Croker and Boyer 1975). cultural treatments. Through identifying the Since longleaf pine is highly sensitive to area, assessing the seed source, monitoring competition, established seedlings develop op- cone crops, controlling competition, preparing timally in the absence of competitors. Al- a suitable seedbed, conducting regeneration though hardwood trees and shrubs comprise surveys of seedlings, appropriately adjusting the most severe competition for young longleaf overstory density, and conducting postestab- pine, these should be largely eliminated during lishment burns and other treatments, the like- preestablishment treatments. While eliminat- lihood of successfully regenerating longleaf ing all competitors is not practical, action to pine forests will be substantially increased. obtain postestablishment competition control is indicated when hardwood densities exceed 2.3 m2/ha (Boyer and White 1990). Herba- Unique Considerations ceous understory plants and adult longleaf Largely as a result of ignorance concerning pine trees in the overstory typically remain as the unique life history of longleaf pine, mis- the major competitors for newly established management may be the rule rather than the 110 II. Ecology exception (Kush 2002). Unlike other south- continuing long-term suppression makes them ern pines, longleaf pine is both resistant to and highly vulnerable to mortality from a vari- dependent on fire and its seedlings must pass ety of individual and cumulative stress factors through a stemless grass stage. If unwise man- (Boyer 1974b; Boyer and White 1990). There- agement or the absence of management allows fore, unless natural or anthropogenic distur- competing species to grow freely while lon- bances remove nearby adults from the canopy, gleaf pine is in the grass stage, then it can be lost it is doubtful that seedlings present beneath from a site and will not become reestablished mature pines will ever ascend to become dom- and regain dominance without intervention inant members of the forest canopy. that provides the requisite disturbance regime. The impoverished (low available moisture Although periodic surface fires are essen- and nutrients) and relatively simple (very tial for sustaining longleaf pine forests, these few woody plants other than longleaf pine) must be prudently timed because longleaf pine ecosystem represented by the well-drained seedlings remain highly vulnerable to fire- and somewhat excessively drained, coarse- caused mortality until they reach a ground- textured soils in north central Florida pro- line diameter of 7.5 mm and a height of 1 m vided a unique opportunity to assess the re- (Kush 2002). Because of their high suscepti- sults of gap-phase regeneration in longleaf bility to competition for aboveground and be- pine that are the cumulative product of re- lowground resources, their slower growth near source competition and repeated fires over adult trees, and the greater needle litter (i.e., several decades. In examining these long-term fuel) accumulation under tree crowns, longleaf patterns of natural longleaf pine regeneration pine seedlings located beneath adult trees are in numerous naturally occurring forest gaps, more vulnerable to mortality from fire than are Brockway and Outcalt (1998) observed that those in openings (Croker and Boyer 1975). most surviving longleaf pine seedlings aggre- Prescribed burning, with the objectives of ade- gated near the center of canopy gaps and were quately reducing fuels and competition while encircled by a zone approximately 12–16 m affording protection for seed-bearing trees and wide from which they were generally absent. seedlings, should be conducted within 2 to This “seedling exclusionary zone” (SEZ) was 3 days following a saturating rain, at tem- found to spatially coincide with an area of peratures no greater than 27◦C, with relative greater fine root biomass and relatively in- humidity higher than 40% and wind speeds tense intraspecific root competition between steady at 8–13 kph (Croker and Boyer 1975). adult and juvenile longleaf pines. The cen- Longleaf pine is a competition-intolerant ter of the canopy gap, where most longleaf species, with its regeneration largely confined pine seedlings cluster, spatially coincides with to canopy gaps (Wahlenberg 1946). Abun- a fine-root gap, where seedlings are under less dance and growth of seedlings have long been competitive stress and thus enjoy higher rates reported to be negatively related to the pres- of survival and growth. Although all three fac- ence of adult longleaf pine (Walker and Davis tors are no doubt of some degree of impor- 1954; Davis 1955; Smith 1961), with the com- tance, in these relatively open longleaf pine petitive influence of individual trees and forest stands (57% cover) growing on xeric sand- edges on seedlings observed to extend up to hills, competition for soil moisture and possibly 16 m (less on better quality sites) from adults competition for nutrients appear to be propor- into forest gaps (Smith 1955; Walker and Davis tionally more important than competition for 1956; Boyer 1963). This competitive effect is so light. pronounced that the most frequently surviving Others studying gap-phase regeneration in and vigorous growing longleaf pine seedlings longleaf pine have typically examined short- typically cluster near the center of canopy gaps term responses of planted seedlings on higher (Brockway and Outcalt 1998). Although lon- quality forest sites (Palik et al. 1997; McGuire gleaf pine seedlings may persist for years be- et al. 2001; Gagnon et al. 2003). On more neath the crowns of adult trees, their growth mesic sites, with richer soils resulting in higher is impeded by competition from adults and stand density and a greater number of woody 4. Longleaf Pine Regeneration Ecology and Methods 111 plants with roots innervating the surface soil landowners. However, this approach is poten- layers, various root clusters are not readily dis- tially quite risky. Many longleaf pine stands tinguishable and a SEZ was not identified. On have been successfully established by direct these better sites, competition for light and nu- seeding, but there have also been many fail- trients is reportedly more important than com- ures resulting from adverse weather, seed pre- petition for soil water. Should the root geom- dation, or simply not following the required etry characteristic of xeric sites eventually de- procedures. The first step in attempting to velop on these mesic sites, it is likely that the achieve successful regeneration by this method resulting SEZ would be considerably narrower is proper seed selection. Longleaf pine has five (i.e., shorter foraging distance for adult roots), seed zones and seed should not be transported perhaps only 5–10 m wide, making its effects for use more than one zone in any direction difficult to distinguish from the influence of (Schmidtling 2001). Since there is no genetic the 4–7 m radius directly beneath the crown evidence for east–west or ecotypic site vari- of a mature longleaf pine tree. ation, seed can be collected from anywhere Survival and growth of longleaf pine within a zone for use in that zone. Seed from seedlings is negatively influenced by the den- one zone warmer should grow faster and seed sity of adult trees in the canopy and the prox- from one zone colder will typically grow slower imity of adults to the seedlings (Palik et al. than local seed. 1997; Brockway and Outcalt 1998; Gagnon Seedlots should be 95% pure and have a et al. 2004). Timely seedling release from in- germination rate of at least 75% (Barnett et al. traspecific competition is essential to achieve 1990). To obtain this germination rate, cones successful natural regeneration. Although re- must not be collected until they are fully ma- lease can be affected through stand treatments ture (Barnett and McGilvray 2002). Mature that thin the forest to a uniform overstory cones have a specific gravity of 0.89, which density (i.e., even-aged management meth- can be determined by the water displacement ods), recent appreciation for the ecological val- method (Barnett and McGilvray 2002). Guide- ues protected by maintaining an uneven-aged lines specify that cones should not be collected structure has increased interest in achieving until 19 of 20 cones have a specific gravity natural regeneration through selection silvi- equal to 0.89, which means the average culture techniques (Farrar 1996). Group se- specific gravity will be about 0.81 (Wakeley lection, which mimics the natural stand re- 1954). Cones must be stored in a dry location placement dynamics of gap-phase regenera- having adequate ventilation. Cone storage for tion, may be one of the most useful ap- a short period will increase seed yields, but proaches for regenerating longleaf pine forests. storage beyond 4 weeks may reduce seed qual- Although canopy gaps created on xeric sand- ity. Cones will dry in a kiln within 24–48 hours hill sites should perhaps range from 0.13 to depending on initial moisture and weather 0.8 ha (radius = 20–50 m) to sufficiently re- conditions. Kiln drying temperatures should duce the competitive influence of adults on not exceed 45◦C (Barnett and McGilvray seedlings (Brockway and Outcalt 1998), gaps 2002). Once cones are dry and open, seed as small as 0.1–0.14 ha (radius = 18–21 m) may be removed by using a tumbler. may be suitable for regenerating longleaf pine Newly extracted seed will have a relatively on mesic sites (Palik et al. 1997; McGuire et al. high moisture content and, therefore, should 2001; Gagnon et al. 2003). be refrigerated and then dried to an 8–10% moisture level. Dewinging the seed requires appropriate equipment and processing tech- Artificial Regeneration niques to prevent damage and a corresponding decline in quality (Barnett and McGilvray Direct Seeding 2002). The wing is not really removed, but Initiating a stand of longleaf pine from seed rather reduced to a stub. Debris is removed sown directly onsite can be a cost-effective by screening and air, followed by removal of regeneration method, especially for small empty and partially filled seeds on a gravity 112 II. Ecology table. Processed seed can then be stored in A scalper attachment may be used with the moisture-proof containers at no more than 2◦C row seeder to remove competition from the for up to 3 years. Storage for longer periods re- area being seeded. Recommended minimum quires a temperature of-18◦C. seeding rates are 32,000 repellent-treated vi- Seed should be tested for its germination able seeds/ha for broadcast application. When rate prior to use or sale. Since longleaf pine row seeding, use 16,800 seeds/ha with one seed is nondormant, stratification is not neces- seed placed every 23–30 cm within a row. sary. However, a 10-minute drench with beno- When spot seeding, 5–6 seeds should be de- myl fungicide prior to sowing will improve posited at each spot with spots arrayed at a germination. Since longleaf pine seed is quite 2 × 2 m spacing. Covering seeds lightly with large and nutritious, some type of repellent is about 6 mm of soil improves germination and needed to deter consumption by birds and ro- establishment success (Barnett et al. 1990). In dents (Nolte and Barnett 2000). On most sites, assessing seeding success, a minimum of 100 seed is best sown during the fall, when mois- survey plots (each 4 m2) per site is recom- ture is adequate and the maximum daytime mended (Mann 1970). If seedling stocking is temperature is below 29◦C (Dennington and greater than 50%, then a suitable stand of lon- Farrar 1991). On other sites, such as those in gleaf pine should develop. northern areas with finer-textured soils prone to frost heaving or those having high rabbit populations that may damage new seedlings, Bare-Root Nursery Seedlings seed should be sown in late winter or early The nursery management objective is to spring. Seed can be broadcast onsite by air or produce high-quality longleaf pine seedlings ground equipment. Since most sites are now with a root-collar diameter of 10 mm, at rather small, aerial seeding is rarely employed, least six primary lateral roots, a stout taproot, but ground application with a cyclone seeder is well-developed terminal bud, and many a commonly used method. Row and spot seed- needle fascicles (Barnett et al. 1990). The first ers operate more slowly, but use less seed per step in this process is to select good-quality unit area and provide increased spacing con- seed with 95% purity and a germination trol. Following dispersal, the seed can be cov- rate of at least 75% (Cordell et al. 1990). ered with soil, which improves seedling stock- A benomyl application to seeds will reduce ing especially on dry sandy sites. fungal pathogens and improve seedling es- Direct seeding should not be attempted on tablishment in the nursery bed (Barnett et poorly drained soils or sites having a high wa- al. 1999). Seedbeds should consist of sandy ter table, because new germinants are highly or loamy sand soils with a pH of 5–6. Weeds, susceptible to mortality from wet conditions insects, and pathogens are reduced by seedbed (Dennington and Farrar 1991). Steep sites fumigation prior to sowing. If mulch is used, it where seed is likely to be displaced down slope should also be fumigated to prevent introduc- and deep sands during drought periods are also tion of pathogens. Seed may be sown in either poor candidates for direct seeding. Preparing fall or spring, although fall sowing will usually the site prior to sowing, by burning, chemical, produce larger seedlings (Barnett et al. 1990). or mechanical means to reduce plant competi- However, a cold winter can damage young tion and expose mineral soil, will increase the seedlings so spring sowing may be favored probability of regeneration success. If compe- at more northerly locations. Longleaf pine tition is mostly composed of herbaceous plants seedlings should be grown at a density of 108– and low shrubs, a prescribed fire in the spring 160/m2. Seedlings produced at higher densi- or summer prior to sowing may be useful ties have lower survival following outplant- (Mann 1970). On sites with substantial compe- ing, while growing them at lower densities tition from woody plants, mechanical or her- increases the per unit cost of production. A pre- bicide treatments may be required, before fire cision sowing machine greatly aids in obtaining can be effectively used to prepare the seedbed. the desired planting density in the seedbed. 4. Longleaf Pine Regeneration Ecology and Methods 113

Irrigation and fertilization must both be health declines rapidly when kept in cold stor- carefully controlled to produce good-quality age longer than 2 weeks and survival becomes longleaf pine seedlings (Cordell et al. 1990). correspondingly low (White 1981). Too much water will promote disease, while too little will impede growth and development. The pH of irrigation water must also be Containerized Nursery controlled, maintaining it at 5–6. Substantial practical experience is needed to develop the Seedlings skill necessary to determine when seedlings re- Superior containerized longleaf pine seedlings quire watering. Fertilization decisions should are best grown outdoors in full sunlight. Dur- be based on a chemical analysis of the soil ing initial germination, a 30% shade cloth is and seedling nutrition requirements. Gen- used to prevent seed wash from containers erally, numerous smaller applications rather (Barnett and McGilvray 2000), but this ma- than fewer larger inputs of nutrients are pro- terial should be removed as soon as germina- vided, with the highest rates applied during the tion is complete. A 1:1 mixture of sphagnum phase of rapid growth. Although most large peat moss and number 2 grade vermiculite is nursery operations use soluble fertilizers ap- an optimum growing medium. Prior to mixing, plied from tank sprays or as injections into the peat moss should be screened to remove the irrigation system, dry fertilizers can also be large sticks and woody fragments. If necessary, used. Standard practices should be followed for medium pH should be adjusted to 4.5–5. Con- reducing fertilization during the seedling hard- tainers having a minimum depth of 11 cm, vol- ening phase. ume of 100 cm3, and density of less than 535 Pruning of both vertical and horizontal seedlings/m2 are recommended for growing roots improves longleaf pine seedling survival longleaf pine (Barnett and McGilvray 1997). If (Cordell et al. 1990). For production of 1-0 possible, good-quality seed with a viability of seedlings, two root-pruning treatments ap- at least 80% should be sown. Treatment with plied in August and October are recom- benomyl prior to sowing will reduce pathogens mended. Root pruning should be followed and improve germination, especially for lower with irrigation to settle the soil and reduce quality seedlots. Seeds should be sown dur- desiccation. Needle clipping just before lifting spring, from March to May, putting one ing can improve survival on dry sites (Barnett seed per cavity for good-quality seedlots or et al. 1990). The best lifting time is January two per cavity if viability is 65–80% (Bar- and early February, but lifting should be co- nett and McGilvray 2000). The seed is then ordinated with planting schedules to reduce covered with a thin 3-mm layer of growing time in storage. Care must be exercised dur- medium, grit, or vermiculite. Seedlings should ing lifting to limit root system damage. Sur- be thinned to one per cavity when seedcoats vival is improved by application of benomyl are being shed. fungicide to the packing medium or as a root Because longleaf normally germinates in coating at the time of packing (Barnett et al. the fall, it will do best if sown during cooler 1988). Grading should normally not be nec- temperatures when the daily range is 15– essary if proper seedbed densities, fertiliza- 27◦C and the mean is 22◦C. During the ger- tion, and irrigation have been used. If prob- mination phase, seedlings need frequent but lems caused by nonuniform growth arise, then light watering. Following germination, water all seedlings with a root-collar diameter less application is based on container weight. A than 7.5 mm should be discarded. Bare-root weight of 80–85% of field capacity during seedlings should be stored and transported in rapid growth and 70–75% during the hard- refrigerated containers maintained at a tem- ening phase is a good guideline for deter- perature of 1–3◦C. Seedlings are best planted mining when water is needed. Sufficient wa- as soon as possible after lifting, with the dura- ter is typically added to wet the entire plug tion in cold storage limited to 1 week. Seedling (Starkey 2002). While many nursery managers 114 II. Ecology now incorporate slow-release fertilizers into seedlings was only 51%, after 5 years (Boyer the growing medium, water-soluble fertiliz- 1989). Although containerized seedlings are ers can also be applied as needed. Starkey more costly, they are clearly the better choice (2002) recommends starting with three fertil- especially on droughty sites prone to moisture izer applications per week, adjusted as needed stress and when planting outside the normal for rainfall amounts, using a balanced N-P-K dormant-season period (Barnett 2002). fertilizer like 20-20-20 or 15-15-15. After ger- Since longleaf pine is sensitive to compe- mination is completed, use 50 ppm for 2 weeks tition, some form of site preparation is re- during the establishment phase, followed by quired before seedlings are planted. Prescribed 200 ppm for 14–16 weeks and then 25–50 burning may be all that is required if the site ppm during the hardening phase. If weeds be- has been regularly burned and is dominated come a problem, oxyfluorfen should be ap- by herbaceous species. On sites with heavy plied at about one-fourth the recommended competition from woody plants, mechanical or rate to prevent seedling damage (Barnett and herbicide treatments will be required. Inten- McGilvray 2000). sive mechanical site preparation is often a poor Watering to field capacity makes plug ex- choice for longleaf pine ecosystems because it traction easier. Poor-quality seedlings, where destroys most native understory plants, can the root plug does not hold together or root- promote rapid growth of annual weeds, and collar diameter is less than 7 mm, should be may lead to soil erosion. Herbicide and fire in culled. Good-quality seedlings are then put a combination sequence of “brown then burn” into cardboard boxes and placed in cold stor- treatments may be used to prepare harvested age at 1–3◦C. Storage intervals in excess of sites for planting. A combination of herbicide 2 weeks should be avoided, as seedling qual- and strip scalping works well on sites with na- ity will decline. Containerized seedlings can be tive understory plants. The herbicide should planted in Florida during July and August be- target woody species, while the scalp reduces cause rainfall is normally plentiful. In other ar- herbaceous competition within the planting eas, seedlings should be planted in the fall as zone. The native grasses will recover and reoc- soon as soil moisture is adequate, to reduce the cupy the scalped strip, providing the fuel nec- risk from freezing winter temperatures (Bar- essary for future management with fire (Out- nett and McGilvray 2000). calt 1995). Herbicide and scalping is also a very effective site preparation technique for oldfield sites (Hainds 2001). Planting Longleaf Pine Seedlings Longleaf pine requires careful handling and In the past, longleaf pine was often not cho- seedlings must be kept in cold storage until sen for reforestation because high mortality they are planted. For small operations, bales showed it to be very difficult to establish us- or boxes of seedlings can be brought to the ing bare-root seedlings and, even if survival planting site a couple of times each day and was acceptable, many seedlings remained in kept in the shade until planted. For large op- the grass stage for extended periods. More re- erations, it is best to have a refrigerated truck cently, however, longleaf pine has been shown present onsite. Both bare-root and container- to survive well and begin height growth quite ized seedlings can be hand or machine planted, rapidly if seedlings are planted properly. Suc- although machine planting is preferable for cessful regeneration with bare-root seedlings bare-root seedlings because of their large root requires healthy, fresh planting stock and pre- system. Bare-root seedlings should be root cision planting on a well-prepared site (Barnett pruned before lifting and not in the field. Ob- 1992). The average survival and growth of vious culls, which are too small or have poorly containerized longleaf pine, however, remains formed root systems or plugs, should be dis- better than those of bare-root seedlings. At five carded. Planting depth is a critical factor in locations in Georgia, containerized seedling survival and growth. The bud of bare-root survival was 76%, while that for bare-root seedlings should be placed at ground level or 4. Longleaf Pine Regeneration Ecology and Methods 115 slightly below to allow for soil settling (Barnett degree the ecological health of an ecosystem. et al. 1990). Shallow planting, with the bud Properly selecting and implementing forest re- above the soil surface, is better for container- production methods that contribute to numer- ized seedlings that have smaller buds more ous conservation and utilitarian goals requires prone to being buried (Hainds 2001). extensive knowledge of longleaf pine ecology, Containerized seedlings can be planted any- silviculture, and the variation in longleaf pine time from July to March, when soil mois- ecosystems across an extensive natural range. ture is adequate. Winter planting should be Although no single approach to obtaining lon- avoided in northern portions of the range be- gleaf pine regeneration can be appropriately cause cold temperatures may lead to mortality prescribed everywhere, methods that seek to from freezing. Bare-root seedlings should be holistically sustain longleaf pine ecosystems planted only during the dormant season from need to be compatible with (1) frequent use December to February, with January the best of surface fires, (2) maintaining native un- month in most areas. The planting objective derstory vegetation, (3) retaining appropriate is typically 740 well-distributed seedlings/ha, numbers of overstory trees onsite well beyond which means 1000–1200 seedlings/ha must economic maturity, and (4) creating and main- be planted, assuming a survival rate of 60– taining small-scale canopy gaps interspersed 75%. If the survival rate is lower, container- among forest patches of varying ages within an ized seedlings can be used to fill in or, if sur- uneven-aged landscape mosaic at larger scale vival is very low, the site can be burned and re- (Landers et al. 1990). planted the next season (Barnett et al. 1990). Seedling surveys, conducted 1 year after planting between October and February, are Clearcutting Method very important. Some sites may require post The clearcutting method consists of removing planting treatments to control competition all overstory and midstory trees, resulting in an and speed seedling emergence from the grass open site that is influenced little by the edge stage. of adjacent forests. The principal objective of clearcutting is to establish a new forest stand that will have an even-aged structure (Daniel et al. 1979). Clearcutting may be prescribed Forest Reproduction for removing stands in poor condition owing Methods to damage by insects, pathogens, or fire and may also be used to quickly remove an over- Concern for renewing the forest, through story composed of undesirable species prior to timely and successful regeneration, is the most reforesting a site with longleaf pine. An area significant difference between silviculture and so cleared may more easily be treated with exploitive logging. Reproduction methods de- mechanical site preparation and subsequently scribe the manner in which forests will be regenerated by artificial means. Clearcutting cut to ensure regeneration and are a com- may also be combined with natural regener- ponent of silvicultural systems that define in ation, through application of alternative-strip a more comprehensive way the manner in or progressive-strip clearcutting. However, ap- which forest stands will be tended, harvested, plication of these techniques is somewhat chal- and replaced (Daniel et al. 1979). Reproduc- lenging, considering the short dispersal dis- tion methods that are effective in achieving tance of large longleaf pine seeds and the care successful regeneration are not only important that must be taken to protect the usually scarce for the continuation of overstory tree species, advanced regeneration (Kush 2002). but also have a profound impact on composi- Certain disadvantages of the clearcutting tion and structure of the understory plant com- method suggest that it should not be applied munity which provides habitat for numerous to perpetuate existing longleaf pine forests in associated species and indicates to a substantial good condition. Clearcutting mature longleaf 116 II. Ecology pine stands can destroy much of their ad- tree method is usually to establish a new for- vanced reproduction (Boyer and Peterson est with an even-aged stand structure, use of 1983), thus requiring reliance on more ex- the deferment-cut variation will result in more pensive planted seedlings rather than estab- than one age-class present in the future forest. lished natural regeneration. Managing lon- The seed-tree method has a number of dis- gleaf pine in accordance with a “clearcut and advantages for longleaf pine. Very rarely will plant” model is very expensive, because ad- the low number of residual overstory trees be vanced regeneration is destroyed and all stand capable of producing sufficient quantities of establishment costs must be carried through a longleaf pine seed to obtain satisfactory natu- long rotation period until harvest. Although ral regeneration. Therefore, a longer time pe- longleaf pine is a competition-intolerant tree riod will typically be needed to achieve re- species, its does not require complete canopy generation success with the seed-tree method. removal to regenerate and its need for sun- The limited seed dispersal distance of longleaf light cannot be used as an appropriate ratio- pine seed requires that cleared areas be within nale for implementing the clearcutting method about 30 m of residual seed-trees (Boyer and (Brockway and Outcalt 1998). Also, complete Peterson 1983). If a good seed crop does not canopy removal through clearcutting can di- occur for a number of years, open areas can be- minish aesthetics and degrades habitat quality come occupied by competing hardwoods and for certain at-risk species. If clearcut longleaf other southern pines. The sparse longleaf pine pine sites are somehow captured by other tree overstory will produce very little needle lit- species, not only does this contribute to further ter, resulting in surface fires that may burn attrition of these endangered ecosystems, but with insufficient intensity to satisfactorily con- plant community succession on such areas will trol this competition (Kush 2002). If these in- be driven along very different trajectories that vaders escape to grow to fire-resistant size, can fundamentally alter important ecological then chemical and/or mechanical treatments processes and structures (Brockway and Lewis may be necessary (Boyer and Peterson 1983). 2003). Also, seed-trees retained onsite represent a timber volume loss to managers interested in maximizing the output of forest products Seed-Tree Method (Kush 2002). However, the greatest problem with application of the seed-tree method in The seed-tree method consists of removing existing longleaf pine forests is that most or most of the mature overstory while leaving all of the large overstory trees will be re- enough good seed-producing trees scattered moved, changing ecosystem structure and de- across a site to ensure acceptable future stock- grading habitat that supports numerous at-risk ing (Daniel et al. 1979). The 20–25 trees/ha species. typically left as residuals afford some degree of site amelioration and ensure an even dis- Shelterwood Method tribution of seed over the area (Kush 2002). The residual seed-trees can provide a more The shelterwood method consists of remov- aesthetic appearance and improve composi- ing a moderate portion of the overstory while tion of the future stand if good-quality phe- retaining onsite sufficient numbers of seed- notypes (and presumably genotypes) are se- bearing trees so that regeneration becomes es- lected. The seed-trees are typically harvested tablished under the protective partial shade of after regeneration becomes established; how- mature residual trees (Fig. 10). Compared with ever, in a variation of this method known as the relatively rigid conditions resulting from “deferment” cutting, these residuals may be application of clearcutting and seed-tree meth- retained onsite along with the regeneration ods, the shelterwood method is the most flexi- throughout the entire next rotation (Smith ble method for establishing new forests of lon- et al. 1989). While the objective of the seed- gleaf pine having an even-aged structure and is 4. Longleaf Pine Regeneration Ecology and Methods 117

FIGURE 10. Longleaf pine forest being regenerated by the shelterwood method; an overstory of 60 trees/ha having a basal area of 7 m2/ha remains following the seed cut. Photo courtesy of the Forestry Images Organization. capable of producing a variable degree of site tially transformed and habitat quality for at- amelioration (Daniel et al. 1979). When ap- risk species may remain reduced for decades. plied in longleaf pine forests, the 50–62 resid- The shelterwood method can be applied only ual trees/ha typically produce an abundance in existing stands with sufficient dominant of seed that is uniformly distributed across a and/or codominant trees of seed-bearing size site (Kush 2002). Selecting high-quality phe- (Boyer and Peterson 1983). The most com- notypes as residual trees provides an oppor- monly used approach is the uniform shelter- tunity to improve genetic composition of the wood method, which may be implemented future stand. With no large areas left barren, with either a two-cut or three-cut technique. a substantial forest canopy deposits sufficient As the name suggests, each cut in the uni- needle litter to support surface fires with inten- form shelterwood method is applied through- sities high enough to control competing woody out the entire stand resulting in a uniform plants (Kush 2002). The more substantial for- appearance and even-aged structure. If a est canopy of this method also enhances envi- forest is highly stocked and requires thin- ronmental aesthetics and improves habitat for ning or improvement cutting, the three-cut numerous wildlife species. The principal dis- technique is necessary. The first cut in the advantages of most forms of the shelterwood three-cut technique is the preparatory cut, method are related to the eventual removal which is optional in regularly tended stands of all overstory trees. During overwood har- and normally performed 10 years before the vest, excessive damage may occur to estab- planned overwood harvest. The preparatory lished longleaf pine seedlings and, following cut should leave a well-distributed population overstory removal, forest structure is substan- of dominant and codominant trees totaling 118 II. Ecology

13.8–16.1 m2/ha (Boyer and Peterson 1983). vival beneath overstory trees may be unaf- This will promote crown development and en- fected for 8 or more years, seedling damage hance cone production in the residual seed- during overwood harvest is minimized if the bearing trees. If the density of competing removal cut is conducted when seedlings are woody plants is greater than 2.3 m2/ha, her- no more than 2 years old and in the stemless bicide and/or fire treatments should be imple- grass stage (Boyer 1975, 1993a). While over- mented (Kush 2002). wood harvest may be scheduled with a reason- The second cut in the three-cut technique able amount of flexibility to meet economic and the first cut in the two-cut technique is the and other management objectives, conduct- seed cut (Croker and Boyer 1975). The seed ing the removal cut on overstories of higher cut is performed 5 years or more before the density above seedlings of increased size will planned overwood harvest to reduce the stand result in substantially greater seedling mor- density to 6.9 m2/ha, leaving well-distributed tality rates (Maple 1977). When the residual high-quality seed-bearing trees in the over- overstory density is equal to or greater than story (Kush 2002). The most desirable resid- 9.2 m2/ha, overwood harvest is best accom- uals are trees equal to or greater than 40 cm plished by two separate removal cuts. The prin- diameter at breast height that have evidence cipal postharvest need is application of pre- of past cone bearing ability. Large hardwood scribed fire to control competing plants and the and other undesirable trees may also be re- brown-spot fungus; however, timing is cru- moved during the preparatory cut, and smaller cial, since mistimed fires can harm vulnera- hardwoods can subsequently be controlled by ble seedlings (Boyer and White 1990). Pre- prescribed fire conducted during the growing scribed burning should be avoided during the season (Boyer and Peterson 1983). first 2 years following the removal cut, because Seed crops are then monitored by annual accumulated logging slash will produce fires springtime counts of flowers and conelets, too hot for newly released seedlings (Boyer watching carefully for crops that produce and Peterson 1983). Periodic burning should 1860–2470 cones/ha, the range typically re- generally not resume until dominant longleaf quired for adequate natural regeneration pine seedlings have attained a root-collar di- (Boyer and Peterson 1983). Once an accept- ameter of 7.5 mm (Boyer 1979). Prescribed able cone crop is noted, a seedbed preparation fire is indicated when surveys reveal a brown- burn can be conducted during the year prior spot fungus infection rate of equal to or greater to seedfall (Kush 2002). While prescribed fire than 20% among dominant seedlings (Boyer during summer or autumn will control woody 1993a). plant competition better than earlier cool- Variations of the shelterwood method in- season burns, increased seed predation by birds clude its application in block, strip, and group has been associated with fire during the fall configurations (Boyer and White 1990; Boyer season. Annual seedling surveys are also con- 1993a). The block approach is associated with ducted to assess the degree of regeneration suc- even-aged forests and identifies stands typi- cess, the principal criterion being a minimum cally from 4 to 40 ha as management units de- of 9880–14820 seedlings/ha that are at least lineated by various natural or artificial bound- 1 year old and well distributed across the site ary features such as creeks and roads. While (Boyer 1993a). This abundance provides suffi- each unit may be regenerated individually, cient seedling numbers to compensate for fu- several adjacent blocks could also be treated ture mortality during overwood removal and at the same time, resulting in a relatively large from other factors (Kush 2002). Once an ad- area being in the same stages of regeneration. equate number and distribution of established The progressive strip approach can convert a longleaf pine seedlings are present, the residual large even-aged stand into a number of smaller seed-bearing trees in the overstory may be har- even-aged stands covering an entire range of vested by the removal cut (Boyer and Peterson age-classes from seedling to mature. Strips 1983). Although longleaf pine seedling sur- created by cutting are long and narrow, not 4. Longleaf Pine Regeneration Ecology and Methods 119 exceeding 60 m in width, so that most or all of of regenerating seedlings establish and develop the strip will be within seed-dispersal range of during ensuing decades. With this method, the adjacent mature trees. Within each strip, overstory residuals may range in age from the uniform shelterwood method is applied as newly mature (50–100 years) to very mature described above, using either the three-step or (300–400 years), depending on the health of two-step technique. When the seed cut is made individual trees and specific management ob- in the first strip, a preparatory cut is made in jectives. Although the growth of longleaf pine the second strip. When seedlings are estab- seedlings will no doubt be slowed by compe- lished in the first strip, a removal cut is made tition from residual adult trees (Table 1), the there, with a seed cut then made in the sec- continuously maintained forest canopy is ben- ond strip and a preparatory cut made in an en- eficial in sustaining species dependent on such tirely new third strip. Thus, a “rolling wave” of structural conditions (Boyer 1993b). The ir- regeneration may proceed indefinitely across regular shelterwood method is the most rapid the landscape. The group approach is applied way to convert an even-aged longleaf pine for- to management units that are too small to be est (or plantation) to a two-aged and even- considered blocks, typically 0.2–2 ha. Using tually an uneven-aged stand structure (Boyer this approach, numerous patches are treated 1999). with either the three-step or two-step technique. When the first patches are treated with a seed cut, the second group of patches is treated with a preparatory cut. Once seedlings Uneven-Aged Silviculture are established in the first patches, they are In the most recent decade, ecosystem man- treated with a removal cut, the second patches agement policies have substantially increased are treated with a seed cut, and a third group interest in managing forests through uneven- of patches are treated with a preparatory cut. aged silviculture (Guldin 1996). The chapter by The substantial dispersion and smaller size of Guldin (this volume) discusses uneven-aged areas under treatment at any one time allow silviculture of longleaf pine forests in greater this shelterwood approach to more closely detail. Protecting the native plant community, achieve management objectives related to aes- maintaining a continuous forest canopy, and thetic and ecological values. However, even- facilitating development of large, old trees are tually removing all canopy trees still raises ob- among the desirable habitat features poten- jections from those concerned about the rar- tially resulting when uneven-aged silviculture ity of older forests and welfare of associated is applied in an adaptive ecosystem manage- species. ment framework. Interest has also grown in By contrast, the irregular shelterwood methods for converting even-aged stands to method retains a moderately light (less than uneven-aged forests, like uniform partial cut- 7m2/ha) canopy of mature trees onsite ting and patch cutting combined with thin- throughout an entire stand rotation. This ning (Nyland 2003). Whether by even-aged method may be applied using any of the above or uneven-aged methods, silviculture is fun- techniques and variations, except the removal damentally applied at varying spatial and tem- cut is mostly or entirely eliminated (Boyer poral scales to mimic disturbance and guide 1999). Unlike the uniform method, the irregu- succession in a forest. Considering the wide lar shelterwood method can maintain a resid- natural variation in frequency, intensity, and ual canopy of mature trees onsite over many area impacted by disturbance, forests tended rotations and potentially in perpetuity. Ma- according to ecosystem management princi- ture overstory residuals plus newly regener- ples might well contain a mixture of even- ated seedlings initially comprise a two-aged aged and uneven-aged stands. Where stages forest that may be maintained by thinning of advanced succession are sought and perpet- from below or allowed to eventually achieve uating those stages is desirable (and ecolog- an uneven-aged structure, as successive waves ically feasible), uneven-aged silviculture can 120 II. Ecology

TABLE 1. Inhibitory influence of residual overstory trees on stand growth during a 40-year period (1957– 1996) for longleaf pine forests treated with the uniform shelterwood (initial basal area = 0) and irregular shelterwood (initial basal area > 0) methods (tree dbh > 8.9 cm). Residual pine Ingrowth pine Total pine Initial stand basal area Basal area Volumea Basal area Volumea Basal area Volumea (m2/ha) (m2/ha) (m3/ha) (m2/ha) (m3/ha) (m2/ha) (m3/ha) Block A 0 0 0 18.2 131.9 18.2 131.9 0 0 0 17.0 125.7 17.0 125.7 2.1 2.2 18.6 11.7 70.4 13.9 89.0 4.1 7.3 62.0 6.8 30.1 14.1 92.1 6.2 8.9 74.7 2.3 6.2 11.2 80.9 8.3 12.9 109.6 0.8 4.4 13.7 114.0 10.4 15.4 126.8 0.2 1.1 15.6 127.9 Block B 0 0 0 20.1 139.9 20.1 139.9 2.1 2.5 21.4 11.8 68.7 14.3 90.1 4.2 6.3 53.7 4.6 17.6 10.9 71.3 6.2 9.3 78.8 1.8 5.2 11.1 84.0 8.3 12.7 106.9 0.4 1.5 13.1 108.4 10.4 16.1 135.8 0.1 0.2 16.2 136.0 a All volumes reported as inside bark and derived from local forest volume tables. be used to achieve such conditions (Guldin Group Selection Method 1996). Uneven-aged silviculture affords a major As an uneven-aged mosaic of even-aged advantage, in that natural regeneration is patches distributed across the landscape more or less continuous, as late succes- (Platt and Rathbun 1993), natural longleaf sional forest dynamics are emulated (Guldin pine forests maintain a continuous over- 1996). The Dauerwald (i.e., continuous for- story canopy while establishing naturally est) of Germany is a prime example of the regenerated seedlings in canopy gaps cre- uneven-aged, continuously regenerating for- ated by lightning and other local disturbance est that can result (Schabel and Palmer 1999). agents. Of all available forest reproduction Loblolly pine and shortleaf pine, growing on methods, group selection most closely mim- gentle terrain, under aggressive or limited ics this natural gap-phase regeneration pat- competition, with the worst trees cut and the tern and is perhaps best suited in the long best trees left to serve as seed-bearing residuals, term for sustaining the ecological character of have been successfully managed with uneven- longleaf pine ecosystems (Brockway and Out- aged methods (Guldin 1996). Although lon- calt 1998). The group selection method can gleaf pine was formerly thought to be too be used to create circular, elliptical, and ir- competition-intolerant for uneven-aged silvi- regularly shaped gaps ranging from 0.1 to 0.8 culture, recent evidence suggests this to be a ha distributed throughout the forest to effec- viable and desirable forest management alter- tively simulate the desired uneven-aged mo- native (Farrar and Boyer 1990; Farrar 1996; saic (Fig. 11). If volume control is preferred, Palik et al. 1997; Brockway and Outcalt 1998; the volume-guiding diameter limit (V-GDL) McGuire et al. 2001; Gagnon et al. 2003). The and basal area-maximum diameter at breast group selection method and the single-tree height-q (BDq) stand regulation procedures in selection method are the principal means by a modified group selection method are suitable which uneven-aged silviculture can be prac- for managing longleaf pine forests, when used ticed. in combination with prescribed fire on a 3-year 4. Longleaf Pine Regeneration Ecology and Methods 121

FIGURE 11. Longleaf pine regeneration freely ascending within a forest canopy gap (background); note numerous seedlings that remain suppressed by intraspecific competition from nearby adults (foreground). Photo courtesy of the Forestry Images Organization. cycle to control competing vegetation and woody competitors. If a 10-year cutting cy- maintain seedbeds in appropriate condition cle is used, the forest will be again entered (Farrar and Boyer 1990; Farrar 1996). Alterna- during year 10 to create a second series of tively, longleaf pine stands may be quite easily gaps. If longleaf pine seedlings are established regulated using an area control procedure. in the first series of gaps, they can be pro- Group selection implementation begins dur- tected by skipping one fire cycle, if necessary, ing year 0, with an entry to create the first and allowing dominant seedlings to reach a series of canopy gaps by cutting groups of more fire-resistant size. Overstory trees imped- overstory trees, placing emphasis on removing ing seedling development along the gap pe- poor-quality trees and retaining good-quality riphery may now be removed. During year 20, seed-bearing trees around the periphery of another entry is made to create a third series of each gap (Table 2). Ideally, gaps should be cut gaps in the canopy. Periodic surface fires need in areas where advanced longleaf pine regen- to be conducted throughout this decade. The eration is already present, thereby decreasing seedlings in the first series of gaps should be the likelihood that newly created gaps will be- sapling size by now and young seedlings may come occupied by competing woody species. be present in the second series of gaps. Com- Gaps should be created of sufficient size (at peting adults may be removed here as needed. least 0.1 ha) to minimize the suppressive ef- During the next entry in year 30, a fourth se- fects of intraspecific competition between lon- ries of canopy gaps is initiated. Seedlings may gleaf pine adults and seedlings. The forest will now be present in the third series of gaps and be burned several times during this decade to competing adults along the periphery may be maintain good-quality seedbeds and control removed where necessary. Seedlings in the 122 II. Ecology

TABLE 2. Schedule for implementing the Group Selection Method using a 10-year cutting cycle during the first 100 years of application. Seedlings established in Create gaps; retain openings; may remove Saplings Sawlogs cone-bearing trees adjacent size adults; Seedlings grown grown to Poles grown continue to Year along edges skip one fire cycle to sapling size pole size to sawlog increase in size gap series 0 1st 10 2nd 1st 20 3rd 2nd 1st 30 4th 3rd 2nd 1st 40 5th 4th 3rd 2nd 1st 50 6th 5th 4th 3rd 2nd 1st 60 7th 6th 5th 4th 3rd 2nd 70 8th 7th 6th 5th 4th 3rd 80 9th 8th 7th 6th 5th 4th 90 10th 9th 8th 7th 6th 5th 100 11th 10th 9th 8th 7th 6th second series of gaps should be sapling size single-tree selection method, (7) full stand reg- by now while saplings in the first gap series ulation is more easily achieved, with area con- should have grown to pole size. By year 40, a trol being an easier-to-apply procedure than fifth series of gaps will be created, seedlings volume control procedures (i.e., BDq, V-GDL), should become established in the fourth se- (8) even small areas can be economically man- ries of gaps, saplings will have developed in aged for regular, even product flow, (9) larger the third series, poles will have grown in the trees may be grown by adjusting the cut- second series, and sawlogs will have emerged ting cycle and/or maximum diameter at breast in the first series of gaps. If retained for a pro- height with no change in stand area, and (10) longed period, these sawlogs will eventually the stand is afforded a greater degree of pro- become very large trees. This continuous se- tection from disturbance-induced losses, be- rial pattern of entry, cutting, and reproduction cause regeneration is present to replace fallen in the longleaf pine forest may be continued, overstory trees (Daniel et al. 1979; Farrar along with application of periodic prescribed and Boyer 1990; Farrar 1996; Kush 2002). burning and other cultural treatments, to cre- Among the disadvantages of group selection ate a potentially infinite series of regenerat- are (1) if healthy groundcover is not maining canopy gaps that indefinitely sustain an tained through periodic prescribed burning, uneven-aged mosaic. gaps without advanced regeneration may be- The principal advantages of the group se- come occupied by woody competitors, neces- lection method include (1) high-forest cover sitating costly mechanical or herbicide treat- is constantly maintained with no degradation ments, (2) need for periodic vegetation control of habitat for at-risk species, (2) the diver- treatments other than prescribed fire (e.g., her- sity of tree size-classes results in an aesthet- bicide application every 20 years) on better ically pleasing environment, (3) regeneration quality sites with more severe woody plant is continuous and not confined to short high- competition, (3) higher management costs, risk periods, (4) regeneration develops under greater inventory information needs and lower even-aged conditions within canopy gaps, (5) product outputs relative to even-aged manage- the larger openings permit establishment of in- ment methods, (4) higher logging costs and tolerant species, (6) more concentrated har- greater logging damage of residuals compared vest results in lower logging costs, less damage with those for even-aged methods, and (5) it to seedlings and residual overstory trees, and can be difficult to apply on sites with severe easier conduct of inventories compared to the understory competition such as saw palmetto 4. Longleaf Pine Regeneration Ecology and Methods 123

(Serenoa repens) dominated flatwoods (Farrar However, since the q factor is the most dif- and Boyer 1990; Farrar 1996; Kush 2002). ficult to control (and least important of the variables), most operational-scale applications result in pine stands consisting of multiple size Single-Tree Selection Method classes, rather than strictly corresponding to In the single-tree selection method, individ- a classically balanced reversed-J distribution. ual trees are removed from the forest canopy With seed-bearing trees greater than 40 cm and regenerating seedlings become established diameter at breast height favored as residual in their place. Since only small openings are overstory trees, basal area should not exceed created by this method, it is best applied to 17 m2/ha to limit the adverse influences of tree species that are tolerant of competition for shading and root competition on regenerating site resources (Daniel et al. 1979). Individual seedlings (Shelton and Cain 2000). Applica- trees are harvested in a series of partial cuts tion of the single-tree selection method results that typically occur in a stand at 10-year inter- in a pine stand with an irregular canopy and vals (though the actual cutting cycle may be many gaps of various sizes, up to 0.1 ha. Al- somewhat shorter or longer). Application of though seedlings become established through- this method results in development and main- out stands managed under single-tree selec- tenance of an uneven-aged forest structure, tion, those in gaps created by harvest are of having a reverse-J size-class distribution that greatest interest since they have the greatest can be quantified by the diminution quotient, potential to eventually ascend to a dominant q. The q factor is derived from de Liocourt’s position in the forest canopy. law, ranges from 1.2 to 1.5, and is applicable Application of the single-tree selection to all uneven-aged forests (de Liocourt 1898; method in longleaf pine forests has received Meyer 1952). less study than other southern pines. The in- Selection was first applied in southern pines frequency of good seed years and inclina- using annual harvests for the initial 15 years, tion of seedlings to be suppressed by com- followed by periodic cuts every 5–7 years petition from nearby adults, appears to make thereafter (Reynolds 1969; Reynolds et al. longleaf pine a less suitable candidate for ap- 1984). Although this approach was neither plication of the single-tree selection method. group selection nor single-tree selection, vol- Guidance for implementing the single-tree se- ume control was successfully achieved in lection method in longleaf pine forests em- loblolly pine and shortleaf pine stands by using phasizes thinning removal of individual trees the V-GDL regulation procedure and following (especially those of poor quality) on a 10- the rule of “cut the worst and leave the best” year cutting cycle, establishment of seedlings trees (Guldin 1996). The keys to successful rein small gaps, and removal of adjacent over- generation with application of the single-tree story trees to progressively enlarge these gaps selection method in this forest type are reg- thereby releasing seedlings from competition ulation of stocking and stand structure, care- with adult longleaf pine trees (Moore 2001). ful logging, and control of competing vegeta- Although each tree is marked and harvested tion (Shelton and Cain 2000). Uneven-aged as an individual, application of this guidance stands are generally easier to create and main- eventually results in the removal of groups of tain on poor-quality sites because of less com- overstory trees, creating canopy gaps that ap- peting vegetation and greater ease of obtain- proach the dimensions of those created by the ing natural pine regeneration. Implementing group selection method. The principal differ- the single-tree selection method with the BDq ence is that two or more stand entries are re- stand regulation procedure calls for postcut- quired to create sufficiently large gaps using ting guidelines of 10–14 m2/ha for basal area, the single-tree selection method, while only 35–55 cm for maximum diameters, and a q fac- a single entry is typically needed using the tor of about 1.2 for 2.5 cm diameter at breast group selection method. Such dependence on height size classes (Shelton and Cain 2000). multiple stand entries for multiple-tree gap 124 II. Ecology creation to achieve successful regeneration, than prescribed fire on better quality sites with raises concerns regarding suitability of the more severe woody plant competition and the single-tree selection method for regenerating difficulty in efficiently applying area-wide me- longleaf pine forests. Nonetheless, a manage- chanical and chemical treatments, and (8) it is ment approach identified as “the single-tree very difficult to apply on sites where intense selection method” has been applied for sev- competition from native plants, such as saw eral decades on quail hunting plantations in palmetto, or exotic plants, especially vines like southern Georgia (Nature Conservancy 2002). kudzu (Pueraria lobata) and Japanese honey- While regeneration has been obtained on these suckle (Lonicera japonica), severely limit pine sites, the absence of an objective stand regu- regeneration success (Daniel et al. 1979; Far- lation procedure, such as V-GDL or BDq, im- rar and Boyer 1990; Farrar 1996; Shelton and pairs reliable replication of management re- Cain 2000; Kush 2002). sults among different practitioners. Since little saleable product results, interme- Advantages of the single-tree selection diate cuts in the younger-age classes necessary method include (1) forest cover and habitat for for maintaining an uneven-aged stand struc- at-risk species is constantly maintained, (2) the ture have a tendency to be neglected (Daniel variety of tree sizes creates high-quality aes- et al. 1979). Also, among the most serious thetics, (3) regeneration is continuous, (4) full dangers inherent in the single-tree selection stand regulation is easily and rapidly achieved, method is the always present temptation to vi- (5) small areas can be economically managed olate the “cut the worst trees first and leave for regular, even product flow, (6) large trees the best” rule. In not observing this discipline, can be produced by increasing the cutting cycle errant practitioners will harvest the best trees or maximum diameter at breast height with- first, thereby allowing the single-tree selection out changing the stand area, and (7) the stand method to degenerate into exploitive timber is afforded a higher degree of protection from “high-grading” (Farrar 1996). Finally, it is im- windthrow, insect, pathogen, and fire losses, portant to note that applications of the selec- since regeneration is always present to replace tion methods in longleaf pine forests are still damaged overstory trees (Daniel et al. 1979; in their infancy and flawless results cannot be Farrar and Boyer 1990; Farrar 1996; Shelton expected without substantially more scientific and Cain 2000; Kush 2002). The disadvantages study and management experience. of single-tree selection are (1) requirement of well-trained staff, highly skilled in applying more difficult stand management concepts Silviculture that Mimics Natural and procedures, (2) large amounts of stand information from inventory examinations and Disturbance growth and yield projections that are difficult The above forest reproduction methods were and time-consuming, (3) crop trees are scat- first developed by foresters in Europe during tered throughout the stand resulting in higher the eighteenth century, when the newly cre- logging costs, greater logging damage to residu- ated profession of forestry was confronted with als, and lower product outputs relative to other the challenge of repairing the extensive dam- forest reproduction methods, (4) very small age done to forests by centuries of exploitive single-tree gaps provide conditions unfavor- logging (Daniel et al. 1979). During the enable for establishment and growth of intoler- suing 300 years, various clearcutting, seed- ant longleaf pine, (5) the resulting forest envi- tree, shelterwood, and selection methods have ronment may be less suitable for some wildlife been employed by foresters in Europe and on species, such as bobwhite quail (Colinus virgini- other continents to regenerate and rehabili- anus) that prefer low vegetation and edge ar- tate many forest types. Although their princi- eas created by other methods, (6) fire dam- pal focus was often limited to renewing the forage of pine seedlings that are often present in est overstory, many early foresters can rightly the youngest and most vulnerable size-classes, be considered among the first practitioners (7) need for periodic vegetation control other of restoration. With ecologists, foresters, and 4. Longleaf Pine Regeneration Ecology and Methods 125 wildlife biologists more recently broadening lar to those caused by a powerful windstorm the focus of forest management to include (e.g., hurricane) or other large-scale distur- understory flora and fauna, appropriate for- bance (Fig. 12). However, tree removal from est reproduction methods applied through the the site is not normally associated with such practice of silviculture remain the fundamen- natural events and the magnitude and pattern tal means by which forest ecosystems will be of accompanying soil disturbance are very dif- restored and sustained. ferent (e.g., skid trails from logging rather than Despite longtime recognition of the inher- tip-up mounds from windthrow). Results of ent efficiency of working in concert with na- the seed-tree method are also like those pro- ture, silviculture has traditionally placed little duced by major windstorms, where a few large emphasis on the importance of natural dis- trees remain scattered across a site. Once again, turbance dynamics in the development and removal of trees from the site and the resulting maintenance of forests. Indeed, forest man- soil disturbance pattern are unnatural conse- agement has often implemented silvicultural quences. The shelterwood method is thought systems that prescribe stand homogeneity to to closely mimic stand replacement dynam- optimize stand-level tree growth. However, as ics following a damaging tropical storm, leav- the primary emphasis on timber production ing the forest canopy moderately intact and has shifted to also include the stewardship facilitating prolific regeneration (Croker and goals of sustaining ecological functions, con- Boyer 1975). While this may be true for the serving biological diversity, and protecting at- irregular shelterwood method (where seed- risk species, natural disturbance has become bearing trees of the overwood are retained on- recognized as a potential source of guidance site as biological legacies), eventual harvest of for forest management (Coates and Burton the overstory trees makes the uniform shel- 1997). The fundamental hypothesis for such terwood method appear less similar to natural an approach is that the species, structures, and disturbance. The two selection methods result processes that characterize natural ecosystems in stand conditions not unlike those created by are more likely to persist in managed forests small-scale disturbances (e.g., lightning, local- if anthropogenic disturbances like logging and ized fire, disease, or insect outbreaks), where prescribed burning mimic the dynamics of nat- individuals or groups of trees suffer mortality ural disturbances (Mitchell et al. 2002). The and fall from the forest canopy. The major dif- greater spatial and temporal heterogeneity of ference between these two methods is related natural disturbance regimes cause higher vari- to gap size and dispersion pattern. Group se- ation in the size and age structure of living lection normally results in canopy gaps of at trees and distribution of dead woody debris least 0.1 ha and well dispersed, while single- than typically result from application of tradi- tree selection produces smaller gaps less than tional silviculture (Franklin et al. 1997; Spies 0.1 ha (and often as small as the area under 1997). Silviculture might achieve a more nat- an individual tree crown about 0.01 ha) re- ural basis by striving to mimic this heterogene- quiring less dispersion. As additional overstory ity and leaving onsite appropriate patterns and trees are removed through time, these smaller suitable numbers of residual trees, snags, logs, gaps may coalesce, forming larger gaps difficult and other woody debris as biological legacies to distinguish from those created by the group (Mitchell et al. 2002). These legacies of the past selection method. Although tree removal from can substantially influence forest development the site and periodic soil disturbance are also following disturbance and contribute to future unnatural aspects of both selection meth- conservation of biological diversity (Palik et al. ods, the abundant biological legacies retained 2002). onsite following each harvest make both of The forest reproduction methods of tradi- these methods appear consistent with the char- tional silviculture are in some ways analo- acteristics of a more natural version of silvicul- gous to various natural disturbance events. ture that mimics natural disturbance. The clearcutting method eliminates or greatly As in numerous tropical and other tem- reduces the forest canopy, with results simi- perate forest types (Brokaw 1985; Denslow 126 II. Ecology

FIGURE 12. Naturally regenerating even-aged longleaf pine stand following Hurricane Camille. Photo courtesy of the Forestry Images Organization.

1987; Spies and Franklin 1989; Veblen 1989; vanced regeneration prior to gap formation. Liu and Hytteborn 1991; Lertzman 1992), the Gap sizes and distribution resulting from ini- importance of canopy gaps as fundamental tial harvest and subsequent stand entries will ecological structures in the natural regenera- have a substantial influence on the colonization of longleaf pine ecosystems has also been tion of individuals and development of ad- recognized (Palik et al. 1997; Brockway and vanced regeneration, therefore the type and Outcalt 1998; McGuire et al. 2001; Gagnon timing of silvicultural treatments will have im- et al. 2003). Canopy gaps are essential in main- portant impacts on forest population dynamics taining the fine-scale variability created by (Coates and Burton 1997). small-scale natural disturbances and are also Canopy gaps in longleaf pine forests nat- generated by forest reproduction methods that urally result from a variety of disturbance remove overstory trees in a patch configura- agents that operate over a wide range of spa- tion (Coates and Burton 1997). Forests hav- tial and temporal scales (Palik and Pedersen ing a wide range of gap sizes afford a diver- 1996). Disturbances that remove single trees sity of microenvironments for regeneration, ultimately create canopy gaps of suitable size with habitat conditions varying among gaps, for unimpaired growth of regeneration (about within gaps (gap centers versus gap edges), 0.14 ha) as effectively as single events that re- and beneath the forest canopy. The resulting move groups of trees, a major difference being variation in available light, soil moisture, and the longer time required for many episodes of nutrient resources will differentially affect re- single-tree removal to occur (Palik et al. 1997). generation success, as will the presence of ad- Although different pathways are followed, 4. Longleaf Pine Regeneration Ecology and Methods 127 the essential result is that similar outcomes on Growing Longleaf Pine in Containers, eds. J.P. for stand structure are eventually obtained Barnett, R.K. Dumroese, and D.J. Moorhead, (Palik et al. 2002). Thus, silviculturists may pp. 5–7. USDA Forest Service, Southern Re- validly apply either the group selection method search Station, General Technical Report SRS–56, or single-tree selection method in achieving Asheville, NC. structures representative of natural forests as Barnett, J.P., and McGilvray, J.M. 1997. Practical guidelines for producing longleaf pine seedlings long as canopy gaps of sufficient size to facili- in containers. USDA Forest Service, Southern Re- tate regeneration are created. search Station, General Technical Report SRS–14, Silviculture, because of social, economic, Asheville, NC. and other constraints, may be unable to per- Barnett, J.P., and McGilvray, J.M. 2000. Growing fectly emulate the forest ecosystem structures longleaf pine seedlings in containers. Native Plants and processes sustained by a natural distur- J. 1(1):54–58. bance regime (Palik et al. 2002). However, Barnett, J.P., and McGilvray, J.M. 2002. Guide- improvements can be obtained by mitigating lines for producing quality longleaf pine seeds. the negative effects of logging on understory USDA Forest Service, Southern Research Sta- plants, forest floor, and soil structure, and in- tion, General Technical Report SRS–52, Asheville, corporating knowledge concerning ecosystem NC. Barnett, J.P., Brissette, J.C., Kais, A.G., and Jones, disequilibrium dynamics, the importance of bi- J.P. 1988. Improving field performance of south- ological legacies and the vital role of gap-phase ern pine seedlings by treating with fungicides be- replacement into appropriately implemented fore storage. South J Appl For 12:281–285. forest reproduction methods. The application Barnett, J.P., Lauer, D.K., and Brissette, J.C. 1990. of silviculture will then result in environ- Regenerating longleaf pine with artificial meth- mental conditions that more closely approx- ods. In Management of Longleaf Pine, ed. R.M. Far- imate those produced by natural disturbance rar, pp. 72–93. USDA Forest Service, Southern and improve the likelihood that multiple Forest Experiment Station, General Technical Re- constituencies (i.e., environmental, conserva- port SO–75, New Orleans, LA. tion, utilitarian, and industrial groups) will be Barnett, J.P., Pickens, B., and Karfalt, R. 1999. Im- served as longleaf pine forests are effectively proving longleaf pine seedling establishment in the nursery by reducing seedcoat microorgan- regenerated and sustained in the future. isms. In Proceedings of 10th Biennial Southern Sil- vicultural Research Conference, ed. J.D. Haywood, pp. 339–343. USDA Forest Service, Southern Re- Acknowledgments search Station, General Technical Report SRS–30, Asheville, NC. The authors express their appreciation to Boyer, W.D. 1963. Development of longleaf pine Becky Estes for assistance in searching the lit- seedlings under parent trees. USDA Forest Ser- erature to identify numerous relevant publica- vice, Southern Forest Experiment Station, Re- tions. We are also grateful to Shibu Jose, Eric search Paper SO–4, New Orleans, LA. Jokela, Dennis Hardin, Larry Bishop, and one Boyer, W.D. 1973. Air temperature, heat sums and anonymous reviewer for comments helpful in pollen shedding phenology of longleaf pine. Ecol- improving this manuscript. ogy 54(2):420–426. Boyer, W.D. 1974a. Longleaf pine cone production related to pollen density. In Seed Yield from South- References ern Pine Seed Orchards, ed. J. Krause, pp. 8–14. Macon: Georgia Forest Research Council. Allen, R.M. 1958. A study of the factors affecting Boyer, W.D. 1974b. Impact of prescribed fire on height growth of longleaf pine seedlings. Ph.D. mortality of released and unreleased longleaf pine dissertation, Duke University, Durham, NC. seedlings. USDA Forest Service, Southern Forest Barnett, J.P. 1992. The South’s longleaf pine, it can Experiment Station, Research Note SO–182, New rise again! For People 41(4):14–17. Orleans, LA. Barnett, J.P. 2002. Longleaf pine: Why plant it? Boyer, W.D. 1975. Timing overstory removal in lon- Why use containers? In Proceedings of Workshops gleaf pine. J For 73(9):578–580. 128 II. Ecology

Boyer, W.D. 1978. Heat accumulation: An easy way Boyer, W.D. 1999. Longleaf pine: Natural regenera- to anticipate the flowering of southern pines. J tion and management. Alabama’s Treasured Forests For 76(1):20–23. 18(3):7–9. Boyer, W.D.1979. Regenerating the natural longleaf Boyer, W.D., and Peterson, D.W. 1983. Longleaf pine forest. J For 77:572–575. pine. In Silvicultural Systems for the Major Forest Boyer, W.D. 1981. Pollen production and dispersal Types of the United States, tech. comp. R.M. Burns, as affected by seasonal temperature and rainfall pp. 153–156. USDA Forest Service, Agricultural patterns. In Pollen Management Handbook, ed. E.C. Handbook No. 445, Washington, DC. Franklin, pp. 2–9. USDA Agricultural Handbook Boyer, W.D., and White, J.B. 1990. Natural regen- 587, Washington, DC. eration of longleaf pine. In Management of Lon- Boyer, W.D. 1987a. Annual and geographic varia- gleaf Pine, ed. R.M. Farrar, pp. 94–113. USDA For- tions in cone production by longleaf pine. In Pro- est Service, Southern Forest Experiment Station, ceedings of the 4th Biennial Southern Silvicultural Re- General Technical Report SO–75, New Orleans, search Conference, comp. D.R. Phillips, pp. 73–76. LA. USDA Forest Service, Southeastern Forest Exper- Boyer, W.D., and Woods, F.W. 1973. Date of pollen iment Station, General Technical Report SE–42, shedding by longleaf pine advanced by increased Asheville, NC. temperatures at strobili. For Sci 19(4):315–318. Boyer, W.D. 1987b. Volume growth loss: A hidden Brennan, L.A., and Hermann, S.M. 1994. Prescribed cost of periodic prescribed burning in longleaf fire and forest pests: Solutions for today and to- pine? South J Appl For 11:154–157. morrow. J For 92(11):34–37. Boyer, W.D. 1989. Response of planted longleaf pine Brockway, D.G., and Lewis, C.E. 1997. Long–term bare–root and container stock to site preparation effects of dormant-season prescribed fire on plant and release: Fifth–year results. In Proceedings 5th community diversity, structure and productivity Biennial Southern Silvicultural Research Conference, in a longleaf pine wiregrass ecosystem. For Ecol ed. J.H. Miller, pp. 165–168. USDA Forest Ser- Manage 96(1,2):167–183. vice, Southern Forest Experiment Station, Gen- Brockway, D.G., and Lewis, C.E. 2003. Influence eral Technical Report SO–47, New Orleans, LA. of deer, cattle grazing and timber harvest on Boyer, W.D. 1990a. Pinus palustris, Mill. longleaf plant species diversity in a longleaf pine bluestem pine. In Silvics of North America, technical coordi- ecosystem. For Ecol Manage 175(1–3):49– nators R.M. Burns and B.H. Honkala, pp. 405– 69. 412. Vol. 1, Conifers Washington, DC: USDA For- Brockway, D.G., and Outcalt, K.W. 1998. Gap-phase est Service. regeneration in longleaf pine wiregrass ecosys- Boyer, W.D. 1990b. Growing-season burns for con- tems. For Ecol Manage 106(2,3):125–139. trol of hardwoods in longleaf pine stands. USDA Brockway, D.G., and Outcalt, K.W. 2000. Restoring Forest Service, Southern Forest Experiment Sta- longleaf pine wiregrass ecosystems: Hexazinone tion, Research Paper SO–256, New Orleans, LA. application enhances effects of prescribed fire. For Boyer, W.D. 1993a. Regenerating longleaf pine with Ecol Manage 137(1–3):121–138. natural seeding. In Proceedings of the 18th Tall Tim- Brockway, D.G., Outcalt, K.W., and Wilkins, R.N. bers Fire Ecology Conference, ed. S.M. Hermann, pp. 1998. Restoring longleaf pine wiregrass ecosys- 299–309. Tall Timbers Research Station, Tallahas- tems: Plant cover, diversity and biomass following see, FL. low-rate hexazinone application on Florida sand- Boyer, W.D. 1993b. Long–term development of re- hills. For Ecol Manage 103(2/3):159–175. generation under longleaf pine seedtree and shel- Brokaw, N.V.L. 1985. Gap-phase regeneration in a terwood stands. South J Appl For 17(1):10–15. tropical forest. Ecology 66(3):682–687. Boyer, W.D. 1996. Anticipating good longleaf pine Coates, K.D., and Burton, P.J. 1997. A gap-based cone crops: The key to successful natural re- approach for development of silvicultural systems generation. Alabama’s Treasured Forests 15(3):24– to address ecosystem management objectives. For 26. Ecol Manage 99:337–354. Boyer, W.D. 1998. Long-term changes in flowering Cordell, C.E., Hatchell, G.E., and Marx, D.H. and cone production by longleaf pine. In Proceed- 1990. Nursery culture of bare-root longleaf pine ings of the 9th Biennial Southern Silvicultural Research seedlings. In Management of Longleaf Pine, ed. R.M. Conference, ed. T.A. Waldrop, pp. 92–98. USDA Farrar, pp. 38–51. USDA Forest Service, South- Forest Service, Southern Research Station, Gen- ern Forest Experiment Station, General Technical eral Technical Report SRS–20, Asheville, NC. Report SO–75, New Orleans, LA. 4. Longleaf Pine Regeneration Ecology and Methods 129

Crofton, E.W. 2001. Flora and fauna of the longleaf search Station, Miscellaneous Publication No. 9, pine–grassland ecosystem. In The Fire Forest: Lon- Tallahassee, FL. gleaf Pine Wiregrass Ecosystem, ed. J.R. Wilson. Geor- Farrar, R.M., and Boyer, W.D. 1990. Managing lon- gia Wildlife 8(2):69–77. gleaf pine under the selection system: Promises Croker, T.C. 1952. Early release stimulates cone pro- and Problems. In Proceedings of the 6th Bien- duction. USDA Forest Service, Southern Forest nial Southern Silvicultural Research Conference, eds. Experiment Station, Southern Forestry Note 79, S.S. Coleman and D.G. Neary, pp. 357–368. New Orleans, LA. USDA Forest Service, Southeastern Forest Exper- Croker, T.C. 1973. Longleaf pine cone production iment Station, General Technical Report SE–70, in relation to site index, stand age and stand den- Asheville, NC. sity. USDA Forest Service, Southern Forest Ex- Franklin, J.F., Berg, D.R., Thornburgh, D.A., and periment Station, Research Note SO–156, New Tappeiner, J.C. 1997. Alternative silvicultural ap- Orleans, LA. proaches to timber harvesting: Variable retention Croker, T.C. 1975. Seedbed preparation aids nat- harvest systems. In Creating a Forestry for the 21st ural regeneration of longleaf pine. USDA Forest Century: The Science of Ecosystem Management, eds. Service, Southern Forest Experiment Station, Re- K.A. Kohm and J.F. Franklin, pp. 111–139. Wash- search Paper SO–112, New Orleans, LA. ington, DC: Island Press. Croker, T.C. 1987. Longleaf pine: A history of man Frost, C.C. 1993. Four centuries of changing land- and a forest. USDA Forest Service, Southern Re- scape patterns in the longleaf pine ecosystem. In gion, Forestry Report R8–FR7, Atlanta, GA. Proceedings of the 18th Tall Timbers Fire Ecology Con- Croker, T.C., and Boyer, W.D. 1975. Regenerat- ference, ed. S.M. Hermann, pp. 17–43. Tall Timbers ing longleaf pine naturally. USDA Forest Service, Research Station, Tallahassee, FL. Southern Forest Experiment Station, Research Gagnon, J.L., Jokela, E.J., Moser, W.K., and Huber, Paper SO–105, New Orleans, LA. D.A. 2003. Dynamics of artificial regeneration in Daniel, T.W., Helms, J.A., and Baker, F.S. 1979. Prin- gaps within a longleaf pine flatwoods ecosystem. ciples of Silviculture. New York: McGraw–Hill. For Ecol Manage 172:133–144. Davis, V.B. 1955. Don’t keep longleaf pine seed trees Gagnon, J.L., Jokela, E.J., Moser, W.K., and Hu- too long! USDA Forest Service, Southern Forest ber, D.A. 2004. Characteristics of gaps and natural Experiment Station, Research Note 98:3, New Or- regeneration in mature longleaf pine flatwoods leans, LA. ecosystems. For Ecol Manage 187:373–380. de Liocourt, F. 1898. De l’amenagement des sap- Gilliam, F.S. ,and Platt, W.J. 1999. Effects of long- inieres. Societe Forestiere de Franche-Comte et Belfort term fire exclusion on tree species composi- Bulletin 6:396–405. tion and stand structure in an old-growth Pinus Dennington, R.W., and Farrar, R.M. 1991. Longleaf palustris (longleaf pine) forest. Plant Ecol 140:15– pine management. USDA Forest Service, South- 26. ern Region, Forestry Report R8–FR3, Atlanta, GA. Grace, S.L., and Platt, W.J. 1995. Effects of adult tree Denslow, J.S. 1987. Tropical forest gaps and tree density and fire on the demography of pregrass species diversity. Annu Rev Ecol Syst 18:431–451. stage juvenile longleaf pine (Pinus palustris Mill.). Dickmann, D.I. 1993. Management of red pine for J Ecol 83:75–86. multiple benefits using prescribed fire. North J Grelen, H.W. 1983. May burning favors survival and Appl For 10(2):53–62. early height growth of longleaf pine seedlings. Engstrom, R.T. 1993. Characteristic mammals and South J Appl For 7:16–20. birds of longleaf pine forests. In Proceedings of the Guldin, J.M. 1996. The role of uneven–aged silvi- 18th Tall Timbers Fire Ecology Conference, ed. S.M. culture in the context of ecosystem management. Hermann, pp. 127–138. Tall Timbers Research West J Appl For 11(1):4–12. Station, Tallahassee, FL. Guyer, C., and Bailey, M.A. 1993. Amphibians and Engstrom, R.T., Kirkman, L.K., and Mitchell, R.J. reptiles of longleaf pine communities. In Proceed- 2001. The natural history of the fire forest. In The ings of the 18th Tall Timbers Fire Ecology Conference, Fire Forest: Longleaf Pine Wiregrass Ecosystem, ed. J.R. ed. S.M. Hermann, pp. 139–158. Tall Timbers Re- Wilson. Georgia Wildlife 8(2):5–11, 14–17. search Station, Tallahassee, FL. Farrar, R.M. 1975. Sprouting ability of longleaf pine. Hainds, M. 2001. Scalping aids survival of longleaf. For Sci 21:189–190. Alabama’s Treasured Forests 20(3):24–27. Farrar, R.M. 1996. Fundamentals of uneven-aged Hardin, E.D., and White, D.L. 1989. Rare vascu- management in southern pine. Tall Timbers Re- lar plant taxa associated with wiregrass (Aristida 130 II. Ecology

stricta) in the southeastern United States. Nat Ar- Forestry and Wildlife Sciences, Auburn Univer- eas J 9:234–245. sity, Auburn, AL. Haywood, J.D. 2000. Mulch and hexazinone herbi- Landers, J.L. 1991. Disturbance inﬂuences on pine cide shorten the time longleaf pine seedlings are traits in the southeastern United States. In Proceed- in the grass stage and increase height growth. New ings of the 17th Tall Timbers Fire Ecology Conference, For 19:279–290. pp. 61–98. Tall Timbers Research Station, Talla- Heyward, F. 1933. The root system of longleaf pine hassee, FL. on the deep sands of western Florida. Ecology Landers, J.L., Byrd, N.A., and Komarek, R. 1990. A 14:136–148. holistic approach to managing longleaf pine com- Hilton, J. 1999. Biological diversity in the lon- munities. In Management of Longleaf Pine, ed. R.M. gleaf pine ecosystem. Alabama’s Treasured Forests Farrar, pp. 135–167. USDA Forest Service, South- 18(4):28–29. ern Forest Experiment Station, General Technical Johnson, R., and Gjerstad, D. 1998. Landscape-scale Report SO–75, New Orleans, LA. restoration of the longleaf pine ecosystem. Restor Landers, J.L., Van Lear, D.H., and Boyer, W.D. 1995. Manag Notes 16(1):41–45. The longleaf pine forests of the Southeast: Re- Johnson, R., and Gjerstad, D. 1999. Restoring the quiem or renaissance? J For 93(11):39–44. longleaf pine forest ecosystem. Alabama’s Trea- Lertzman, K.P. 1992. Patterns of gap-phase replace- sured Forests 18(4):18–19. ment in a subalpine old-growth forest. Ecology Jose, S., Merritt, S., and Ramsey, C.L. 2003. Growth, 73(2):657–669. nutrition, photosynthesis and transpiration re- Lipscomb, D.J. 1989. Impacts of feral hogs on lon- sponses of longleaf pine seedlings to light, water gleaf pine regeneration. South J Appl For 13:177– and nitrogen. For Ecol Manage 180:335–344. 181. Kantola, T.A., and Humphrey, S.R. 1990. Habitat use Liu, Q., and Hytteborn, H. 1991. Gap structure, dis- of Sherman’s fox squirrel (Sciurus niger shermanni) turbance and regeneration in a primeval Picea in Florida. J Mammal 71:411–419. abies forest. J Veg Sci 2:391–402. Kelly, J.F., and Bechtold, W.A. 1990. The longleaf Lynch, K.D. 1980. A phenotypic study of se- pine resource. In Management of Longleaf Pine, lected variable in longleaf pine. Ph.D. dissertation, ed. R.M. Farrar, pp. 11–22. USDA Forest Ser- School of Forestry, Auburn University, Auburn, vice, Southern Forest Experiment Station, Gen- AL. eral Technical Report SO–75, New Orleans, LA. Mann, W.F. 1970. Direct seeding longleaf pine. Komarek, E.V. 1968. Lightning and lightning ﬁres USDA Forest Service, Southern Forest Experi- as ecological forces. In Proceedings of the 9th Tall ment Station, Research Paper SO–57, New Or- Timbers Fire Ecology Conference, pp. 169–198. Tall leans, LA. Timbers Research Station, Tallahassee, FL. Maple, W.R. 1977. Planning longleaf pine regen- Kush, J.S., comp. 1996. Longleaf pine: A regional eration cuttings for best seedling survival and perspective of challenges and opportunities. Pro- growth. J For 75:25–27. ceedings of the 1st Longleaf Alliance Conference. Mattoon, W.R. 1922. Longleaf pine. U.S. Depart- Longleaf Alliance Report No. 1, Auburn, AL. ment of Agriculture, Bulletin No. 1061, Washing- Kush, J.S., comp. 1998. Proceedings of the longleaf ton, DC. pine ecosystem restoration symposium. Longleaf McGuire, J.P. 2001. Living on longleaf: How hu- Alliance Report No. 3, Auburn, AL. mans shaped the piney woods ecosystem. In The Kush, J.S., comp. 1999. Longleaf pine: A forward Fire Forest: Longleaf Pine Wiregrass Ecosystem, ed. J.R. look. Proceedings of the 2nd Longleaf Alliance Wilson. Georgia Wildlife 8(2):42–53. Regional Conference. Longleaf Alliance Report McGuire, J.P., Mitchell, R.J., Moser, E.B., Pecot, No. 4, Auburn, AL. S.D., Gjerstad, D.H., and Hedman, C.H. 2001. Kush, J.S., comp. 2001. Restoration and manage- Gaps in a gappy forest: Plant resources, longleaf ment of longleaf pine ecosystems: Silvicultural, pine regeneration and understory response to tree ecological, social, political and economic chal- removal in longleaf pine savannas. Can J For Res lenges. Proceedings of the 3rd Longleaf Alliance 31:765–778. Regional Conference, Longleaf Alliance Report McKee, W.H. 1982. Changes in soil fertility follow- No. 5, Auburn, AL. ing prescribed burning on Coastal Plain pine sites. Kush, J.S. 2002. Natural regeneration of longleaf USDA Forest Service, Southeastern Forest Exper- pine: Adaptations to site conditions and man- iment Station, Research Paper SE–234, Asheville, agement systems. Ph.D. dissertation, School of NC. 4. Longleaf Pine Regeneration Ecology and Methods 131

McLemore, B.F. 1977. Strobili and conelet losses Outcalt, K.W., Williams, M.E., and Onokpise, O. in four species of southern pines. USDA For- 1999. Restoring Aristida stricta to Pinus palustris est Service, Southern Forest Experiment Sta- ecosystems on the Atlantic Coastal Plain, USA. tion, Research Note SO–226, New Orleans, Restor Ecol 7:262–270. LA. Palik, B.J., and Pedersen, N. 1996. Overstory mor- Means, D.B., and Grow, G. 1985. The endangered tality and canopy disturbances in longleaf pine longleaf pine community. ENFO, Florida Conser- ecosystems. Can J For Res 26:2035–2047. vation Foundation, Inc., Winter Park, FL 85(4):1– Palik, B.J., Mitchell, R.J., Houseal, G., and Peder- 12. sen, N. 1997. Effects of canopy structure on re- Meyer, H.A. 1952. Structure, growth and drain in source availability and seedling responses in a balanced uneven-aged forests. J For 50:85–92. longleaf pine ecosystem. Can J For Res 27:1458– Mitchell, R.J., Palik, B.J., and Hunter, M.L. 2002. 1464. Natural disturbance as a guide to silviculture. For Palik, B.J., Mitchell, R.J., and Hiers, J.K. 2002. Ecol Manage 155:315–317. Modeling silviculture after natural disturbance Moore, J.H. 2001. Managing the forest and the to sustain biodiversity in the longleaf pine (Pi- trees: A private landowner’s guide to conserva- nus palustris) ecosystem: Balancing complexity tion management of longleaf pine. The Nature and implementation. For Ecol Manage 155:347– Conservancy, East Gulf Coastal Plain Ecoregional 356. Team and Southeast Conservation Science Team, Peet, R.K., and Allard, D.J. 1993. Longleaf pine– Baton Rouge, LA, Theo Davis and Sons, Inc., Ze- dominated vegetation of the southern Atlantic bulon, NC. and eastern Gulf Coast region, USA. In Proceedings Mulligan, M.K., Kirkman, L.K., and Mitchell, R.J. of the 18th Tall Timbers Fire Ecology Conference, ed. 2002. Aristida beyrichiana (wiregrass) establish- S.M. Hermann, pp. 45–81. Tall Timbers Research ment and recruitment: Implications for restora- Station, Tallahassee, FL. tion. Restor Ecol 10(1):68–76. Pessin, L.J. 1938. The effect of vegetation on the Nature Conservancy. 2002. Conserving Greenwood growth of longleaf pine seedlings. Ecol Monogr Plantation. Eye on Nature, Georgia Chapter, The 8:115–149. Nature Conservancy, Atlanta, GA. Summer Issue, Platt, W.J., Evans, G.W., and Rathbun, S.L. 1988. p. 4. The population dynamics of a long-lived conifer Nolte, D.L., and Barnett, J.P. 2000. A repellent to (Pinus palustris). Am Nat 131(4):491–525. reduce mouse damage to longleaf pine seed. Int Platt, W.J., and Rathbun, S.L. 1993. Dynamics of an Biodeterior Biodegr 45:169–174. old-growth longleaf pine population. In Proceed- Noss, R.F. 1989. Longleaf pine and wiregrass: Key- ings of the 18th Tall Timbers Fire Ecology Conference, stone components of an endangered ecosystem. ed. S.M. Hermann, pp. 275–297. Tall Timbers Re- Nat Areas J 9:211–213. search Station, Tallahassee, FL. Noss, R.F., LaRoe, E.T., and Scott, J.M. 1995. Endan- Provencher, L., Herring, B.J., Gordon, D.R., gered ecosystems of the United States: A prelim- Rodgers, H.L., Galley, K.E. M., Tanner, G.W., inary assessment of loss and degradation. USDI Hardesty, J.L., and Brennan, L.A. 2001a. Effects of National Biological Service, Biological Report 28, hardwood reduction techniques on longleaf pine Washington, DC. sandhill vegetation in northwest Florida. Restor Nyland, R.D. 2003. Even- to uneven-aged: The chal- Ecol 9:13–27. lenges of conversion. For Ecol Manage 172: 291– Provencher, L., Litt, A.R., Gordon, D.R., Rodgers, 300. H.L., Herring, B.J., Galley, K.E.M., McAdoo, Outcalt, K.W. 1995. Maintaining the native plant J.P., McAdoo, S.J., Bobris, N.M., and Hard- community during longleaf pine establishment. esty, J.L. 2001b. Restoration fire and hurricanes In Forest Research Institute, Bulletin No. 192, pp. in longleaf pine sandhills. Ecol Restor 19(2):92– 283–285. Rotorua, New Zealand. 98. Outcalt, K.W. 2000. The longleaf pine ecosystem Pyne, S.J. 1997. Fire in America: A Cultural History of the South. Native Plants J 1(1):42–44, 48– of wildland and Rural Fire. Seattle: University of 53. Washington Press. Outcalt, K.W., and Sheffield, R.M. 1996. The lon- Ramsey, C.L., Jose, S., Brecke, B.J., and Mer- gleaf pine forest: Trends and current conditions. ritt, S. 2003. Growth response of longleaf pine USDA Forest Service, Southern Research Station, (Pinus palustris Mill.) seedlings to fertilization Resource Bulletin SRS-9, Asheville, NC. and herbaceous weed control in an old field 132 II. Ecology

in southern USA. For Ecol Manage 172:281– Smith, L.F. 1955. Development of longleaf pine 289. seedlings near large trees. J For 53(4):289–290. Reynolds, R.R. 1969. Twenty-nine years of selection Smith, L.F. 1961. Growth of longleaf pine seedlings timber management on the Crossett Experimen- under large pines and oaks in Mississippi. USDA tal Forest. USDA Forest Service, Southern Forest Forest Service, Southern Forest Experiment Sta- Experiment Station, Research Paper SO–40, New tion, Research Paper 189, New Orleans, LA. Orleans, LA. Smith, L.F. 1962. Growth of longleaf pine seedlings Reynolds, R.R., Baker, J.B., and Ku, T.T. 1984. Four under large pines and oaks in Mississippi. USDA decades of selection management on the Crossett Forest Service, Southern Forest Experiment Sta- Farm Forestry Forties. Arkansas Agricultural Ex- tion, Paper 189, New Orleans, LA. periment Station Bulletin 872. Snyder, E.B., Dinus, R.J., and Derr, H.J. 1977. Ge- Schabel, H.G., and Palmer, S.L. 1999. The Dauer- netics of longleaf pine. USDA Forest Service, Re- wald: Its role in the restoration of natural forests. search Paper WO-33, Washington, DC. J For 97(11):20–25. Spies, T. 1997. Forest stand structure, composition Schmidtling, R.C. 1999. Longleaf pine genetics. In and function. In Creating a Forestry for the 21st Cen- Longleaf Pine: A Forward Look. Proceedings of the 2nd tury: The Science of Ecosystem Management, eds. K.A. Longleaf Alliance Regional Conference, Longleaf Al- Kohm and J.F. Franklin, pp. 11–30. Washington, liance Report No. 4, comp. J.S. Kush, pp. 24–26. DC: Island Press. Auburn, AL. Spies, T.A., and Franklin, J.F. 1989. Gap characteris- Schmidtling, R.C. 2001. Southern pine seed tics and vegetation response in coniferous forests sources. USDA Forest Service, Southern Re- of the Pacific Northwest. Ecology 70(3):543–545. search Station, General Technical Report SRS–44, Starkey, T.E. 2002. Irrigation and fertilization type, Asheville, NC. rate and frequency of application. In Proceedings Schmidtling, R.C., and Hipkins, V. 1998. Genetic di- of Workshops on Growing Longleaf Pine in Containers, versity in longleaf pine (Pinus palustris): Influence eds. J.P. Barnett, R.K. Dumroese, and D.J. Moor- on historical and prehistorical events. Can J For head, pp. 30–34. USDA Forest Service, Southern Res 28:1135–1145. Research Station, General Technical Report SRS– Schmidtling, R.C., Hipkins, V., and Carroll, E. 2000. 56, Asheville, NC. Pleistocene refugia for longleaf and loblolly pines. Stout, I.J., and Marion, W.R. 1993. Pine flatwoods J Sustain For 10(3/4):349–354. and xeric pine forests of the southern lower Schopmeyer, C.S., tech. coord. 1974. Seeds of Woody coastal plain. In Biodiversity of the Southeastern Plants in the United States. USDA Agricultural United States: Lowland Terrestrial Communities, eds. Handbook 450, Washington, DC. W.H. Martin, S.G. Boyce, and A.C. Echternacht, Schwarz, G.F. 1907. The Longleaf Pine Virgin Forest: A pp. 373–446. New York: John Wiley & Sons. Silvical Study. New York: John Wiley & Sons. Taylor, A.R. 1974. Ecological aspects of lightning in Seamon, G. 1998. A longleaf pine sandhill restora- forests. In Proceedings of the 13th Tall Timbers Fire tion in northwest Florida. Restor and Manage Notes Ecology Conference, pp. 455–482. Tall Timbers Re- 16:46–50. search Station, Tallahassee, FL. Shelton, M.G., and Cain, M.D. 2000. Regenerat- Veblen, T.T. 1989. Tree regeneration responses to ing uneven-aged stands of loblolly and shortleaf gaps along a transandean gradient. Ecology 70 pines: The current state of knowledge. For Ecol (3):541–543. Manage 129:177–193. Wade, D.D., and Lewis, C.E. 1987. Managing south- Shoulders, E. 1967. Fertilizer application, inherent ern grazing ecosystems with fire. Rangelands fruitfulness and rainfall affect flowering of lon- 9(3):115–119. gleaf pine. For Sci 13:376–383. Wade, D.D., and Lundsford, J. 1990. Fire as a for- Simberloff, D. 1993. Species-area fragmentation ef- est management tool: Prescribed burning in the fects on old-growth forests: Prospects for longleaf southern United States. Unasylva 162(41):28–38. pine communities. In Proceedings of the 18th Tall Wahlenberg, W.G. 1946. Longleaf pine: Its use, Timbers Fire Ecology Conference, ed. S.M. Hermann, ecology, regeneration, protection, growth and pp. 1–13. Tall Timbers Research Station, Tallahas- management. C.L. Pack Forestry Foundation and see, FL. USDA Forest Service, Washington, DC. Smith, H.C., Lamson, N.I., and Miller, G.W. 1989. Wakeley, P.C. 1954. Planting the southern pines. U. An esthetic alternative to clearcutting. J For S. Department of Agriculture, Agricultural Mono- 87(3):14–18. graph 18, Washington, DC. 4. Longleaf Pine Regeneration Ecology and Methods 133

Walker, J.L. 1993. Rare vascular plant taxa associ- Walker, L.C. 1954. Early scrub-oak control helps ated with the longleaf pine ecosystem. In Proceed- longleaf pine seedlings. J For 52:939–940. ings of the 18th Tall Timbers Fire Ecology Conference, Walker, L.C., and Davis, V.B. 1954. Forest walls reed. S.M. Hermann, pp. 105–125. Tall Timbers Re- tard young longleaf pine. USDA Forest Service, search Station, Tallahassee, FL. Southern Forest Experiment Station, Research Walker, J.L. 1995. Longleaf pine ecosystem Note 93:3, New Orleans, LA. restoration: Toward a regional strategy. USDA Walker, L.C., and Davis, V.B. 1956. Seed trees retard Forest Service, Southern Research Station, longleaf pine seedlings. J For 54(4):269. Asheville, NC and Southern Region, Atlanta, Ware, S., Frost, C.C., and Doerr, P.D. 1993. South- GA. ern mixed hardwood forest: The former longleaf Walker, J.L. 1999. Longleaf pine forests and wood- pine forest. In Biodiversity of the Southeastern United lands: Old growth under ﬁre! In The Value of Old States: Lowland Terrestrial Communities, eds. W.H. Growth Forest Ecosystems of the Eastern United States, Martin, S.G. Boyce, and A.C. Echternacht, pp. ed. G.L. Miller, pp. 33–40. Asheville: University 447–493. New York: John Wiley & Sons. of North Carolina. White, J.B. 1981. The inﬂuence of seedling size and Walker, J.L., and Boyer, W.D. 1993. An ecologi- length of storage on longleaf pine survival. Tree cal model and information needs assessment for Planters’ Notes 32(4):3–4. longleaf pine ecosystem restoration. In Silvicul- White, T.L., Harris, H.G., and Kellison, R.C. 1977. ture from the Cradle of Forestry to Ecosystem Manage- Conelet abortion in longleaf pine. Can J For Res ment, comp. L.H. Foley, pp. 138–147. USDA For- 7:378–382. est Service, Southeastern Forest Experiment Sta- Wright, H.A., and Bailey, A.W. 1982. Fire Ecology of tion, General Technical Report SE–88, Asheville, the United States and Southern Canada. New York: NC. John Wiley & Sons. Chapter 5 Plant Competition, Facilitation, and Other Overstory–Understory Interactions in Longleaf Pine Ecosystems

Timothy B. Harrington

Introduction understory interactions will provide a sound basis for prescribing treatments designed to re- Many of the stand structural characteristics store and maintain longleaf pine communities. of longleaf pine (Pinus palustris Mill.) forests Overstory trees in forest stands affect that existed prior to European colonization understory vegetation by modifying growing have been altered or lost from past disturbance conditions, either directly or indirectly. These histories (Frost this volume). For example, modifications are manifested in a variety often missing are the widely spaced, large- of ways, including consumption of growth- diameter trees, the all-aged stand structure limiting resources (i.e., light, soil water, and that included a vigorous cohort of grass-stage nutrients) and alteration of other physical longleaf pine seedlings, and the understory characteristics that impact growing conditions community composed of numerous woody (i.e., temperature, litterfall accumulation, and and herbaceous species of short stature em- fire behavior). In a similar way, understory bedded within the flashy fuels of a wiregrass vegetation can influence the growing condi- (Aristida beyrichiana Trin. & Rupr.) or bluestem tions of overstory trees, potentially affecting (Andropogon spp.) matrix. Some of these struc- their survival, stem growth, and crown mor- tural features, such as the understory commu- phology. nity, can be restored through modern silvicul- This chapter will focus on two common tural methods, vegetation management, pre- interactions in forest communities, competi- scribed fire, pine thinning, and artificial regen- tion and facilitation, and their potential influ- eration (i.e., planting or seeding) (Johnson and ences on overstory and understory responses Gjerstad this volume; Walker and Silletti this in southern pine forests, with emphasis on lon- volume). Other structural features, such as an gleaf pine. First I will discuss basic concepts of all-aged distribution of longleaf pines, must be plant interactions, including types and associ- allowed to develop over time given appropri- ated responses, and attempt to classify over- ate disturbance regimes and the presence of story and understory interactions commonly keystone species (i.e., longleaf pine and wire- observed in longleaf pine forests. Implications grass or bluestem) to “jump-start” the system. of plantation silviculture to these interactions A mechanistic understanding of overstory and will be considered. Next I will review previous

r Timothy B. Harrington USDA Forest Service, Paciﬁc Northwest Research Station, Olympia, Washington 98512.

135 136 II. Ecology research on overstory and understory interac- to abundance of the intermediary (Goldberg tions with emphasis on southern pine com- 1990). munities. Finally I will discuss overstory and Interference includes those negative plant understory interactions observed in two case interactions that result either from competi- studies conducted in longleaf pine plantations tion for limited resources or allelopathy (pro- at the Savannah River Site, a National Envi- duction of chemical toxins by one plant that ronmental Research Park near Aiken, SC. The inhibit the functions of another). Because of discussion will conclude with implications of differences in plant size or other traits, com- these interactions to restoration and mainte- petition is often asymmetrical such that one nance of longleaf pine communities. plant is negatively affected while the other The plant interactions that are the primary may show little or no signs of a response focus of this chapter imply a stand structure (Grace 1990). Resource gradients, such as spa- of at least two canopy layers, the overstory tial differences in soil nitrogen availability, and the understory. To simplify the discus- can change competitive relationships that ex- sion, I will define “overstory” as the pine trees ist among plants to favor species or individu- that comprise the upper canopy of a forest als that are most effective at resource capture stand, where the minimum height of a “tree” is (Goldberg and Miller 1990; Kalmbacher and equal to 6 m (Daniel et al. 1979). “Understory” Martin 1996). Such modifications in compet- will be defined as herbaceous species (forbs itive relationships often result in declines in and grasses) plus woody species (vines, shrubs, plant species diversity because of dominance hardwoods, and pines) that have a stem diam- by a few species. eter at breast height (1.37 m) less than 2.5 cm Table 1 provides examples of plant inter- growing under or in the proximity of overstory actions that have been observed in southern trees. Midstory layers also are possible strata pine communities, classified according to the in this hypothetical stand structure but are not conceptual framework proposed by Goldberg the focus of this chapter. (1990). The simple case of exploitation of a limiting resource has been defined as “uptake effects” of competition. “Nonuptake effects” of competition occur when an intermediary, such Concepts of Plant as litterfall from overstory trees, changes re- Interactions source availability indirectly. In southern pine communities, both uptake and nonuptake ef- Plant interactions encompass a broad variety fects of competition from overstory pines play of positive and negative relationships that can a prominent role in limiting abundance and exist when plants are grown in close proxim- species diversity of the understory (Monk ity such that they influence each other’s sur- and Gabrielson 1985; Harrington and Edwards vival, growth, or reproduction (Harper 1977). 1999; Harrington et al. 2003). These effects In this section, I will rely on the conceptual on the understory occur largely through di- framework proposed by Goldberg (1990) to rect exploitation of light, soil water, and ni- characterize plant interactions, with empha- trogen resources by the overstory, but also in- sis on competition and facilitation. Most indirectly by accumulation of overstory needle teractions between plants are indirect (i.e., litter that limits light availability to forest floor they do not physically injure one another) plants. and occur through an intermediary such as “Apparent competition” results when the in- resources, natural enemies, or plant-produced termediary is not a growth-limiting resource toxins. The net result from a plant interaction but rather it is a natural enemy, such as fire, (i.e., is it positive or negative and what is the disease, or herbivory, that is promoted by one magnitude of the effect?) is the combination plant so that the net effect of the enemy on of one plant’s effect on abundance of the in- the target plant is similar to that resulting termediary and the “target” plant’s response from exploitation competition. For example, 5. Plant Competition, Facilitation 137 a (1990). overstory interactions → overstory interactions → Hendricks and Boring 1999; Wilson et al. 1999; Hiers et al. 2003 or –) of effects (E) of the overstory on + ( Understory and injury of hardwoods, and (2) understory , (4) provides nitrogen inputs to the soil, and (5) E R N References −+− Harrington and Edwards 1999 −+− (2) Richardson and Williamson 1988 +−− Richardson and Williamson 1988 +++ (6) Hendricks and Boring 1992; +++ (7) Platt et al. 1988a plants (6) provide nitrogen inputs to the soil through to the understory. Understory et al. 1998; Anderson 2003 Harrington communities as classified according to the framework of Goldberg understory interactions) are shown as the direction → understory interactions → to abundance of the intermediary, and net effect (N) (overstory soil structure needed to maintain its water-holding capacity , or nitrogen) restricts its availability to “target” plants. vegetation chemically suppress growth of pine seedlings. Overstory 2003 and Edwards 1999; McGuire etet al. al. 2001; 2003 Harrington et al. 2003 Glitzenstein et al. 1995 and Menges 1997 increases fire intensity causing mortality of longleaf pine seedlings References ground-layer vegetation. and flowering of herbaceous species, and when understory ground fires needed to regenerate longleaf pine. needed to regenerate longleaf pine. are explained in the footnotes. −+− Palik et al. 1997; McGuire et al. 2001; Harrington −+− Monk and Gabrielson 1985; Means 1997; Harrington −+− Monk and Gabrielson 1985; Harrington et al. −+− Monk and Gabrielson 1985; Shelton 1995; +−− (1) Grace and Platt 1995; Brockway and Outcalt +++ (3) Ginter et al. 1979; Boyer and Miller 1994 +++ (4) Harrington et al. 2003 +++ (5) Streng et al. 1993; Brewer and Platt 1994; nitrogen soil water light light fire toxin soil water nitrogen fire d f e Examples of plant interactions observed in southern pine b c 1. competition (nonuptake effects) competition facilitation competition (uptake effects) Apparent competition results when (1) pine litterfall Plant consumption of a limiting resource (light, soil water ABLE The influence of overstory pines on understory vegetation Above- or belowground exudates from understory scrub Litterfall of overstory pines limits light availability to Positive facilitation results when pine litterfall (3) protects Exploitation Apparent Allelopathy Positive InteractionExploitation Intermediary E R N follow in a similar manner. Numbers in parentheses abundance of an intermediary, response (R) of the understory T a b c d e f symbiotic nitrogen fixation and (7) promote frequent promotes frequent ground fires that stimulate growth scrub vegetation decreases frequency of ground fires 138 II. Ecology accumulations of needle litter near overstory bance agent in these communities, has a dual trees can support fire intensities capable of role in regulating interactions between over- killing grass-stage seedlings of longleaf pine story and understory vegetation. It has a neg- to create a vegetation-free zone similar in ap- ative effect (apparent competition described pearance to what would occur from intense above) in directly injuring or killing understory resource competition (Grace and Platt 1995; species incapable of tolerating its influence Brockway and Outcalt 1998). (Glitzenstein et al. 1995; Grace and Platt 1995; “Allelopathy” is a direct interaction in which Brockway and Lewis 1997; Brockway and plants produce a toxin in their foliage or roots Outcalt 1998). But fire also has a positive effect that chemically alters the functions of the tar- (facilitation) in stimulating vegetative growth get plant. For example, in scrub vegetation and flowering of specific understory species communities of Florida, rosemary (Ceratiola (Streng et al. 1993; Brewer and Platt 1994; An- ericoides Michx.) and false rosemary (Conrad- derson and Menges 1997) and in maintaining ina canescens [Torr. & Gray] A. Gray) shrubs wiregrass, a critical keystone species in longleaf can limit seedling growth of bluestem, wire- pine communities (Platt et al. 1988a). grass, longleaf pine, and sand pine (Pinus clausa Overstory and understory interactions may [Chapm. ex Engelm.] Vasey ex Sarg.) when operate differently in even-aged plantations grown together in noncompetitive environ- than in natural stands because of differences in ments (Richardson and Williamson 1988). This structural attributes (Oliver and Larson 1996). allelopathic relationship fosters the develop- Even-aged plantations generally have uniform ment of a fire-tolerant community that favors spacing and size of trees, similar crown mor- survival of shrubs at the expense of fire- phologies and shapes largely dictated by spac- dependent species, such as wiregrass and lon- ing, and complete crown coverage when the gleaf and sand pines. stands are at full stocking. Stem size distri- “Facilitation” is a plant interaction in which butions are usually unimodal, and with time, one or both plants benefit from their relation- they develop increasing positive skewness in- ship with each other. For example, needle lit- dicative of stand differentiation into different ter from overstory pines is a source of nitrogen crown classes. Depending on spacing among to understory plants that may aid their survival trees, rates of crown closure and overall pro- and growth (Harrington et al. 2003). Similarly, ductivity of even-aged plantations often ex- needle litter can act as a mulch to conserve soil ceed those of natural stands. water (Ginter et al. 1979) and it can protect soil Structure of understory vegetation in even- structure needed to preserve water-holding ca- aged forest plantations is often symptomatic pacity (Boyer and Miller 1994). Greater first- of overstory structure and site history, includ- year survival of longleaf pine seedlings under ing previous land use and silvicultural treat- uncut forest versus seedlings in experimentally ments. Thus, an understory in an even-aged created gaps suggests a beneficial effect of the plantation is often uniform in plant size and overstory (McGuire et al. 2001). Note that, in species composition due to the homogeneity contrast to the common plant interaction of of the overstory and associated limitations in facilitation, “mutualism” implies an obligatory understory resource availability. At crown clo- relationship between plant species such that sure of the overstory, the stand enters the each benefits when grown in proximity to the “stem exclusion stage” during which recruit- other and each suffers when grown separately ment of new trees into the overstory ceases (Radosevich and Holt 1984). and understory abundance of shade-intolerant As indicated by the majority of the plant species and vigor of tolerant species decline interactions listed in Table 1, competition in (Oliver and Larson 1996). Invigoration of ex- southern pine communities is asymmetric isting species and recruitment of new species with the overstory having the predominant in the understory may not occur until the influence in its relationship with understory “understory reinitiation stage” when canopy vegetation. However, fire, as a common distur- coverage of the overstory begins to decline. 5. Plant Competition, Facilitation 139

The transition to the understory reinitiation inversely with pine seedling density and also stage may be delayed in even-aged plantations according to the presence (1–7 cm yr−1) ver- because their uniform spacing and size of trees sus absence (2–31 cm yr−1) of herbaceous veg- prevents large canopy gaps from forming until etation. Emergence of pine seedlings from the late in stand development. grass stage clearly was limited by availability Given these differences in structural at- of belowground resources, because light avail- tributes and rates of development, it seems ability (estimated by evaporation rates) was likely that some interactions between over- 60% of maximum intensity or greater in all story trees and understory vegetation are likely but the scrub oak plot (34% of maximum). Not to occur sooner, at greater intensity, and with surprisingly, height growth of longleaf pine increased duration in even-aged plantations seedlings in the scrub oak plot (1 cm yr−1) versus natural stands. Exclusion or suppres- was among the lowest values observed in this sion of understory species as a result of over- study. story competition probably will occur more Research on production of grazing forage rapidly than in natural stands, depending on in longleaf pine rangelands has quantified the spacing among trees. However, understory extent to which overstory trees limit abun- reinitiation may be delayed substantially be- dance and biomass of understory herbaceous cause of the prolonged uniformity of stand species. Abundance of understory vegetation structure, even with the onset of density- varied inversely with increasing density of dependent mortality of overstory trees. Al- overstory longleaf pines (Wolters 1973, 1981). though plantation silviculture can provide an Pine thinning and prescribed burning com- effective means for reestablishing stands of a binations increased forage yields in natural desired species composition and spacing, in stands of longleaf pine to about half that its conventional usage it may impede restora- observed for treeless rangeland (Grelen and tion of understory species that rely on a het- Enghardt 1973). Although these studies pro- erogeneous stand structure. Thus, a complete vide empirical evidence of the competitive ef- understanding of overstory and understory in- fects of overstory longleaf pines on understory teractions is needed to properly direct develop- vegetation, they do not identify the primary ment of longleaf pine plantations toward the mechanisms responsible for them. desired stand structure of the overstory and Much of the research on overstory and un- species composition of the understory. derstory interactions in forest communities has focused on the effects of three primary factors: shade, root competition, and litterfall. Over- Previous Research on story and understory competitive interactions have been studied effectively by trenching Overstory and Understory around experimental plots to eliminate below- Interactions ground competition for soil water and nutrients while maintaining light availability in the Early research on overstory and understory understory at nominal levels. In an early study, interactions in longleaf pine communities fo- Fricke (1904, cited in Spurr and Barnes 1992) cused on factors influencing the rate at which cut the roots of overstory Scots pine (Pinus longleaf pine seedlings exited the grass stage. sylvestris L.) in a closed canopy forest and ob- Pessin (1938) compared 3-year growth of served dramatic increases in growth of under- grass-stage longleaf pines growing at various story pines, indicating that soil moisture, and densities (2470 to 247,000 seedlings ha−1) perhaps nutrients, were limiting their growth. with or without manual removal of herba- Other trenching studies have demonstrated ceous vegetation. Scrub oaks, primarily black- similar results with eastern white pine (Pinus jack (Quercus marilandica Muenchh.) and post strobus L.) in New Hampshire (Toumey and oaks (Quercus stellata Wangenh.), were re- Kienholz 1931), loblolly pine (Pinus taeda L.) moved in all but one plot. Height growth varied in North Carolina (Korstian and Coile 1938), 140 II. Ecology grand fir (Abies grandis [Dougl. ex D. Don] fall stimulated the greatest reductions in plant Lindl.) in Montana (McCune 1986), and pon- density. Litterfall probably limited germination derosa pine (Pinus ponderosa Dougl. ex Laws) of annuals; however, the upright growth form in eastern Oregon (Riegel et al. 1992). Us- of many of the perennial species enabled them ing a different experimental approach, Riegel to shed litter, especially when grown in full and Miller (1991) demonstrated for ponderosa sun. In trenched plots, presence versus absence pine that eliminating some of the overstory of litterfall strongly limited abundance of an- competition for soil water and nitrogen via ir- nuals but it had no detectable influence on rigation and fertilization, respectively, stimu- perennials. However, in the absence of trench- lated a 36% increase in aboveground biomass ing, litterfall did not influence abundance of ei- of the understory relative to nontreated areas. ther annuals or perennials because apparently Of the three factors mentioned previously, their populations were already being regulated litterfall is perhaps the one least studied in for- by shade and root competition. est communities. Southern pines shed their Recent research on longleaf pine communi- needles throughout the year, and this accumu- ties has attempted to identify the critical factors lation forms a physical barrier that can prevent limiting pine seedling development; however, seeds from reaching mineral soil and seedlings as Harper (1977) points out, study design can from emerging into sunlight (Shelton 1995). In strongly influence experimental outcomes. For general, herbaceous species with erect, semi- example, in both natural and experimentally woody growth habits can tolerate moderate created gaps within longleaf pine forest, com- levels of litterfall (Sydes and Grime 1981a,b). petition for light was identified as the pri- Litterfall can intercept, absorb, and facilitate mary factor limiting growth of longleaf pine evaporation of rainfall before it reaches min- seedlings (Palik et al. 1997; McGuire et al. eral soil layers, but it also can act as a mulch 2001). Nitrogen availability increased with de- to reduce evaporation from the soil and as a creasing density of overstory trees, but the substrate for protecting soil structure for re- magnitude of its effect on seedling growth was tention of its water-holding capacity (Ginter secondary to light availability. Soil water avail- et al. 1979; Boyer and Miller 1994). Litter- ability did not vary in a systematic way with fall can reduce temperature fluctuations in the overstory density, perhaps because of broad surface soil layers, and it can act as both a fluctuations in rainfall. In these studies, un- source and sink of nitrogen for plant nutri- derstory vegetation either was not manipu- tion. In addition, needle litter of longleaf pines lated (McGuire et al. 2001) or it was eliminated is an important fuel component that can influ- up to 1.2 m around individual pine seedlings ence fire intensity and spread, especially where (Palik et al. 1997). Perhaps soil water availabil- it accumulates around individual trees (Grace ity did play a more prominent role in affecting and Platt 1995; Brockway and Outcalt 1998). seedling responses in these studies, but back- Monk and Gabrielson (1985) studied dynam- ground effects of understory competition pre- ics of species turnover in old-field herbaceous vented its detection, either as edge effects in communities using treatments that combined the study by Palik et al. (1997) or by varying artificial shade, presence or absence of litterfall, inversely with overstory density in the study and trenching around isolated loblolly pines. by McGuire et al. (2001). The normal progression of changes in species In another example, abundance of longleaf composition during the 2-year study was an pine natural regeneration in an uneven-aged increase in density of perennials and a de- stand did not increase substantially until the crease in density of annuals. Presence of shade distance from overstory trees exceeded 12 m, (4% of full sunlight) or litterfall accelerated the and maximum seedling densities occurred at turnover or loss of annuals but only slowed distances of 16 m or greater (Brockway and the rate of succession to perennials resulting in Outcalt 1998) (Fig. 1). Absence of strong dif- fewer species and lower densities of individual ferences in light availability and the limited size plants. Combined effects of shade and litter- of the zone of higher fire intensity from needle 5. Plant Competition, Facilitation 141

1997). Such stand structures provide a higher percentage of area with sufficient availabilities of light and nitrogen to support regeneration of longleaf pine seedlings (Battaglia et al. 2002; Palik et al. 2003). The shape and orientation of individual gaps also will influence the duration of sunlight and spatial distribution of root competition from overstory trees. Fire is an essential feature of longleaf pine forests because of their pyrogenic characteristics of needle drape, grass-stage pine seedlings, and uniformly distributed wiregrass and bluestem (Andropogon spp.) (Platt et al. FIGURE 1. Relationships of average biomass (data 1988a). Numerous studies have affirmed the from McGuire et al. 2001) and density (data benefits of fire for reducing competition from from Brockway and Outcalt 1998) of longleaf pine hardwoods and shrubs and for reducing nee- seedlings growing within a gap versus distance from the edge of mature longleaf pine forest. Maximum dle litter accumulations that can impede es- values for these variables were approximately 9 g tablishment of longleaf pine seedlings (Boyer and 6700 seedlings ha−1 for biomass and density, 1990, 1993; Glitzenstein et al. 1995; Brockway respectively. and Lewis 1997). Where wiregrass forms a dense and continuous cover under longleaf litter accumulations (up to 4 m away from indi- pines, variability in fire intensity may play a vidual trees) prompted the authors to conclude role in generating gaps in ground-layer veg- that competition for belowground resources etation that promote expansion of surviving was the primary factor limiting pine regener- species and colonization of new species, such ation within gaps. In the study by McGuire as golden aster (Pityopsis graminifolia [Michx.] et al. (2001), growth of planted longleaf pine Nutt.) (Brewer et al. 1996). Because they af- seedlings was maximized when they occurred fect burn intensity, frequency and timing of at least 18 m from overstory trees (Fig. 1). Re- prescribed fire influence the abundance, size, sults of these studies imply that gaps of radius and composition of understory species (Boyer 16 m or greater (0.08 ha or larger) are needed 1995) as well as the timing of their flower- to eliminate overstory influences. Palik et al. ing and seed production (Platt et al. 1988b; (1997) advocated gap sizes of 0.14 ha or larger Brewer and Platt 1994). Hardwood mortality (i.e., a radius of 21 m or larger for circular also varies with frequency and timing of pre- gaps) to promote growth of pine seedlings free scribed fire. Over an 18-year period, Boyer of overstory influences. However, in each of (1993) found that stand basal area of mid- these examples, the experimental approaches story hardwoods increased following winter did not adequately separate and quantify over- biennial burns (from 0.8 to 2.2 m2 ha−1), story and understory effects to determine if while it decreased following spring biennial consumption of belowground resources by un- burns (from 0.9 to 0.3 m2 ha−1). Van Lear and derstory vegetation was a key factor limiting Waldrop (1991) reported that over 80% of pine seedling development. oak (Quercus spp.) and sweetgum (Liquidambar Results of the canopy gap research on lon- styraciflua L.) rootstocks were killed by 10 an- gleaf pine indicate that group selection is an ap- nual burns, while only 50% of rootstocks were propriate method of regeneration. For a given killed by 10 biennial burns. stand basal area, forest stand structures that The research reviewed here indicates a have an aggregated distribution of overstory complex suite of factors regulates overstory trees will provide a higher percentage of area in and understory interactions in longleaf pine larger gaps than those that retain trees evenly forests. Overstory effects on understory veg- dispersed across a given area of land (Palik et al. etation can be direct, such as competition for 142 II. Ecology limited resources or physical smothering from longleaf pine plantations focused on commu- needle litter. These effects also can be indirect, nity responses (understory vegetation abun- such as spatial variation in fire intensity that dance and diversity) to thinning of over- results from variable rates of needle litter ac- story pines and control of hardwoods and cumulation. Two case studies are discussed be- shrubs with herbicides (Harrington and Ed- low to illustrate some of the overstory and un- wards 1999). These silvicultural treatments derstory interactions that occur in even-aged were selected for study because they provided plantations of longleaf pine. As discussed pre- a wide range of light, soil water, and needle viously, the uniform size and spatial distribu- litter conditions in which to study understory tion of overstory trees in even-aged plantations responses. Six 8- to 11-year-old plantations of probably create a more homogeneous environ- longleaf pine growing on sandhill sites were ment for studying effects of competition and selected having average stand basal areas of facilitation than would be expected in a natu- 9.3 and 1.1 m2 ha−1 of pines and hardwoods, rally regenerated, uneven-aged stand. respectively. Soils included loamy sands of the Blanton, Lakeland, or Troup series that were Case Studies on Overstory well drained to excessively well drained resulting in low to very low available water-holding and Understory capacities (Rogers 1990). A prescribed fire of Interactions in Longleaf moderate to high intensity was applied to each site in February 1994 and 1998. Each of the Pine Plantations six sites was divided into four treatment areas of similar size (3 to 7 ha) and one of the Two case studies are presented to illustrate following treatments was randomly assigned responses to asymmetrical competition that to each: (1) nontreated, (2) pine thinning in occur between the overstory and understory May 1994 to leave approximately half of the of longleaf pine plantations. The studies pro- original stem density, (3) control of hardwoods vide a basis for prescribing silvicultural treat- and shrubs with herbicides in 1995 (grid ap- ments aimed at restoring plant species native plication of hexazinone) and 1996 (spot treat- to longleaf pine communities. In the first study, ments of triclopyr, imazapyr, and glyphosate), community-level responses (abundance and and (4) combined treatments of pine thinning diversity) of understory vegetation were in- and hardwood and shrub control. The experi- vestigated in response to pine thinning and mental design is a randomized complete block hardwood and shrub control (Harrington and with six replications (sites) of the four treat- Edwards 1999). In the second study, a con- ments arranged as a 2 × 2 factorial. trolled experiment was established to sepa- Within each of the 24 treatment areas, rately quantify competition and needle litter 10 sample points spaced on a 40-m grid were effects of a longleaf pine overstory on fitness permanently marked for periodic vegetation and fecundity of planted populations of several measurements. At each sample point, cover perennial herbaceous species native to longleaf was estimated for each understory species by pine forests (Harrington et al. 2003). the line intercept method (Mueller-Dombois and Ellenberg 1974) in August 1994–1996 and Study I: Understory by visual estimation within 10 m2 plots in August 1998. The data were pooled by cate- Community Responses to gories of herbaceous (forbs and grasses) and Pine Thinning and woody species (hardwoods and shrubs). At the end of the 1994–1996, 1998, and 2002 grow- Hardwood and Shrub ing seasons, stem diameter at breast height Control (millimeters at 1.37 m above ground) was measured on each pine and hardwood stem Initial research at the Savannah River Site rooted within 6 m of a sample point. Height on overstory and understory interactions in (centimeters) and crown width (centimeters 5. Plant Competition, Facilitation 143 in north–south and east–west directions) also were measured annually starting in 1995 on approximately 20% of measurement trees per sample point. Periodic annual increments in individual- tree stem basal area, height, and crown width were calculated separately for pines and hardwoods. Understory vegetation and tree growth data were averaged first by sample point and then by treatment area. Data for each measurement year were subjected to analysis of variance to identify whether main effects (pine thinning or hardwood and shrub control) or their interaction were significant (α = 0.05). Multiple comparisons of means were conducted with either Bonferroni adjusted probabilities (for significant interactions) or Tukey’s test (if only main effects were significant) (Sokal and Rohlf 1981).

Understory Plant Abundance Abundance of herbaceous vegetation responded dynamically to changes in stand structure, with initial decreases following the herbicide treatments, moderate increases fol- FIGURE 2. Average cover (± standard error) of (A) lowing pine thinning, and, ultimately, large nonpine woody species (hardwoods and shrubs) increases following the combination treat- and (B) herbaceous species during 4 to 5 years af- ment (Fig. 2). In 1995, cover of both woody ter combinations of thinning of overstory pines and and herbaceous species varied significantly control of hardwoods and shrubs with herbicides. as a result of the interaction of pine thin- Letters along the x-axis indicate factors and their ning and hardwood and shrub control treat- interactions from the analysis of variance that were significant (P ≤ 0.05) for a given year of the study: ments. Of the four treatments, the combina- P = pine thinning and H = hardwood and shrub tion treatment had the lowest overall abun- control. dance of vegetation probably because activity of hexazinone herbicide (a photosynthetic inhibitor) increased as a result of the greater control was associated with increases in soil light availability in the pine-thinning treat- water throughout the entire 1996 growing ment. In 1996 and 1998, cover of under- season. Relative differences in the magnitude story hardwoods and shrubs was substan- of herbaceous cover responses to these treat- tially less in the presence versus absence of ments in 1998 indicated that light availability the herbicide treatments (Fig. 2A). Cover of was the most influential factor limiting herba- herbaceous species in 1996 and 1998 demon- ceous cover (21% absolute cover reduction), strated strong increases in response to pine although its effects were similar in magnitude thinning. In addition, herbaceous cover in to those resulting from differences in soil water 1998 was greater in the presence versus ab- availability (16% cover reduction) (Fig. 2B). sence of hardwood and shrub control. Pine Because needle litter accumulations were lim- thinning increased light availability through- ited by the prescribed burns of 1994 and 1998, out the study duration; however, it increased this factor did not appear to play a strong soil water availability only during May 1995 role in limiting development of herbaceous and 1996. In contrast, hardwood and shrub cover. 144 II. Ecology Herbaceous Species Diversity Herbaceous species density (number of species per 40 m2 sample area) in 1998 varied according to the interaction of pine thinning and hardwood and shrub control (Harrington and Edwards 1999). Pine thinning alone had a greater effect on species density (33 species) than either of the hardwood and shrub control or combination treatments (30 species), and each of these responses was greater than observed for nontreated areas (25 species). A comparison of relative differences in species density resulting from pine thinning versus hardwood and shrub control main effects indicated that only thinning (increased light availability) stimulated increased diversity of herbaceous species.

Tree Growth Overstory and understory interactions were found to operate in both directions. That is, not only did the overstory inﬂuence understory vegetation abundance and species diversity, but the understory also inﬂuenced overstory tree growth. In each of the measurement years after 1994, pine thinning and/or hardwood and shrub control treatments were associated with growth increases in stem basal area and crown width of longleaf pine trees

(Fig. 3). However, in 1995 and 1996, pine thin- FIGURE 3. Average periodic annual growth (± stan- ning was associated with reductions in height dard error) in (A) stem basal area, (B) height, and growth. Hardwood and shrub control was as- (C) crown width of individual longleaf pine trees sociated with marginal increases (P = 0.07) in 7 to 8 years after combinations of thinning of over- height growth in 1996, a year noted for sus- story pines and control of hardwoods and shrubs tained increases in soil water from this treat- with herbicides. Letters along the x-axis indicate fac- ment. However, 1998 height growth was less tors from the analysis of variance that were signif- ≤ . = in the presence versus absence of hardwood icant (P 0 05) for a given year of the study: P = and shrub control. Increased allocations of tree pine thinning and H hardwood and shrub control. growth to stem diameter and crown width at the expense of growth in height also have been observed for loblolly pine soon after thinning creases in average basal area growth of in- an 8-year-old plantation (Ginn et al. 1991). dividual hardwood trees (P ≤ 0.08; data not Such shifts in growth allocation have been shown). In 2002, pine thinning was associated attributed to tree responses associated with with a 117% increase in average height growth the capture of newly available growing space of hardwoods (P = 0.05); however, high lev- rather than those associated with “thinning els of variability among plots prevented detec- shock.” In 1995, 1996, and 1998, pine thin- tion of signiﬁcant growth increases for stem ning was associated with 67% to 91% in- basal area (P = 0.11). 5. Plant Competition, Facilitation 145

Conclusions six sites from Study I were used in Study II (i.e., areas away from the sample points in the Results of Study I indicate that availabilities hardwood and shrub control and combination of both light (via pine thinning) and below- treatments). Four 0.09-ha plots (30 m × 30 m) ground resources (via hardwood and shrub were established in each of the 13- to 15-year- control) play prominent roles in maintaining old plantations. Each plot was randomly as- abundance of herbaceous species, while per- signed a thinning to 0, 25, 50, or 100% of haps only availability of light influences diver- the average basal area of nonthinned stands sity of herbaceous species. Similarly, availabil- (20 m2 ha−1). In order to focus the research ities of both light and belowground resources on overstory effects, plots were kept free of all played similar roles in stimulating growth in- nonpine vegetation with periodic applications creases in stem basal area and crown width of of non-soil-active herbicides and hand weed- longleaf pine trees. Height growth responses to ing. treatment varied inversely with the stem and In the interior 20-m × 20-m area within crown responses. Although this research may each plot, four split-plot treatments were es- only be applicable to longleaf plantations of tablished to vary belowground resources and similar age, it demonstrates a useful approach needle litter independently of pine stocking. for studying community dynamics in response Each of the following split-plot treatments was to silvicultural treatments. randomly assigned to one of four 1.2-m × 13.7-m strips in each plot: (a) trenching plus needle litter, (b) trenching minus needle lit- Study II: Effects of Above- ter, (c) absence of trenching plus needle litter, and (d) absence of trenching minus nee- and Belowground dle litter. Trenching was conducted in October Competition and Needle 1998 with a Ditch WitchR (Perry, OK), alu- minum flashing 0.51 m wide was installed Litter from Overstory Pines along the vertical wall of the trench to an aver- on Fitness and Fecundity age depth of 0.43 m to prevent encroachment of pine roots, and the soil was replaced. In nee- of Reintroduced dle litter-present split plots, stand-produced Herbaceous Species needle litter was supplemented with monthly applications to result in a standardized rate A limitation of Study I was the confound- (7893 kg ha−1 yr−1) equal to twice that pre- ing that existed between overstory competi- dicted for nonthinned stands of 20 m2 ha−1 tion (availability of light versus belowground basal area using data from Study I. In needle resources) and needle litter because treatments litter-absent split plots, needle litter was re- were conducted at the stand level, making it moved manually each month. The experimen- impossible to separate the relative influence of tal design of Study II was a randomized, com- each factor on understory vegetation. For ex- plete block with three replications (sites) of ample, understory conditions in nonthinned the split-plot arrangement of treatments. Pine stands included growth limiting effects of com- stocking was the whole-plot factor and split- petition for both light and belowground re- plot factors were trenching and needle litter. sources, making it impossible to distinguish Fourteen perennial herbaceous species na- which had the greatest influence on herba- tive to longleaf pine forests were selected for ceous vegetation. To isolate above- and be- intensive study to represent different growth lowground components of overstory compe- forms (e.g., upright versus prostrate) of forbs tition from needle litter effects, a controlled and grasses. One herbaceous species was ran- experiment was established in 1998 accord- domly assigned to each of 10 1-m2 quadrats ing to methods described in Harrington et al. per split plot. Containerized seedlings of the (2003). Nonsampled portions of three of the 14 species were grown in a greenhouse and 146 II. Ecology planted at a fixed density (11 to 36 plants m−2, Anthaenantia villosa Responses depending on success of greenhouse propagation) in May 1999, 2000, or 2001. Container- Average survival of Anthaenantia villosa ex- ized seedlings of longleaf pine were purchased ceeded 90% in all treatments at the end of the from International Forest Company (States- second year after planting, although a slight re- boro, GA) and planted at a density of 36 plants duction was observed in the presence of nee- m−2 within a randomly assigned quadrat of dle litter (Fig. 4A). In the absence of trenching each split plot. (i.e., in the presence of root competition from Fitness of each herbaceous species was as- overstory pines), cover, number of inflores- sessed as survival (%) and visually estimated cences, and number of seeds decreased with in- cover (percentage) in October. Fecundity of creasing pine stocking at a more rapid rate than each herbaceous species was assessed as in- in the presence of trenching (Fig. 4B–D). Cover florescence count (number of flower clusters in trenched plots averaged about 80% and it per square meter), seed production (num- did not vary among pine stockings. Cover was ber of seeds produced per square meter; es- greater in the presence of needle litter, an ef- timated from subsamples of 20 or more infect likely attributable to nitrogen inputs as- florescences), seed weight (grams per 1000 sociated with this treatment (Harrington et al. seeds), and seedling emergence (percentage) 2003). Seed weight did not vary significantly from standard germination tests (45 days of among treatments (Fig. 4E). Seedling emer- cold stratification followed by 3 weeks of ger- gence was less at 0% pine stocking than at mination in a greenhouse). In fall 2001, sur- 25, 50, or 100% stockings (Fig. 4F). Although viving longleaf pine seedlings growing in their plant size, flowering, and seed production each assigned quadrat were severed, counted, and were greatest in the absence of overstory pines returned to the laboratory for biomass esti- (0% stocking), seed viability was not. Appar- mation. Fitness of longleaf pine was assessed ently some level of overstory competition stim- as survival (percentage), average stem height ulated increased seed viability; thus, total re- (centimeters), average biomass (grams), and productive effort (i.e., the product of seed pro- percentage of surviving trees that were exiting duction and seedling emergence) was greatest the grass stage, defined as those with stems 12 at 25% pine stocking. A. villosa had an interac- cm or greater in height (Haywood 2000). tive response pattern in which its fitness and To highlight the basic findings of Study II, fecundity were maximized at specific combi- this chapter covers responses of two perennial nations of above- and belowground competi- grasses (green silkyscale, Anthaenantia villosa tion. The species was able to survive in a wide [Michx.] Beauv., and pineywoods dropseed, range of competitive environments, it pro- Sporobolus junceus [Michx.] Kunth, planted duced the most vegetation when belowground May 1999 and assessed in October 2000), a resources were not limiting, and it produced perennial forb (beggar’s ticks, Desmodium cil- the most viable seed in a mildly competitive iare [Muhl. Ex Willd.] DC, planted May 2001 environment. and assessed October 2001), and seedlings of longleaf pine (planted December 1998 and assessed October 2001). Analysis of variance was Sporobolus junceus Responses conducted to identify significant (α = 0.05) In contrast to Anthaenantia villosa, Sporobolus treatment effects. Multiple comparisons of junceus displayed a very different set of re- means were conducted with either Bonferroni sponses. Average survival of this species ex- adjusted probabilities (for significant interac- ceeded 60% for all treatments, but it decreased tions) or Tukey’s test (if only main effects were in the presence of either trenching or needle significant) (Sokal and Rohlf 1981). Linear re- litter (Fig. 5A). Needle litter most likely caused gression was used to quantify the relationship shading and eventual death of suppressed in- between average aboveground biomass and dividuals. Cover of S. junceus did not vary average height of longleaf pine seedlings. significantly among any of the treatments, 5. Plant Competition, Facilitation 147

FIGURE 4. Average indices of fitness and fecundity (± standard error) of the perennial grass Anthaenantia villosa as influenced by stocking of overstory pines (basal area as a percentage of nonthinned stands) and presence or absence of trenching and needle litter. Letters in the lower left corner of each graph indicate factors and their interactions from the analysis of variance that were significant (P ≤ 0.05): S = pine stocking, T = trenching, and N = needle litter. suggesting that the species was relatively tol- In the presence of needle litter, inflorescence erant to the experimental range of resource number was less than a third of that ob- availabilities and needle litter levels (Fig. 5B). served in its absence (Fig. 5C). In the ab- For the trenching effects on survival, increases sence of trenching, inflorescence number at in belowground resources might have stim- 100% pine stocking was less than at the ulated greater amounts of density-dependent other stocking levels; however, in the pres- mortality (i.e., “self-thinning”). This explana- ence of trenching, inflorescence number did tion was supported by the fact that, although not vary significantly among stocking levels. survival declined slightly from trenching and Seed number at 100% stocking was less than needle litter effects, surviving individuals oc- at the other stocking levels (Fig. 5D). Pres- cupied the newly available growing spacing ence of needle litter greatly limited seed num- resulting in the absence of treatment effects ber, seed weight (Fig. 5E), and seedling emer- on cover. These fitness responses contrasted gence (Fig. 5F). S. junceus demonstrated a fit- sharply with the species’ fecundity responses. ness response that was relatively tolerant of 148 II. Ecology

FIGURE 5. Average indices of fitness and fecundity (± standard error) of the perennial grass Sporobolus junceus as influenced by stocking of overstory pines (basal area as a percentage of nonthinned stands) and presence or absence of trenching and needle litter. Letters in the lower left corner of each graph indicate factors and their interactions from the analysis of variance that were significant (P ≤ 0.05): S = pine stocking, T = trenching, and N = needle litter. competition from overstory pines. The species able to thrive vegetatively in the wide range was able to recover growing spacing through of growing conditions present in the study. increases in plant size and fully occupy the site Observed increases in nitrogen concentration regardless of density-dependent mortality or of S. junceus foliage in the presence of nee- needle litter. However, fecundity was greatly dle litter (Harrington et al. 2003) suggested impacted primarily by needle litter (via re- that improvements in plant nutrition might ductions in flowering, seed production, seed have stimulated vegetative growth at the ex- size, and seed viability) and secondarily by pense of reproductive allocation. However, a stocking (via reductions in flowering and seed more likely explanation for the fecundity reproduction). The mechanism underlying the sponses could be the shading effects of nee- strongly negative effects of needle litter on fe- dle litter. Fecundity responses to stocking in- cundity was not clear since similar effects on dicated that a threshold in light availability plant fitness were not apparent until October existed below which flowering and seed pro- 2001 (Harrington et al. 2003). The species was duction were strongly curtailed. This threshold 5. Plant Competition, Facilitation 149

FIGURE 6. Average indices of fitness and fecundity (± standard error) of the perennial forb Desmodium ciliare as influenced by stocking of overstory pines (basal area as a percentage of nonthinned stands) and presence or absence of trenching and needle litter. Letters in the lower left corner of each graph indicate factors and their interactions from the analysis of variance that were significant (P ≤ 0.05): S = pine stocking, T = trenching, and N = needle litter. occurred between 50% and 100% stocking of suggesting that the species was able to respond overstory pines, because at lesser stockings, to a variety of growing conditions. Although responses of these variables were relatively survival of D. ciliare declined somewhat with stable. increasing pine stocking, these responses were not statistically significant (Fig. 6A). Cover decreased systematically with increasing pine Desmodium ciliare Responses stocking and it was especially limited at 100% Desmodium ciliare responses to overstory condi- pine stocking (Fig. 6B). Trenching effects on tions were similar to those observed for other cover were marginally significant (P = 0.06), perennial forbs in that they indicated a strong indicating a mild additive effect resulting degree of intolerance to both shade and root from increased availability of belowground re- competition from overstory pines (Harring- sources. Flowering and seed production re- ton et al. 2003). In general, these experimen- sponses were similar to those for cover (Fig. tal effects were additive and not interactive, 6C,D): a decline with increasing stocking and 150 II. Ecology a small to moderate increase from trenching. Seed weight did not vary significantly among treatments (Fig. 6E). The interaction of stocking and trenching was significant for seedling emergence because increases from trenching were observed only at 50% stocking; all other stockings demonstrated neutral effects from trenching (Fig. 6F). In addition, there was less seedling emergence in the presence versus absence of needle litter. These results suggested that seed viability of D. ciliare was regulated largely by light availability determined by either or both of overstory density and needle litter. In summary, D. ciliare had a strong requirement for light to survive, grow, and reproduce; its potential responsive- ness to enhanced belowground resources was secondary to that resulting from increased light availability.

Pinus palustris Seedling Responses Pinus palustris was able to survive but grew very little in response to shade and root competition from overstory pines. In the third year after planting (2001), seedling survival was greater in the presence versus absence of trenching, and it was reduced in the presence versus absence of needle litter (Fig. 7A). Stocking of overstory pines did not signifi- FIGURE 7. Average (A) survival, (B) stem height, cantly influence either of these survival re- and (C) percentage of survivors exiting the grass sponses, suggesting that light availability was stage (± standard error) for seedlings of the tree Pi- not a limiting factor for survival except when nus palustris as influenced by stocking of overstory smothering from needle litter occurred. The pines (basal area as a percentage of nonthinned fact that stocking had no detectable effect stands) and presence or absence of trenching and needle litter. Letters in the lower left corner of on survival of P. palustris seedlings probably each graph indicate factors and their interactions was the result of adequate carbohydrate stor- from the analysis of variance that were significant age in the taproot of the nursery-grown (P ≤ 0.05): S = pine stocking, T = trenching, and N seedlings that permitted the plant to toler- = needle litter. ate severe growth-limiting conditions for several years. However, it remains uncertain About 97% of the variation in average how much longer these seedlings could have aboveground biomass (grams) of longleaf pine sustained themselves under these growing seedlings was explained by its linear relation- conditions. The moderate degree of survival ship with average height (centimeters) (R2 = (59% to 68%) observed in the absence of 0.97; sy.x = 12.2; n = 48): overstory competition and needle litter sug- Y =−2.755 + 6.090(X) gests that density-dependent mortality had occurred. Because of this high correlation, height and 5. Plant Competition, Facilitation 151 aboveground biomass had essentially identi- belowground competition and needle litter. cal responses to the interaction of stocking and However, of all the species’ responses, only trenching (only height responses are shown one indicated a positive effect of overstory in Fig. 7B). Both of these variables exhibited shade (facilitation): seed viability of A. villosa steep declines with increasing stocking but the was greater at pine stockings of 25% to 100% steepness of the decline was much greater in than it was at 0%. Therefore, a primary finding the absence versus presence of trenching. This from Study II is that the growing conditions in result indicated a strong threshold response gaps (i.e., full sunlight) provided the resources to light availability because average height needed to maximize fitness and fecundity of (and aboveground biomass) declined precipi- reintroduced herbaceous species. In addition, tously as stocking increased from 0% to 25% root competition had a critically negative ef- with subsequent increases in stocking causing fect on most of the species’ responses, indi- only slight decreases in growth. The thresh- cating that restoration plantings will be most old growth response to light availability was successful in the absence of both overstory moderated only slightly by increases in be- and understory vegetation. The vigorous re- lowground resources from trenching. By the sponses of herbaceous species observed in the third year after planting, over half of long- relatively small gaps (30 m × 30 m or 0.09 leaf pine seedlings growing in 0% stocking ha) of Study II indicated that the zone of over- were exiting the grass stage, while less than story root competition from neighboring 13- 20% of seedlings were doing so in stockings to 15-year-old pines extended less than 10 m of 50 and 100% (Fig. 7C). In contrast to the into the gap. Surprisingly, plant responses to average height responses, trenching strongly needle litter varied from positive effects (facil- increased the percentage of trees exiting the itation via nitrogen addition) to negative ef- grass stage for stockings of 25 and 50%. This fects (apparent competition via reduced light result is in agreement with previous research availability). Mulching effects from needle lit- in which increased availability of belowground ter (i.e., increased retention of soil water) were resources, such as resulting from control of not detected (Harrington et al. 2003). competing herbaceous vegetation with herbicides or mulching (Haywood 2000), enabled a higher percentage of longleaf pine seedlings to exit the grass stage. However, Study II demon- Implications to strated that light availability had a much Maintenance and greater effect than availability of belowground resources since very few of the seedlings grow- Restoration of Longleaf ing in trenched split plots at 100% stocking Pine Communities were exiting the grass stage. In terms of providing new saplings of P. palustris to ultimately Previous research and the two case studies re- replace overstory trees, only the lowest stock- viewed here illustrate the multiple interactions ings (0 and 25%) averaged more than 20% that occur between overstory pines and un- of seedlings emerging from the grass stage to derstory vegetation of longleaf pine forests. As support their growth into the upper canopy. indicated initially in this chapter, a mechanistic understanding of overstory and under- Conclusions story interactions will provide a sound basis for prescribing treatments designed to restore Results from Study II confirm some of the find- and maintain longleaf pine communities. In ings from Study I: effects of overstory competi- this concluding section I will discuss the im- tion for light and for belowground resources on plications of overstory and understory compe- understory vegetation can be of similar mag- tition and facilitation to silvicultural regimes nitude. Study II further illustrates the great that are currently in use or being considered variability in species’ responses to above- and for longleaf pine forests. Three conditions will 152 II. Ecology be considered: new longleaf pine communities; of herbaceous will rely on absence of a se- existing herbaceous communities within natu- vere frost and presence of adequate rainfall ral, uneven-aged longleaf pine stands; and ex- (Harrington et al. 2003); winter planting of isting herbaceous communities within even- dormant, containerized herbaceous seedlings aged, planted longleaf pine stands. is an untested approach that may avoid these pitfalls of spring weather. Each ground-layer gap will have sufficient New Longleaf Pine availabilities of light and soil water, at least during the first growing season, to enable es- Communities tablishment of the new seedlings. Alterna- Restoring all of the elements of a longleaf tively, specific soil-active herbicides that are pine community requires coordinated estab- safe for pines and that provide residual con- lishment of longleaf pine seedlings and herba- trol of competing vegetation (e.g., hexazinone ceous species. Both vegetation components re- or imazapyr) can be used specifically to create quire an environment with an abundance of ground-layer gaps in which to plant longleaf direct sunlight and soil water. Within existing pine seedlings. Herbaceous species susceptible forest, gaps 21 m in radius or larger provide to these herbicides can be established 1–2 years such an environment. Recently harvested for- later after soil concentrations of such herbi- est sites or old fields will have adequate light cides have declined sufficiently. availability but not necessarily adequate soil Nitrogen is probably the nutrient most limit- water availability because of a potentially high ing to establishment and growth of new herba- abundance of competing vegetation. In order ceous and pine seedlings. However, fertilizer to increase soil water availability, competing should only be applied directly to the plant- vegetation abundance must be reduced signif- ing hole to avoid stimulating growth and po- icantly, at least in the individual spots where tential dominance of other, more aggressive seedlings are to be established. plant species, and the prescribed fertilizer rate Spot (or banded) treatments provide a should be safe for seedling roots. means of reducing competition without the Fire is a key ingredient for maintenance of widespread disruption in the ground layer new plantings, because it will suppress or kill community that can result from broadcast nonadapted plant species and allow adapted treatments (Brockway and Outcalt 2000). species to flourish. Field observations from In addition, spot treatments enable existing Study II indicated that, by the end of the desirable species growing between spots to first or second year after planting, most peren- be retained in an undisturbed condition. One nial herbaceous species native to longleaf pine potential technique is to create gaps in the communities can withstand dormant season ground layer vegetation 1 meter in radius fires. or larger by chemical (e.g., nonsoil active herbicides, such as glyphosate or triclopyr) or mechanical methods (e.g., “Wombat” spot Existing Herbaceous cultivator, Savannah Forestry Equipment, Communities within Natural, Savannah, GA). Uneven-Aged Longleaf Pine By locating these ground-layer gaps at a Stands spacing of 3 m or greater, the desired community components can be established over In this forest community, many of the stand a broad area. Half of the gaps can be planted structural components already exist, and only with a single longleaf pine seedling and no enrichment plantings are needed to reestab- other vegetation, while perennial herbaceous lish the desired native herbaceous species. species can be established in the remaining Understory light availability should be ade- gaps by sowing seeds or planting containerized quate, although not optimum, if the uneven- seedlings. The success of a spring planting aged structure of the stand is composed of 5. Plant Competition, Facilitation 153 widely spaced individual trees (spacing of 11 trees to allow pine seeds to germinate in ar- m or greater; Platt et al. 1988a). If light avail- eas of minimal competition and accumulation ability is potentially limiting to longleaf pine of needle litter from overstory trees. As dis- seedlings and their associates, excess over- cussed previously, root competition from overstory trees should be harvested or girdled (to story pines can create zones that exclude nat- create wildlife habitat). As discussed previ- ural pine regeneration, and areas within 4 m ously, tree removals that result in aggregated, of overstory pines have the potential for in- rather than dispersed, stand structures provide tense ground fires capable of killing young pine a higher percentage of growing space with high seedlings. light availability. Additional canopy layers of Reinstatement of prescribed fires in the dor- hardwoods and shrubs cause further limita- mant season or spring is essential for main- tions to understory light availability, and their taining the vigor of the new plantings and to abundance should be reduced significantly by avoid overtopping by resprouting hardwoods spot treatments that do not endanger desir- and shrubs. able herbaceous species, such as injection with non-soil-active herbicides, girdling, or cutting. Inherent limitations in availability of below- Existing Herbaceous ground resources, primarily water and nitro- Communities within Planted, gen, are typical of longleaf pine communities because often they occur on sandy soils with Even-Aged Longleaf Pine Stands low water-holding capacities and limited ni- Extensive plantations of longleaf pine have trogen retention. Because the pools of below- been established on public lands and, as part ground resources can be in such short supply, of the Conservation Reserve Program, on for- full occupancy of belowground resources by mer agricultural land. Often these plantations overstory pines may occur at stem densities were established at spacings typical for loblolly that do not provide complete crown coverage. pine (e.g., up to 1800 trees ha−1). As men- Therefore, moderate densities of longleaf pine tioned previously, belowground resources may can be highly competitive for belowground not be adequate to support long-term develop- resources. Soil water availability in this com- ment of longleaf pine established at close spac- munity also will be limited by understory veg- ings, especially on sandhill sites. Under these etation. Although this recommendation con- conditions, mortality from bark beetles (Den- flicts with the findings of Palik et al. (1997) droctonus spp. and Ips spp.) is likely, as has and McGuire et al. (2001) in which soil wa- been observed recently in nonthinned stands ter availability did not vary systematically with of Study I. To foster long-term health of the overstory density, I hypothesize that if over- forest community, densely stocked plantations story and understory components were ma- can be thinned to wide spacings (5 m or more) nipulated independently in natural stands of prior to engaging in understory restoration. longleaf pine, their separate and significant Within the widely spaced stands, gaps 21 m in competitive effects on soil water could be de- radius or larger can be established for under- tected, as found in the longleaf pine planta- story restoration. Conversely, an aggregated tions of Study II. stand structure can be created to maintain As discussed previously, spot (or banded) some areas of higher stocking for timber pro- treatments can be used to create ground-layer duction while providing open areas for under- gaps in areas at least 21 m from overstory pines. story restoration. If advanced regeneration of longleaf pine is Because of uniformity of spacing and size needed, mineral soil seedbeds can be created of trees, plantations will have regions of re- to foster germination and seedling develop- stricted light availability even after the preced- ment in years of adequate seed crops. Some of ing stocking guidelines have been followed. To the ground-layer gaps can be located approxi- provide adequate light and soil water avail- mately 16 m from potential longleaf pine seed ability, ground-layer gaps can be established 154 II. Ecology away from overstory trees and on the north applied properly. Use all pesticides selectively side of overstory gaps. Results from Study II and carefully. Follow recommended practices suggest that the critical zone of root com- for the disposal of surplus pesticides and pesti- petition extends less than 10 m away from cide containers. 13- to 15-year-old longleaf pines, unlike the 18-m zone observed for mature trees (Brock- way and Outcalt 1998). Note that restrictions References in light availability in longleaf pine plantations are likely to become more intense with de- Anderson, R.C., and Menges, E.S. 1997. Effects of creasing space among trees and increasing site fire on sandhill herbs: Nutrients, mycorrhizae, quality. and biomass allocation. Am J Bot 84:938–948. Battaglia, M.A., Mou, P., Palik, B., and Mitchell, Based on Study II, reintroduced herbaceous R.J. 2002. The effect of spatially variable over- and longleaf pine seedlings will have a wide story on the understory light environment of an range of fitness and fecundity responses in open-canopied longleaf pine forest. Can J For Res the understory of a longleaf pine plantation. 32:1984–1991. Responses are likely to range from increases Boyer, W.D. 1990. Growing-season burns for con- in vegetative growth at the expense of re- trol of hardwoods in longleaf pine stands. USDA productive performance to gradual declines in Forest Service, Southern Forest Experiment Sta- response to needle litter accumulations. This tion, Research Paper SO-256. diversity of plant responses indicates that a di- Boyer, W.D. 1993. Season of burn and hardwood versity of growing conditions would be needed development in young longleaf pine stands. Pro- to maintain a full gamut of species in the un- ceedings of the Seventh Biennial Southern Silviculture Research Conference pp. 511–515. derstory. Applications of prescribed fire, vege- Boyer, W.D. 1995. Responses of groundcover un- tation management, and pine thinning can be der longleaf pine to biennial seasonal burning and scheduled at appropriate times to create pock- hardwood control. Proceedings of the Eighth Biennial ets of enhanced resource availability that foster Southern Silviculture Research Conference pp. 512– flowering, seed production, and recruitment 516. of new seedlings. In time and given appro- Boyer, W.D., and Miller, J.H. 1994. Effect of burning priate maintenance treatments, reintroduced and brush treatments on nutrient and soil phys- plant species are likely to form stable popula- ical properties in young longleaf pine stands. For tions that are self-perpetuating. Ecol Manage 70:311–318. Brewer, J.S., and Platt, W.J. 1994. Effects of fire season and herbivory on reproductive success in a clonal forb, Pityopsis graminifolia. J Ecol 82:665– USDA/Forest Service 675. Brewer, J.S., Platt, W.J., Glitzenstein, J.S., and Disclaimers Streng, D.R. 1996. Effects of fire-generated gaps on growth and reproduction of golden aster (Pity- The use of trade or firm names in this publi- opsis graminifolia). Bull Torrey Bot Club 123:295– cation is for reader information and does not 303. imply endorsement by the U.S. Department of Brockway, D.G., and Lewis, C.E. 1997. Long-term Agriculture of any product or service. This pub- effects of dormant-season prescribed fire on plant lication reports research involving pesticides. community diversity, structure and productivity It does not contain recommendations for their in a longleaf pine wiregrass ecosystem. For Ecol use, nor does it imply that the uses discussed Manage 96:167–183. Brockway, D.G., and Outcalt, K.W. 1998. Gap-phase here have been registered. All uses of pesti- regeneration in longleaf pine wiregrass ecosys- cides must be registered by appropriate state tems. For Ecol Manage 106:125–139. or federal agencies, or both, before they can be Brockway, D.G., and Outcalt, K.W. 2000. Restoring recommended. CAUTION: Pesticides can be in- longleaf pine wiregrass ecosystems: Hexazinone jurious to humans, domestic and wild animals, application enhances effects of prescribed fire. For and desirable plants if they are not handled or Ecol Manage 137:121–138. 5. Plant Competition, Facilitation 155

Daniel, T.W., Helms, J.A., and Baker, F.S. 1979. Prin- ecosystems of the southeastern United States. For ciples of Silviculture, 2nd ed. New York: McGraw– Ecol Manage 113:167–177. Hill. Hiers, J.K., Mitchell, R.J., Boring, L.R., Hendricks, Ginn, S.E., Seiler, J.R., Cazell, B.H., and Kreh, J.J., and Wyatt, R. 2003. Legumes native to lon- R.E. 1991. Physiological and growth responses gleaf pine savannas exhibit capacity for high N2- of eight-year-old loblolly pine stands to thinning. fixation rates and negligible impacts due to timing For Sci 37:1030–1040. of fire. New Phytol 157:327–338. Ginter, D.L., McLeod, K.W., and Sherrod, C., Jr. Kalmbacher, R., and Martin, F. 1996. Shifts in 1979. Water stress in longleaf pine induced by botanical composition of flatwoods range follow- litter removal. For Ecol Manage 2:13–20. ing fertilization. J Range Manage 49:530–534. Glitzenstein, J.S., Platt, W.J., and Streng, D.R. 1995. Korstian, C.F., and Coile, T.S. 1938. Plant competi- Effects of fire regime and habitat on tree dynam- tion in forest stands. Duke University School of ics in north Florida longleaf pine savannas. Ecol Forestry Bulletin 3, Durham, NC. Monogr 65:441–476. McCune, B. 1986. Root competition in a low- Goldberg, D.E. 1990. Components of resource com- elevation grand fir forest in Montana: A trenching petition in plant communities. In Perspectives on experiment. Northwest Sci 60:52–54. Plant Competition, eds. J.B. Grace and D. Tilman, McGuire, J.P., Mitchell, R.J., Moser, E.B., Pecot, pp. 27–49. New York: Academic Press. S.D., Gjerstad, D.H., and Hedman, C.W. 2001. Goldberg, D.E., and Miller, T.E. 1990. Effects of dif- Gaps in a gappy forest: Plant resources, longleaf ferent resource additions on species diversity in pine regeneration, and understory response to an annual plant community. Ecology 71:213–225. tree removal in longleaf pine savannas. Can J For Grace, J.B. 1990. On the relationship between plant Res 31:765–778. traits and competitive ability. In Perspectives on Means, D.B. 1997. Wiregrass restoration: Probable Plant Competition, eds. J.B. Grace and D. Tilman, shading effects in a slash pine plantation. Restor pp. 51–65. New York: Academic Press. Manage Notes 15:52–55. Grace, S.L., and Platt, W.J. 1995. Effects of adult tree Monk, C.D., and Gabrielson, F.C., Jr. 1985. Effects density and fire on the demography of pregrass of shade, litter, and root competition on old-field stage juvenile longleaf pine (Pinus palustris Mill.). vegetation in South Carolina. Bull Torrey Bot Club J Ecol 83:75–86. 112:383–392. Grelen, H.E., and Enghardt, H.G. 1973. Burning and Mueller-Dombois, D., and Ellenberg, H. 1974. Aims thinning maintain forage in a longleaf pine plan- and Methods of Vegetation Ecology. New York: John tation. J For 71:419–425. Wiley & Sons. Harper, J.L. 1977. Population Biology of Plants. New Oliver, C.D., and Larson, B.C. 1996. Forest Stand Dy- York: Academic Press. namics. Update edition. New York: John Wiley & Harrington, T.B., and Edwards, M.B. 1999. Under- Sons. story vegetation, resource availability, and litter- Palik, B.J., Mitchell, R.J., Houseal, G., and Pederson, fall responses to pine thinning and woody vege- N. 1997. Effects of canopy structure on resource tation control in longleaf pine plantations. Can J availability and seedling responses in a longleaf For Res 29:1055–1064. pine ecosystem. Can J For Res 27:1458–1464. Harrington, T.B., Dagley, C.M., and Edwards, M.B. Palik, B., Mitchell, R.J., Pecot, S., Battaglia, M., 2003. Above- and belowground competition from and Mou, P. 2003. Spatial distribution of over- longleaf pine plantations limits performance of story retention influences resources and growth reintroduced herbaceous species. For Sci 49:681– of longleaf pine seedlings. Ecol Appl 13:674– 695. 686. Haywood, J.D. 2000. Mulch and hexazinone herbi- Pessin, L.J. 1938. The effect of vegetation on the cide shorten the time longleaf pine seedlings are growth of longleaf pine seedlings. Ecol Monogr in the grass stage and increase height growth. New 8:115–149. For 19:279–290. Platt, W.J., Evans, G.W., and Rathbun, S.L. 1988a. Hendricks, J.J., and Boring, L.R. 1992. Litter qual- The population dynamics of a long-lived conifer ity of native herbaceous legumes in a burned (Pinus palustris). Am Nat 131:491–525. pine forest of the Georgia Piedmont. Can J For Res Platt, W.J., Evans, G.W., and Davis, M.M. 1988b. 22:2007–2010. Effects of fire season on flowering of forbs and Hendricks, J.J., and Boring, L.R. 1999. N2-fixation shrubs in longleaf pine forests. Oecologia 76:353– by native herbaceous legumes in burned pine 363. 156 II. Ecology

Radosevich, S.R., and Holt, J.S. 1984. Weed Ecology. ings 18th Tall Timbers Fire Ecology Conference, pp. New York: John Wiley & Sons. 227–263. Richardson, D.R., and Williamson, G.B. 1988. Al- Sydes, C., and Grime, J.P. 1981a. Effects of tree lelopathic effects of shrubs of the sand pine scrub leaf litter on herbaceous vegetation in deciduous on pines and grasses of the Sandhills. For Sci woodland. I. Field investigations. J Ecol 69:237– 34:592–605. 248. Riegel, G.M., and Miller, R.F. 1991. Understory veg- Sydes, C., and Grime, J.P. 1981b. Effects of tree etation response to increasing water and nitrogen leaf litter on herbaceous vegetation in decidu- levels in a Pinus ponderosa forest in northeastern ous woodland. II. An experimental investigation. Oregon. Northwest Sci 65:10–15. J Ecol 69:249–262. Riegel, G.M., Miller, R.F., and Krueger, W.C. 1992. Toumey, J.W., and Kienholz, R. 1931 (cited in Spurr Competition for resources between understory and Barnes 1992). Trenched plots under forest vegetation and overstory Pinus ponderosa in north- canopies. Yale University School of Forestry Bul- eastern Oregon. Ecol Appl 2:71–85. letin 30. Rogers, V.A. 1990. Soil survey of Savannah River Van Lear, D.H., and Waldrop, T.A. 1991. Pre- Plant area, parts of Aiken, Barnwell, and Allen- scribed burning for regeneration. In Forest Re- dale Counties, South Carolina. USDA Soil Con- generation Manual, eds. M. L. Duryea and P.M. servation Service, Washington, DC. Dougherty, pp. 235–250. Dordrecht, The Nether- Shelton, M.G. 1995. Effects of the amount and com- lands: Kluwer Academic Publishers. position of the forest ﬂoor on emergence and Wilson, C.A., Mitchell, R.J., Hendricks, J.J., and early establishment of loblolly pine seedlings. Can Boring, L.R. 1999. Patterns and controls of ecosys- J For Res 25:480–486. tem function in longleaf pine – wiregrass savan- Sokal, R.R., and Rohlf, F.J. 1981. Biometry, 2nd ed. nas. II. Nitrogen dynamics. Can J For Res 29:752– New York: W.H. Freeman and Company. 760. Spurr, S.H., and Barnes, B.V. 1992. Forest Ecology, Wolters, G.L. 1973. Southern pine overstories inﬂu- 3rd ed. Malabar, FL: Krieger Publishing Co. ence herbage quality. J Range Manage 26:423–426. Streng, D.R., Glitzenstein, J.S., and Platt, W.J. 1993. Wolters, G.L. 1981. Timber thinning and prescribed Evaluating effects of season of burn in longleaf burning as methods to increase herbage on grazed pine forests: A critical literature review and some and protected longleaf pine ranges. J Range Man- results from an ongoing long-term study. Proceed- age 34:494–497. Chapter 6 Vertebrate Faunal Diversity of Longleaf Pine Ecosystems

D. Bruce Means

Introduction Before describing longleaf pine ecosystems, something must be said about how we know In the southeastern U.S. Coastal Plain, all what we think we know about them. A good landscapes can be conceptually divided into deal of knowledge is available about vertebrate aquatic, wetland, and upland habitats. Aquatic species in longleaf ecosystems from numerous and wetland habitats account for a substan- autecological studies. However, knowledge tial percentage of the Coastal Plain, especially about longleaf pine communities as function- near the coast and in Louisiana and Florida, ing ecosystems has been difficult to obtain be- but overall from southeastern Virginia to east cause in the twentieth century, when plant Texas, uplands constitute the largest propor- ecology became a scientific discipline, most of tion of the terrain. It has been estimated that, the longleaf pine habitat had already disap- upon the arrival of Europeans and Africans peared, and that which remains had been drain North America, upland ecosystems domi- matically impacted by man. As recently as 13 nated by a single tree species, longleaf pine years ago, it was calculated that only 3% of (Pinus palustris), accounted for about 60% of the original extent of longleaf pine ecosystems the Coastal Plain landscape (Ware et al. 1993). remained (Ware et al. 1993). A few forestry In other words, longleaf pine ecosystems were statistics tell the story. In Florida, longleaf pine the principal ecosystems in a belt of land forests declined from 30,756 km2 (7.6 million stretching about 2000 miles along the south- acres) in 1936 to only 3845 km2 (0.95 mil- eastern margin of the North American conti- lion acres) in 1989, an 88% decrease (Cerulean nent. Most of the range of longleaf pine was 1991). In southeastern Georgia, the longleaf in the Coastal Plain, a gently undulating, low- pine forest declined 36% between 1981 and elevation (0–200 m), sedimentary landform 1988 (Johnson 1988). The first inventory of with soils developed from sandy clays (clayhills what was left of old-growth longleaf reported and some flatwoods) or pure sand (sandhills a meager 9975 acres out of an estimated 85 and flatwoods), sometimes underlain by lime- million acres, or about 0.01% (Means 1996). stone (Brown et al. 1990; Martin and Boyce Only 5 years later, those valuable acres had 1993). Longleaf pine ecosystems and their ver- dwindled further by 43% (Varner and Kush tebrate faunas are the focus of this chapter. 2001). Our knowledge about plant species

r D. Bruce Means Coastal Plains Institute and Land Conservancy, Tallahassee, Florida 32303.

157 158 II. Ecology composition, structure, community function, mary productivity takes place high in the trees, and even geographical distribution of its com- vertebrates are volant, scansorial, and arbo- ponents, therefore, has been pieced together real. Herbivores that feed far above ground in from historical records and recent research the branches are limited in body size. Larger, in the few longleaf habitats that remain. ground-dwelling mammalian herbivores are Rather than possessing direct observational ev- mainly browsers. Terrestrial amphibians and idence for such information as fire frequency, reptiles must feed on detrital invertebrates and stand density, groundcover species composi- are therefore specialized to burrow in litter tion, and other properties, we have had to and feed upon a detritivore-based food web. infer them from brief and incomplete histor- In open-canopied savannas and prairies, by ical accounts, anecdotal information, and ex- contrast, a large percentage of ecosystem pri- perimental manipulations of small remaining mary productivity takes place at ground level, tracts whose naturally functioning ecological supporting a diverse herbivore fauna charac- processes have been interrupted for possibly terized by large grazing mammals and even more than several centuries. In discussions reptiles (tortoises). Insectivorous amphibians that follow about the ecology of longleaf pine and other reptiles can feed on a contiguous ecosystems, I will sometimes use the past tense supply of arthropods on the ground and in to refer to truly pristine, unaltered longleaf the low herbaceous vegetation. Vertebrates pine ecosystems, because none are left and can be so specialized for the types of habitats probably haven’t been since about the 1920s. in which they live that paleontologists regu- What remains (less than a million acres) is a larly deduce the general types of vegetation small amount of second-growth longleaf pine that existed in the vicinity of a fossil site by ecosystems mostly on publicly owned lands. whether the fossil fauna is dominated by an- Longleaf pine ecosystems are not well de- imals whose bones and teeth give evidence scribed as forest because they more read- that they were browsers, grazers, carnivores, ily fit the definition of “savannas” (Vogl walked, burrowed, or climbed. Supple-bodied, 1973), which are clumps of trees or sparsely climbing snakes such as boas, for example, are distributed trees not forming continuous indicative of forest habitats whereas elongate, canopies, and groundcover dominated by fast-moving snakes such as racers and whip- abundant heliophilic herbaceous vegetation, snakes suggest open, grassland habitats. That particularly warm-season grasses (Werner longleaf pine savannas were open-canopied 1991; McPherson 1997; Scholes and Archer grasslands or prairies as opposed to closed- 1997; Platt 1999). Historical accounts of early canopy hardwoods or closed-canopy conifer- explorers, travelers, and botanists reveal that ous forests, therefore, had a great deal of the rich grassland or prairie ground covers influence on both the kinds and diversity of of pine savannas stretched across the Coastal vertebrates that lived in these ecosystems. Plain and Piedmont from the Atlantic to the The ground cover of Coastal Plain longleaf Gulf of Mexico and continued beyond the lim- pine savannas contained a diverse mixture of its of longleaf pine in east Texas into the tall- grasses, forbs, and shrubs (Walker and Peet grass prairies of the Great Plains (reviews in 1983; Christensen 1988; Bridges and Orzell Vogl 1973; Platt 1999; Frost this volume). To 1989; Harcombe et al. 1993; Peet and Allard understand the vertebrate faunas of longleaf 1993; Peet this volume). Platt (1999) listed pine savannas, it is crucial to appreciate that the most dominant grasses, most of which longleaf pine ecosystems are forests to only belong to warm-season genera: Andropogon, a handful of species, but that for most verte- Aristida, Ctenium, Schizachyrium, Sorghastrum, brates, they are, or were, grasslands. Sporobolus. Hundreds of forb species occurred This is important because terrestrial verte- in longleaf pine savannas, with Asteraceae, brates (amphibians, reptiles, mammals, birds) Lamiaceae, Fabaceae, Liliaceae, Scrophulari- respond to vegetation in specific ways. In aceae, Ericaceae, Orchidaceae, and Xyridaceae closed-canopy forests where most of the pri- occurring as dominants (Christensen 1988; 6. Vertebrate Faunal Diversity 159

Platt et al. 1988; Hardin and White 1989; and the mammalian fauna is exceedingly im- Walker 1993). Walker (1993) proposed that poverished (Layne 1974). Below, I review the endemic species were distributed among four scant literature on vertebrate faunas of lon- different regions within the overall range of gleaf pine savannas, then discuss the ecological longleaf pine: south Atlantic Coastal Plain, roles of each vertebrate species that is a special- peninsular Florida, and the east and west Gulf ist adapted for living in longleaf pine savannas. Coastal Plains. About 200 of these species At the end of this chapter, I discuss why the are rare (Hardin and White 1989; Walker mammalian fauna is depauperate and spec- 1993). Shrubs and subshrubs were common ulate on what effect the impoverished mam- in the ground cover of pine savannas, includ- malian fauna may have had on longleaf pine ing sprouts of trees, especially oaks (Quercus savannas. laevis, Q. margaretta, Q. incana, Q. marilandica), Only a few overviews of the vertebrate fau- as well as runner oaks (Q. minima, Q. pumila), nas of longleaf pine savannas have been pub- sumac (Rhus), ericaceous shrubs (Vaccinium, lished. The amphibians and reptiles of Georgia Gaylussacia), palms (Serenoa, Sabal), wax myr- and Florida longleaf pine sandhills commu- tle (Myrica), and hollies (Ilex) (Peet and Allard nities were discussed by Landers and Speake 1993; Olson and Platt 1995). (1980) and Campbell and Christman (1982), Peet and Allard (1993) found that the num- respectively. Means and Campbell (1982) ad- bers of plant species present in individual sam- dressed the effects of fire on amphibians and ples in longleaf savannas across the Coastal reptiles worldwide, but most of their examples Plain were among the highest reported for came from longleaf pine savannas. The most the temperate Western Hemisphere. Values thorough reviews of amphibians and reptiles for vascular plants ranged from 140/1000 were done by Stout and Marion (1993) and m2 and 90+/10 m2 to counts of more than Guyer and Bailey (1993). A shorter review of 40 species/m2. Several hundred plant species amphibians and reptiles was done by Dodd may occur within an area smaller than a (1995b). hectare (Platt et al. 1988; Peet and Allard Bird communities of Southeastern pinelands 1993; Stout and Marion 1993). This species were reviewed by Jackson (1988), but he richness, the highest reported from North cautioned that no data were available on American savannas (Peet and Allard 1993), species composition or population abundances rivals that of tropical forests, but on a two- of birds in old-growth longleaf pine savannas. dimensional, ground-level, square-meter spa- Stout and Marion (1993) discussed bird and tial scale. Moreover, diversity is also high on mammal community assemblages of pine flat- a regional scale. Peet and Allard (1993) recog- woods and longleaf pine-turkey oak sandhills. nized 23 different longleaf pine savanna com- Engstrom (1993) prepared the most compre- munities across the range of longleaf pine east hensive review of mammals and birds of lon- of the Mississippi River. gleaf pine savannas, recognizing those species With such high plant species richness in such that characteristically live but are not special- a large natural area as the Coastal Plain, one ists in longleaf savannas and species that are would expect vertebrate species richness to be largely sympatric with longleaf savannas and commensurately high. And indeed, it is for presumably specialists in it. Echternacht and amphibians and reptiles. The Coastal Plain of Harris (1993) reviewed the vertebrate fauna of the southeastern United States has the highest the entire southeastern United States but did species richness of turtles, frogs, and snakes in not mention habitat associations. the United States and Canada (Kiester 1971), The four vertebrate classes possess entirely as well as a large salamander fauna (Means different morphologies, physiologies, vagili- 2000). Only lizards, which flourish in arid en- ties, life history characteristics, and ecologies, vironments, are not superabundant. On the so a review of all the vertebrate species that live other hand, bird species richness is not ex- in longleaf pine savannas is best done by dis- ceptionally great (Stout and Marion 1993), cussing each class separately. We must define, 160 II. Ecology however, criteria for how species were cho- phosed individuals can be intercepted by drift sen for discussion. Some species in each class fences as populations move in and out of na- may be widely ubiquitous across many differ- tal ponds (Dodd 1992; Semlitsch et al. 1996). ent habitat types throughout the Coastal Plain, However, in every case among the 35 species of including living in longleaf pine savannas, or amphibians that live in longleaf pine savannas have geographic distributions that overlap that as adults, little is known about their terrestrial of longleaf pine savannas, although the species lives, because field studies of them are diffi- never lives in them. Other species may occupy cult to conduct. Both juveniles and adults are longleaf pine savannas in those parts of their usually fossorial or semifossorial, do not vocal- much wider geographic distributions in North ize away from water, and are small and cryptic America. And some species are strict endemics morphologically and behaviorally. that live exclusively in longleaf pine savan- Throughout the Coastal Plain, there are nas. Guyer and Bailey (1993) and Engstrom numerous kinds of ponds in the longleaf (1993) defined two groups of longleaf pine sa- pine landscape, including 500,000 Carolina vanna vertebrates: (1) species that are charac- Bays (Prouty 1952; Savage 1982; Sharitz teristic, that is, they occur in longleaf savannas and Gibbons 1982), 7000 Citronelle ponds but whose distributional limits (both ecological (Folkerts 1997), and probably a similar num- and geographical) are not confined to the dis- ber of limestone dissolutional basins in flat- tribution of longleaf pine (i.e., residents) and woods and sandhills. Ponds may be forested (2) species whose limits are confined within with cypress (Taxodium), tupelos (Nyssa), and/or closely associated with those of longleaf myrtle-leaved holly (Ilex myrtifolia), pop-ash pine (i.e., specialists). Below I adopt this classi- (Fraxinus carolina), titi species (Cyrillaceae), fication scheme with slight modifications and or other shrubs, or ponds may be open and then briefly discuss details of longleaf pine sa- dominated by emergent grasses (Panicum) and vanna specialist and resident species for each sedges (Cladium). They may have very short vertebrate class. Throughout the text, com- hydroperiods of only a few months in win- mon names are used; corresponding scientific ter or be nearly permanent. What seems most names are listed in Tables 1–5. important to the 17 species of longleaf pine amphibian specialists is the absence of fish, so that the preferred ponds in which they lay Amphibians their eggs have short hydroperiods (Moler and Franz 1988). To illustrate the importance of the Most amphibians have a biphasic life cycle in fishless, temporary nature of ponds, some 26 which their natal stages (eggs, larvae) live in species of anurans are known to characteristi- water, and they then take up permanent res- cally breed in those ponds found in longleaf idence on land as adults after metamorphosis pine savannas of the Coastal Plain (Table 1) (Duellman and Trueb 1986). Of the 9 salaman- compared to only 8 species that breed in num- der and 26 frog species that live in longleaf bers (bullfrog, bronze frog, pig frog, southern pine savannas, 6 salamanders and 11 frogs leopard frog, carpenter frog, green treefrog, have adults that are specialists in these ecosys- southern cricket frog, southern toad) in the tems (Table 1). All require lentic situations— extensive permanent lakes and marshes of the small ponds, often with short hydroperiods— Coastal Plain. in their natal stages (Moler and Franz 1988). Two breeding seasons are present in the Many others, stream-dwelling species, also oc- Coastal Plain range of longleaf pine savan- cur in the Coastal Plain but live in floodplains nas, because this region is characterized by and swamps as adults. Much is known about a distinctive precipitation regime of sum- the breeding biology and larval life in ponds mer thundershowers (June–September) and of longleaf pine savanna amphibians, because rains brought on by the passage of win- calling males of frogs are easy to locate, lar- ter cold fronts (December–March). The rainy vae can be seined or dipnetted, and metamor- seasons are typically separated by autumn 6. Vertebrate Faunal Diversity 161

, TABLE 1. Characteristic resident amphibians of longleaf pine savannas.a b Salamanders 1. Ambystoma cingulatum Flatwoods salamander s w 2. Ambystoma mabeei Mabee’s salamander s w 3. Ambystoma talpoideum Mole salamander b w 4. Ambystoma tigrinum Tiger salamander s w 5. Notophthalmus perstriatus Striped newt s w 6. Notophthalmus viridescens Eastern newt b w 7. Eurycea quadridigitata Dwarf salamander s w 8. Eurycea n. sp. Bog salamander s w 9. Plethodon glutinosus “complex” Slimy salamander c m Frogs 1. Bufo americanus American toad a w 2. Bufo quercicus Oak toad s m 3. Bufo terrestris Southern toad b m 4. Bufo valliceps Gulf coast toad a m 5. Bufo woodhousii Woodhouse’s toad a m 6. Acris crepitans Northern cricket frog c m 7. Acris gryllus Southern cricket frog d m 8. Hyla andersonii Pine barrens treefrog c m 9. Hyla chrysoscelis Cope’s gray treefrog c m 10. Hyla cinerea Green treefrog d m 11. Hyla femoralis Pinewoods treefrog s m 12. Hyla gratiosa Barking treefrog s m 13. Hyla squirella Squirrel treefrog s m 14. Pseudacris brachyphona Mountain chorus frog a w 15. Pseudacris brimleyi Brimley’s chorus frog s w 16. Pseudacris crucifer Spring peeper c w 17. Pseudacris nigrita Southern chorus frog s w 18. Pseudacris ocularis Little grass frog s m 19. Pseudacris ornata Ornate chorus frog s w 20. Pseudacris triseriata Western chorus frog c m 21. Gastrophryne carolinensis Eastern narrowmouth toad b m 22. Gastrophryne olivacea Great plains narrowmouth toad a m 23. Rana areolata Crawﬁsh frog s w 24. Rana capito Gopher frog s w 25. Rana sphenocephala Southern leopard frog d w 26. Scaphiopus holbrookii Eastern spadefoot s w a Adapted, with modiﬁcations, from Moler and Franz (1988), Stout and Marion (1993), Guyer and Bailey (1993), Dodd (1995b). b a = small overlap with longleaf pine savanna range; b = ubiquitous among Coastal Plain habitats; c = visitor from other habitats; d = aquatic; s = endemic in longleaf pine savanna, a specialist; w = winter breeder; m = summer breeder.

(October–November) and spring (April–May) (Taxodium ascendans) and blackgum or black droughts (Wolfe et al. 1988; Martin and Boyce tupelo (Nyssa sylvatica). Adult flatwoods sala- 1993). manders migrate to breeding ponds in October and November and then move out of ponds in Longleaf Pine Amphibian late November and December (Palis and Means 2005). The flatwoods salamander is a true en- Specialists demic, highly specialized for living in longleaf All six salamander specialists in longleaf pine savanna, but only in flatwoods, not sand- pine savannas are winter breeders. Flatwoods hills (Goin 1950; Means et al. 1996). Whether (Fig. 1a) and Mabee’s salamanders breed in it is something about the quality of the terres- small, fishless ponds in longleaf pine savan- trial habitat of the adults or the natal ponds nas that usually have a canopy of pond cypress that confines the flatwoods salamander to 162 II. Ecology

FIGURE 1. (a) Flatwoods salamander, Ambystoma cingulatum, a specialist in longleaf pine ﬂatwoods. (b) Striped newt, Notophthalmus perstriatus, a specialist in longleaf pine sandhills, terrestrial stage called “eft.”

flatwoods is unknown, but a similar, although with cobalt-60 radioactive tags (Ashton and reverse, proclivity exists in the striped newt Ashton 2005), indicate that the species lives (Fig. 1b), another strict longleaf pine savanna its adult life in longleaf pine savanna sev- endemic species that is mainly found in sand- eral hundred meters away from natal ponds. hills rather than flatwoods (Franz and Smith Over a 7- to 9-day period, Ashton and Ash- 1999; Johnson 2002; Dodd et al. 2005). ton (2005) recorded moves of 1566, 1702, Observations of migrating flatwoods sala- and 1708 m in three adults emigrating from mander adults found in the field (Means et the breeding pond, although not necessar- al. 1996), and a study of four individuals ily in a straight-line direction. Likewise, the leaving breeding ponds that were monitored smaller striped newt emigrates more than 500 6. Vertebrate Faunal Diversity 163 m from its natal ponds into sandhill longleaf documented living in old-growth longleaf pine pine-turkey oak vegetation (Johnson 2003). clayhills on the Wade Longleaf Forest in Grady Striped newts have lived 12 years in captiv- County, GA (Means and Campbell 1982). Lit- ity (Snider and Bowler 1992; Linda LaClaire, tle remains of clayhills longleaf pine savanna personal communication). No studies of the in first- or second-growth character (Means upland habitat of Mabee’s salamander have 1996), but the eastern tiger salamander some- been published, although the kinds of ponds times is found in clayhills ruderal habitats such it breeds in (Mosimann and Rabb 1948; Hardy as farmland, pastures, and oldfield successional 1969; Mitchell and Hedges 1980), its geo- stands of mixed hardwoods and pines where graphic range (Petranka 1998), and morpho- old-growth longleaf pine savanna once grew logical similarity to the flatwoods salaman- (Means and Campbell 1982; Travis 1992). It der (Hardy and Olmon 1974) suggest that it, is found in sandhills and xeric hammock habi- too, is a longleaf pine specialist as an adult. tats in peninsular Florida (Kevin Enge personal Early accounts of the flatwoods salamander de- communication), but in the Coastal Plain por- scribed the adult habitat as slash pine flatwoods tion of its large geographic distribution, it is (Goin 1950; Martof 1968; Conant and Collins not known from southern temperate (beech- 1998), but Means et al. (1996) and Palis (1996) magnolia) hardwood forests (as defined in demonstrated that the primary adult habitat is Platt and Schwartz 1990) or sand pine (Pinus longleaf pine savannas. clausa) and coastal scrubs, the only other up- During autumn breeding migrations, the land habitats in the Coastal Plain presettlement flatwoods salamander has been observed mov- landscape. Like many ambystomatid salaman- ing to breeding ponds out of longleaf pine sa- ders, it is rarely found outside the breeding sea- vanna (Means et al. 1996), but since few adults son, a time when individuals are seen crossing have been taken outside the breeding season, roads or scooped from swimming pools during microhabitat utilization is virtually unknown winter migrations (November–January). This except that the species has been reported us- large salamander probably occupies the bur- ing crayfish burrows when in the breeding rows or subterranean cavities of moles, ro- habitat (Neill 1951b; Ashton 1992; Ashton and dents, and other animals, as well as excavates Ashton 2005). Observations of tagged adults its own tunnels (Semlitsch 1983a,b). Studies of during emigration from the breeding pond in- its upland habitat use and behavior have not dicated tagged salamanders were never found been conducted, not even a determination of anywhere inside a large slash pine plantation whether it comes to the ground surface at night except in pine duff up to 2 cm deep even when to forage. rotting logs were present. Three post-breeding Larvae of the dwarf salamander (Eurycea flatwoods salamanders completed emigration quadridigitata) have been found in the field in 7–9 days and occupied underground bur- from late January to March following breed- rows of undetermined origin in a mesic area ing migrations to ponds and oviposition in at the edge of the pine plantation (Ashton and ponds in October and November (Brimley Ashton 2005). The mesic area was vegetated 1923; Harrison 1973; Semlitsch and McMillan with Florida maple (Acer barbatum), sweet gum 1980). Recently, populations in the Piedmont (Liquidambar styraciflua), black gum, gallberry and upper Coastal Plain that previously were (Ilex glabra), wax myrtle (Myrica cerifera), and referred to this species have been found to be grape (Vitis rotundifolia). No reports of micro- a distinct species, E. chamberlaini, that breeds habitat use are available during the terrestrial in streams and whose adults probably live lives of the striped newt, common newt, or in hardwood forests (Harrison and Guttman Mabee’s salamander besides mention of an oc- 2003). The remaining populations that occur casional individual found under logs (Petranka in flatwoods and sandhills of the lower Coastal 1998). Plain from east Texas to North Carolina consist Here, I classify the eastern tiger salamander of three distinct species (Richard Highton and as a longleaf pine savanna specialist. It has been Carla Hass personal communication). At least 164 II. Ecology two of these species breed in temporary ponds, species (Enge 1997, 2002; Means and Jensen including E. quadridigitata, and another unde- unpublished; David Printiss personal commu- scribed species breeds in seepage bogs and wet nication). Very little has been published about flats (Enge 2002; D. B. Means and J. Jensen the ecology of these species. unpublished). Adults of these three species live No less than 11 species of frogs are longleaf in, or spend time in, longleaf savannas, espe- pine specialists (Table 1). Five of these are cially mesic or wet flatwoods. Drift fence stud- winter breeders—southern chorus frog, or- ies of longleaf pine savanna amphibian faunas nate chorus frog, crawfish frog, gopher frog have reported captures of one or two of these (Fig. 2), eastern spadefoot—and the remaining

FIGURE 2. (a) Gopher frog, Rana capito, a specialist in longleaf pine sandhills. (b) Pine barrens treefrog, Hyla andersonii, dependent upon hillside seepage bogs maintained by fires sweeping downhill from longleaf pine savannas 6. Vertebrate Faunal Diversity 165 six species breed in summer. Of the winter recent clearcut. After 2 days, frogs became breeders, little is known of the habitats of stationary in underground retreats associated the adult southern chorus frog, but drift fence with stump holes, root mounds of fallen trees, studies of the ornate chorus frog clearly indi- or mammal burrows (Richter et al. 2001). cate that this species migrates long distances The use of gopher tortoise burrows by the into all types of longleaf pine savannas, in- gopher frog has been known for a long time, cluding those in flatwoods, clayhills, and sand- ergo the common name of the species (Test hills (Means and Campbell 1982; Brown and 1893). Carr (1940) reported that the gopher Means 1984). In fact, the ornate chorus frog frog utilizes burrows of animals other than actively burrows forward using its robust arms the gopher tortoise: cotton mouse (Peromyscus and sometimes backward using its hind legs polionotus), stumpholes, tip-up mounds, and (Brown and Means 1984). Neill (1958) found other subterranean cavities. Dispersal dis- a few ornate chorus frogs in sand while prob- tances of up to 200 m from breeding ponds ing at the bases of wiregrass clumps. In a study have been recorded for the gopher frog (Means of the use by both species of chorus frogs of and Means 2005). Gopher and crawfish frogs Carolina Bay ponds in South Carolina over a are found almost exclusively in sandhills and 4-year period, Caldwell (1987) concluded that rarely in flatwoods (Wright 1932; Enge 1997). both were short-lived as adults, because ju- This would seem to be due to the dependency veniles mature early and disperse to nearby of adults on the friable soils of sandhills and ponds when ponds are abundant in most years. the availability in sandhills of suitable subter- She did not gather information about the habi- ranean refugia such as burrows of the gopher tat use of metamorphosed juveniles and adults, tortoise, which is itself restricted to well- however. drained soils vegetated with longleaf pine- Three other winter-breeding frogs (craw- oak, xeric hammocks, sand pine-oak ridges, fish frog, gopher frog, eastern spadefoot) have and ruderal successional vegetations that have adults that are long-lived (for the latter two replaced these vegetations (Auffenberg and species, 9 and 12 years in captivity, respec- Franz 1982). The sandy ridges and slopes of tively, Snider and Bowler 1992) and breed only longleaf pine clayhills vegetation support pop- after exceptional rainfall bouts (Hansen 1958; ulations of the gopher tortoise but apparently Palis 1998; Bailey and Means 2004). Their lar- not the gopher frog (Means and Campbell vae develop rapidly (Semlitsch et al. 1995; 1982; Means and Means 1998). Palis 1998; Richter 1998; Means unpublished), Summer-breeding frogs that are longleaf and juvenile recruitment can be very high pine savanna specialists include little grass (Greenberg 2001; Means unpublished). Breed- frog, oak toad, and pinewoods, barking, and ing events are less regular in these species and squirrel treefrogs. All of these species can be can occur over a longer span of time, mostly in heard calling from the same temporary ponds winter from January through March but occa- in March–August (Wright 1932; Wright and sionally at any time of the year (Franz 1991; Wright 1949). Adults of each of these species Semlitsch et al. 1995; Franz and Smith 1999; have been reported in longleaf savannas from Means and Means 2005). Dispersal of juveniles miscellaneous observations (Carr 1940; Wright into longleaf pine savannas and adult use of and Wright 1949) and drift fence studies (Enge longleaf pine savanna have been documented 1997), but few special investigations of feed- for all three species. Richter et al. (2001) ing habits, nocturnal behavior, or microhabi- showed that adult crawfish frogs moved rel- tat utilization have been conducted. The large atively short distances (less than 300 m) out of barking treefrog has been found during day- a pond surrounded by longleaf pine savanna time in terrestrial burrows in sand at the base in southern Mississippi. However, the land ap- of wiregrass clusters, some nestled in con- proximately 200 m north of the pond was a cavities and others in burrows 5–8 cm deep privately managed pine plantation, so further (Neill 1952). How much of its activity is re- movements may have been constrained by a stricted to the ground versus trees and shrubs 166 II. Ecology is unknown, but at least some of its activity is 1997), as well as in longleaf pine savannas, and arboreal (Neill 1958). The pinewoods treefrog they have been taken in drift fences in longleaf is found throughout longleaf pine savannas sandhills 200 m from breeding ponds (Means and often occurs in crevices in the bark of and Means 2005). The eastern newt breeds in longleaf pine trees (Means personal observa- an even greater diversity of aquatic habitats, tion). Squirrel treefrogs are common in lon- including marshy streams, canals, lakes with gleaf pine ground cover, especially when it fish, and temporary ponds (Petranka 1998). Its contains a woody component of saw palmetto, ubiquity probably is due to its noxious skin se- turkey oak, or gallberry. The tiny oak toad for- cretions, which deter predation. Metamorphs ages on the ground and primarily eats ants and adults of populations that utilize ponds (Punzo 1995). Oak toads were abundant in in flatwoods and sandhills spend the terres- old-growth longleaf pine in clayhills but not trial phase of their lives in longleaf pine savan- in nearby shortleaf/loblolly pine oldfield suc- nas, presumably very similarly to the striped cessional forest on the same soil type (Means newt. Both species can be found using the and Campbell 1982), suggesting that the oak same breeding ponds, as well as migrating into toad is highly specialized for living in wiregrass the same longleaf pine sandhills more than 200 ground cover or is unable to recolonize old lon- m away from the pond (Means and Means gleaf pine sites that have been abandoned af- 2005). ter agriculture. Very little is known about the The 15 frogs listed as residents but not spe- adult habitat of the tiniest frog in the United cialists can be grouped into four categories: States, the little grass frog. From March to Au- (1) those whose large ranges just enter gust, it can be heard calling throughout pine the distribution of longleaf pine (American flatwoods. toad, gulf coast toad, mountain chorus frog, Brimley’s chorus frog); (2) ubiquitous over a range of habitats including longleaf pine sa- Other Longleaf Pine vanna (southern toad, eastern narrowmouth toad); (3) visitors that are more common in Amphibians other habitats (woodhouse’s toad, northern Three other salamanders and 15 frogs are res- cricket frog, pine barrens treefrog, Cope’s gray idents that commonly occur but are not ex- treefrog, spring peeper, western chorus frog); clusively found in longleaf pine savannas. The and (4) more-or-less aquatic species (cricket slimy salamander is occasionally found in rot- frog, green treefrog, southern leopard frog). ting logs in flatwoods and clayhills, especially All of these species can be found in longleaf where the longleaf pine savanna transitions pine savannas as adults while crossing through into the southern temperate hardwood for- these habitats to reach breeding ponds. Some est on slopes and moist bottomlands (listed by (southern toad, eastern narrowmouth toad) Guyer and Bailey 1993). It is uncommon in are as abundant in longleaf pine savannas as longleaf pine ecosystems and probably should the specialist frogs and might be classified as not be considered a resident species. The mole specialists were they not also found in other salamander breeds in a wide diversity of pond habitats. Others are clearly occasional visitors habitats, including forested as well as open- (all frogs in category 3, southern cricket frog, canopied, grassy ponds in flatwoods, clayhills, green treefrog). The southern leopard frog is an and sandhills (Shoop 1960; Semlitsch et al. interesting case because it probably breeds in 1981). It breeds in farm ponds and ponds with more ponds in the Coastal Plain landscape than or without fish and commonly breeds in the any other frog, produces a huge annual crop same ponds as most of the longleaf pine sa- of metamorphs, and is highly obvious during vanna specialists, both in sandhills and in flat- rainy nights dispersing through the uplands in woods (Means and Means 2005). Adults have long jumps. Strangely, the proportion of time been captured in drift fence studies in bottom- the adult leopard frog spends away from the land and floodplain hardwood forests (Enge edges of ponds is unknown, but the evidence 6. Vertebrate Faunal Diversity 167 from drift fence studies indicates a substantial Amphibians make three major contributions amount (Enge 1997). to longleaf pine savanna ecology. First, nutri- Amphibians probably contribute more to ents acquired in natal ponds are transported the ecology of longleaf pine savannas than in body tissues into the rather nutrient-poor any other group of vertebrates. Huge num- uplands (Platt 1999) where, upon death or bers of metamorphosed salamanders and frogs ingestion by predators, the nutrients even- have been recorded emerging from natal ponds tually wind up in the savanna soils. Sec- and migrating into adjacent uplands. Over a ond, all juvenile and adult amphibians are 16-year period, 216,251 metamorphs of five carnivores that prey upon invertebrates and species of salamanders and eight species of even small vertebrates (Wright 1932; Petranka frogs emerged from Rainbow Bay, a 1-ha 1998), so that by their sheer numbers, they pond with maximum depth of about 1 m in are important predators in the longleaf pine Barnwell County, SC (Semlitsch et al. 1996). savanna food webs. Third, the high numbers From a very small, 0.02-ha temporary pond in of amphibians that annually take up life in Leon County, FL, metamorphs of 5895 mole longleaf pine savannas provide a large prey salamanders, 49,664 eastern spadefoots, 3154 base for all the carnivores (snakes, shrikes, southern leopard frogs, and 2235 gopher frogs hawks, owls, skunks, weasels, bobcats, opos- emerged as a result of single breeding events sums, raccoons, shrews, moles, etc.) that eat during the study (Means 2001; Means unpub- them. lished). From a large 1.2-ha pond in Santa Rosa County, FL, about 150 female gopher frogs deposited approximately 322,660 eggs during Reptiles one winter’s breeding season (Palis 1998). The percentage that survived to metamorphosis With 13 specialist species—nine snakes, two was not studied, but if even 10% survived, the lizards, one worm lizard, and one turtle— number of recruits into the adjacent longleaf reptiles are the second most diverse group of pine sandhills would have been substantial. vertebrate specialists in longleaf pine savannas Over a 2-year period, Johnson (2002) captured after amphibians; altogether, 56 reptiles are 5731 recently transformed larvae (efts and pe- residents or spend a significant portion of their domorphs) of the striped newt emigrating from lives in longleaf pine savannas (Table 2). Rep- a small pond in Putnam County, FL. Pearson tiles are very different from amphibians in sig- (1955) calculated densities of between 1000 nificant ways. By laying a shelled egg or giving and 1250 eastern spadefoots per hectare that live birth, reptiles are freed from dependency occurred on his longleaf pine sandhill study upon an aquatic habitat during development. area in north-central Florida. And during a 5- They hatch or are born as fully developed year drought, Dodd (1992) captured a total of miniatures of the adults, do not require a 5740 adult eastern narrowmouth toads mov- biphasic life cycle, and take up residence in ing in and out of a 0.16-ha temporary pond the same habitat that will be occupied when in Putnam County, FL, with only one success- adult. Body size in reptiles is generally larger ful reproduction yielding metamorphs. Inter- because many reptiles (especially snakes) feed estingly, there was little variation in numbers upon amphibians. Insectivorous lizards such as trapped in each of the 5 years in spite of the low the scincids, iguanids, and teiids are about the rainfall, which would indicate that a substan- same size as salamanders and frogs, with the tial population of adults can survive in longleaf exception of the elongated glass lizards. Snakes pine sandhills even during long-term drought that feed on earthworms and other inverte- (Dodd 1995a). These data on only a few species brates are also small in size. Some snakes that indicate that amphibian biomass in longleaf feed on mammals (pine snake, indigo snake, pine savannas is on the order of magnitude of eastern diamondback rattlesnake) are as large, that reported for salamanders in a New Hamp- by weight, as some of the smaller mammalian shire woodland (Burton and Likens 1975). carnivores. 168 II. Ecology

, TABLE 2. Characteristic reptiles resident in longleaf pine savannas.a b Snakes 1. Carphophis amoenus Eastern worm snake c 2. Cemophora coccinea Scarlet snake s 3. Coluber constrictor Black racer p 4. Drymarchon corais Eastern indigo snake s 5. Elaphe guttata Corn snake b 6. Elaphe obsoleta Rat snake b 7. Heterodon platirhinos Eastern hognose snake p 8. Heterodon simus Southern hognose snake s 9. Lampropeltis calligaster Mole kingsnake p 10. Lampropeltis getula Common kingsnake b 11. Lampropeltis triangulum Scarlet kingsnake p 12. Masticophis flagellum Coachwhip p 13. Pituophis melanoleucus Pine snake s 14. Rhadinaea flavilata Pine woods snake s 15. Stilosoma extenuatum Short-tailed snake s 16. Storeria dekayi Brown snake b 17. Storeria occipitomaculata Red-bellied snake b 18. Tantilla coronata Crowned snake b 19. Tantilla relicta Florida crowned snake s 20. Thamnophis sauritus Eastern ribbon snake d 21. Thamnophis sirtalis Common garter snake d 22. Virginia striatula Rough earth snake b 23. Virginia valeriae Smooth earth snake b 24. Micrurus fulvius Eastern coral snake s 25. Agkistrodon contortrix Copperhead b 26. Agkistrodon piscivorus Cottonmouth d 27. Crotalus adamanteus Eastern diamondback rattlesnake s 28. Crotalus horridus Timber rattlesnake b 29. Sistrurus miliarius Pygmy rattlesnake p Lizards 1. Ophisaurus attenuatus Slender glass lizard b 2. Ophisaurus compressus Island glass lizard b 3. Ophisaurus mimicus Mimic glass lizard s 4. Ophisaurus ventralis Eastern glass lizard b 5. Sceloporus undulatus Eastern fence lizard b 6. Sceloporus woodi Florida scrub lizard c 7. Anolis carolinensis Green anole b 8. Eumeces anthracinus Coal skink d 9. Eumeces egregius Mole skink s 10. Eumeces fasciatus Five-lined skink d 11. Eumeces inexpectatus Southeastern five-lined skink b 12. Eumeces laticeps Broadhead skink c 13. Scincella lateralis Ground skink b 14. Cnemidophorus sexlineatus Six-lined racerunner b Amphisbaenian 1. Rhineura floridana s Turtles 1. Deirochelys reticularia Chicken turtle d 2. Pseudemys floridana Florida cooter d 3. Pseudemys nelsoni Florida redbelly turtle d 4. Trachemys scripta Yellowbelly slider d 5. Terrapene carolina Eastern box turtle b 6. Terrapene ornata Ornate box turtle a 7. Kinosternon baurii Striped mud turtle d 8. Kinosternon subrubrum Eastern mud turtle p 9. Sternotherus odoratus Stinkpot d 10. Gopherus polyphemus Gopher tortoise s a Modified from Guyer and Bailey (1993). b a = just enters longleaf pine savanna range; b = ubiquitous among habitats; c = visitor from other habitats; d = aquatic or semiaquatic; s = endemic in longleaf pine savanna; p = prairie or grassland habitat preferences. 6. Vertebrate Faunal Diversity 169

Some turtles are completely terrestrial (box Florida crowned snakes are Florida endemic turtles and tortoises) but most in the Coastal species, both of which are highly fossorial Plain are aquatic as adults. Nine aquatic species in deep, well-drained sandy soils of longleaf live in the same ponds as those discussed above pine-turkey oak sandhills, scrub oak, or xeric for amphibians and also spend critical times in hammocks (Campbell and Christman 1982; their lives in the uplands (Table 2). The eggs Cambell and Moler 1992) and it is thought that and hatchlings spend variable periods in the the Florida crowned snake is the main prey uplands (up to 9 months) during development of the fossorial short-tailed snake (Mushinsky and during migration to ponds (Bennett et al. 1984). The pine woods snake ranges from the 1978). Adult females, eggs, and hatchlings are Florida parishes of Louisiana to North Carolina vulnerable to predation and desiccation. And but requires open-canopied grasslands main- juveniles and adults of all species must either tained by periodic fire and cannot tolerate disperse from drying ponds in search of other the invasion of these habitats by hardwoods water bodies during droughts or take refuge in (Ernst and Ernst 2003). Very little is known the uplands until water levels are restored in about the natural foods and ecology of any ponds (Burke and Gibbons 1995). Because of of these species, including especially what size these dependencies of many aquatic turtles on and qualities an area must have to sustain min- longleaf pine savannas, they must be included imal effective populations. in any consideration of longleaf pine savanna The southern hognose snake (Fig. 3b) is faunas. most commonly found in xeric sandy habi- Snakes are the largest group of reptiles resi- tats such as longleaf pine sandhills (Martof et dent in longleaf pine savannas (Table 2). I have al. 1980; Palmer and Braswell 1995; Jensen apportioned snakes into four groups based on 1996; Tennant 1997), sand pine–rosemary different aspects of how they use longleaf pine scrub, xeric hammock (Ashton and Ashton savannas. Each snake group is discussed sepa- 1981), and associated ruderal habitats (Enge rately, and then the lizards and turtles are dis- and Wood 2003). It is one of the most xeric- cussed. adapted snakes in the eastern United States, in spite of unsubstantiated comments that it is sometimes found in dry river floodplains and Longleaf Pine Snake Specialists hardwood hammocks (Tuberville et al. 2000). Among the nine snake specialists are species It uses its upturned snout to dig up its prin- that represent a wide diversity of feeding cipal food, toads (genus Bufo) and the eastern strategies ranging from the egg-eating scar- spadefoot (Carr 1940; Goin 1947; Wright and let snake and small burrowers such as the Wright 1957) and a few other longleaf pine sa- Florida crowned snake and pine woods snake vanna vertebrates (Deckert 1918; Neill 1958; to two large predators on warm-blooded prey, Mount 1975; Beane et al. 1998). Data gath- the pine snake and eastern diamondback ered by Tuberville et al. (2000) suggest that rattlesnake, and culminating in the apex the southern hognose snake has disappeared predator that feeds on both warm- and cold- or declined in a substantial portion of its his- blooded prey, the indigo snake. Five of these torical range. A highly fossorial snake (Palmer snakes’ geographical distributions are entirely and Braswell 1995), little information is avail- nested inside the range of longleaf pine savan- able on its life history and ecology. nas and the other four have extralimital dis- The principal native habitat of the eastern di- tributions in the southwestern United States amondback rattlesnake in presettlement times (Ernst and Ernst 2003). was longleaf pine savannas, including sand- The five snakes whose ranges are encom- hills, clayhills, and flatwoods varieties, but passed entirely inside the range of longleaf since so much of what is left of the original pine are the eastern diamondback rattlesnake, longleaf pine habitat is now ruderal vegetation pine woods, short-tailed, Florida crowned, and such as shortleaf/loblolly pine oldfield succes- southern hognose snakes. The short-tailed and sional forest or pine plantations, the species 170 II. Ecology

FIGURE 3. (a) Mole skink, Eumeces egregius, a specialist in longleaf pine sandhills. (b) Southern hognose snake, Heterodon simus, a specialist in longleaf pine savanna. survives mainly in any open-canopied pine on barrier islands along both Atlantic and Gulf woodland (Martin and Means 2000). In sand- coasts, high densities of up to one to six per hills vegetation, Timmerman (1995) found hectare can be achieved (Allen 1968; Means that two females had home ranges of 46.5 ha 1986). The eastern diamondback has declined and four males averaged 84.3 ha, but he had no less dramatically than some other large lon- density data. In 80-year oldﬁeld successional gleaf pine savanna snakes probably because shortleaf/loblolly pine forest that had been it can tolerate high densities, and dense pop- burned annually in late winter, Means (1986) ulations can be maintained in small areas. found a density of one adult per 20 acres, but However, data from rattlesnake roundups in 6. Vertebrate Faunal Diversity 171

Georgia and Alabama show that the upper 2003). It is mostly active at night (Palmer and end of the size class distribution of adults has Tregumbo 1970; Nelson and Gibbons 1972) declined significantly over the past 20 years and is capable of actively burrowing into the (Means, in preparation). substrate (Wilson 1951). Reynolds (1980) es- Four other longleaf pine savanna specialist timated the mean home ranges of males and snakes (scarlet, eastern indigo, pine, and east- females to be about 1627 and 1395 m2, respec- ern coral snakes) all have large portions of tively, in North Carolina sandhills. The scar- their distributions in the range of longleaf pine let snake is one of the most abundant snakes (Ernst and Ernst 2003). Three of these (scar- captured in drift fence studies in longleaf pine- let, eastern indigo, eastern coral) have disjunct dominated habitats in Florida (Enge 1997), but populations or close relatives in the arid south- most captures occur during a limited activity western United States, indicating that their period from April through August (Enge and ancestors may have come from arid habitats. Sullivan 2000). And, in fact, all three species prefer the dri- The eastern coral snake is mostly active in est habitats in longleaf pine savannas (Ernst early morning and late afternoon (Neill 1957; and Ernst 2003). The eastern indigo snake is Jackson and Franz 1981), and rarely after dark. here classified as a longleaf pine savanna spe- It is an inveterate snake-eater, with nearly ev- cialist because, in the northern four-fifths of ery longleaf pine savanna snake recorded in its its range, the indigo snake is restricted to the diet plus all the terrestrial lizards, the Florida vicinity of xeric longleaf pine–turkey oak sand- worm lizard, and an occasional eastern coral hills inhabited by the gopher tortoise (Diemer snake (Ernst and Ernst 2003). The surface ac- and Speake 1983; Moler 1985a; Stevenson tivity of the eastern coral snake has been stud- et al. 2003). This, the longest snake in the ied, but almost nothing is known about its sub- United States and Canada, is the top preda- terranean habits, home range, or the acreage of tor among snakes, eating any vertebrate it can suitable longleaf pine savanna habitat that will overpower, especially other snakes, including sustain a minimal viable population (Jackson venomous snakes (Keegan 1944; Belson 2000; and Franz 1981). Ernst and Ernst 2003). Summer home ranges The pine snake (Pituophis melanoleucus in- have been calculated at about 20–200 ha in cluding the races mugitus, lodingi, and popula- the Gulf Hammock region of northwest penin- tions of melanoleucus in the Coastal Plain of the sular Florida (Moler 1985b, 1992), to nearly Carolinas) is a large-bodied, spectacular diur- 400 ha in northern Georgia (Natalie Hyslop nal snake that makes its home in longleaf pine personal communication). Densities of the in- savannas where the southeastern pocket go- digo snake have not been determined, but pher lives. In Florida it lives in xeric habitats there is some indication that its ophiophagy such as longleaf pine–turkey oak, sand pine and agonistic behavior might prevent many scrub, pine flatwoods on well-drained soils, from living together in the same area (Means and oldfields on former sandhill sites (Franz in preparation). Behavior-mediated low densi- 1992). Females occupy home ranges of about ties of indigo snake populations might be a ma- 12 ha and males between 24 and 96 ha, with jor factor responsible for the federally threat- greatest activity in May, June, July, and Oc- ened status of the indigo snake (Means in tober (Franz 1992). It digs open southeast- preparation). ern pocket gopher mounds using its pointed Scarlet and eastern coral snakes are two nose and remains underground up to 85% small fossorial snakes that are rarely seen of the time in the tunnels of the southeast- but not necessarily uncommon. The scarlet ern pocket gopher, and to a lesser extent, the snake specializes on eating predominantly rep- burrows of the gopher tortoise (Franz 1992; tile eggs including those of the southeastern Tennant 1997). The diet of the pine snake is five-lined skink, six-lined racerunner, ring- ground-dwelling birds and their eggs, south- neck snake, corn snake, pine snake, turtles, eastern pocket gophers, mice, woodrats, and and even eggs of its own kind (Ernst and Ernst immature rabbits (Ernst and Ernst 2003). 172 II. Ecology

Resident Snakes Common information is available on the spatial relation- in Other Habitats ships (home range, movements) of this diurnal species, but Stickel and Cope (1947) reported Eleven other snakes are found to varying de- that a recaptured individual had moved only grees in longleaf pine savannas: rat snake, corn 100 m in 1.8 years. It preys on a wide range of snake, common kingsnake, brown snake, red- small vertebrates, especially snakes and lizards, bellied snake, southeastern crowned snake, but rats, mice, terrestrial birds, and the eggs of rough earth snake, smooth earth snake, birds, lizards, and turtles are of secondary im- copperhead, timber rattlesnake, and pigmy portance (Ernst and Ernst 2003). rattlesnake. The wide-ranging rat snake comes Two other large snakes, the timber rat- in several color varieties in the Coastal Plain, tlesnake and copperhead, are found in longleaf ranging from the Texas rat snake in Texas and pine savannas in the upper Coastal Plain, do Louisiana; the gray rat snake in Mississippi, not range very far into the Florida peninsula Alabama, and the Florida Panhandle; and the (Conant and Collins 1998), but do reach the yellow rat snake in peninsular Florida to the Atlantic and Gulf coasts both northeast and black rat snake in the Carolinas (Conant and west of Florida. Both of these species include Collins 1998). It is the most arboreal snake in more mesic hardwood forests in their habi- the United States and Canada (Jackson 1970, tat repertoires. The timber rattlesnake feeds 1974, 1977, 1978; Jackson and Dakin 1982; mostly on mammals but occasionally takes Dodd and Franz 1995) and is the most com- birds (less than 15%), whereas the copper- mon snake found entering homes (Dodd and head feeds on anything it can overpower with Franz 1995). Its diet consists mainly of small its venom, including vertebrates and inverte- mammals and birds (Brown 1979; Ernst and brates (Ernst and Ernst 2003). Studies of the Ernst 2003). A number of studies have been spatial relationships and population biology in published on the habitat and behavior outside longleaf pine savannas are wanting for both of the range of longleaf pine savannas, but Franz these species. (1995) found that a north-central Florida pop- The Coastal Plain contains most of the distri- ulation used winter and summer activity areas bution of the small pigmy rattlesnake, which connected by extensive migratory corridors. lives in all the longleaf pine savannas, turkey The corn snake, also wide ranging but with oak sandhills, and scrubs but is never very far no presently recognized subspecies, is a close from water (Ernst and Ernst 2003). Pigmys eat relative that also has the ability to climb, but a wide range of prey from small mammals, it is more fossorial in its spatial habits in lon- small snakes, and frogs to invertebrates (Ernst gleaf pine savanna habitats (Dodd and Franz and Ernst 2003). Their numbers are often quite 1995). It also feeds on mammals, birds, and abundant locally (May et al. 1996, 1997). eggs, but takes a high percentage of lizards and Five longleaf pine savanna resident snakes frogs (Ernst and Ernst 2003). The corn snake are small species that have not been exten- moved long distances between summer and sively studied (brown, red-bellied, southeast- winter activity areas, much like the rat snake ern crowned, rough earth, and smooth earth on the same study area in north-central Florida snakes). Because of their small size, the acreage (Franz 1995). The common kingsnake, a large required to maintain a breeding population is terrestrial constrictor, is fossorial in its habits probably much smaller than for larger snakes. (Krysko and Smith 2005). It is even more wide Earthworms seem to be the principal dietary ranging than the two previous species (Pacific item for all five species, and they all proba- to Atlantic coasts and dipping into northern bly eat other invertebrates, but more complete Mexico) and has four subspecies in the range diet studies are needed (Ernst and Ernst 2003). of longleaf pine, all of which principally live Population densities of these small snakes in in longleaf savannas (Mount 1975; Dundee longleaf pine savannas seem to be relatively and Rossman 1989; Palmer and Braswell 1995; sparse in the brown snake (Clark 1949; Ford Krysko 2001; Means and Krysko 2001). Little et al. 1991; Dalrymple et al. 1991) and 6. Vertebrate Faunal Diversity 173 red-bellied snake (Semlitsch and Moran 1984; ish mostly in deserts and xeric conditions (Pi- Ford et al. 1991); not especially dense in anka 1973, 1975). Fifteen lizards and one am- the smooth earth snake (Ernst and Ernst phisbaenian (a worm lizard, closely related to 2003); locally abundant in both the south- lizards) are common residents of longleaf pine eastern crowned snake (Neill 1951a; Brode savannas (Table 2). Of the three specialists, and Allison 1958; Semlitsch et al. 1981) and the mole skink (Fig. 3a) is a small, slender, Florida crowned snake (Campbell and Christ- fossorial lizard with a long tail. Its burrowing man 1982; Mushinsky 1984); and unstudied habits restrict it to friable, well-drained soils of and poorly known in the rough earth snake longleaf pine–turkey oak sandhills, scrub, and (Ernst and Ernst 2003). xeric hammocks (Mount 1963). Because it is so Eight snakes that are common residents of fossorial, it is rarely seen but can be raked from longleaf pine savannas have large geographi- bare soil exposed in mounds of the gopher tor- cal distributions in the eastern United States toise, southeastern pocket gopher, and scarab with extensions of their range into the prairies beetles (Geotrupinae) or found occasionally or grasslands of the Midwest. These snakes— under logs, boards, pieces of tin, and other ob- black racer and coachwhip (both active on jects (Mount 1963). Mount (1963) collected ground surface), corn snake, eastern hognose 326 of 422 (77%) specimens from push-up snake, scarlet snake, mole kingsnake, scarlet mounds of the southeastern pocket gopher and kingsnake, and pine snake (all burrowers or noted that the geographical distribution of the fossorial)—highlight the fact that longleaf pine mole skink almost exactly overlaps that of the savannas are prairielike grasslands. The black southeastern pocket gopher, both longleaf pine racer and coachwhip are fast-moving hunters savanna specialists. The mimic glass lizard is that actively hunt their prey. The remaining six found in longleaf pine flatwoods with wire- species are fossorial animals that spend a lot of grass ground cover (Moler 1992) in a narrow their time underground. strip of the Coastal Plain from North Carolina A number of species that are mostly asso- to Mississippi, excluding peninsular Florida ciated with wetland margins are frequently (Palmer 1987). It seems to be most common found in longleaf pine savannas. The eastern in the heliophilic ground cover of seepage bogs ribbon snake and common garter snake may (Paul Moler personal communication), but lit- forage for small vertebrates (frogs) in uplands tle else is known about it. The Florida worm far from water, as deduced from their pres- lizard is found principally in longleaf pine– ence in many longleaf pine savanna drift fence turkey oak sandhills and sometimes in xeric studies (Enge 1997). Likewise, the banded wa- hammocks and sand pine scrub, where it bur- ter snake (Nerodia fasciata) is often encoun- rows underground and feeds on earthworms, tered in longleaf pine savannas. Whether it small insects, and possibly termites and ants forages there for frogs, which are part of its (Ashton and Ashton 1985). Its relatively small diet (Ernst and Ernst 2003), or is merely pass- geographic distribution in the Coastal Plain is ing through en route to other water bodies, restricted to the northern half of the Florida as may also be the case for garter and rib- peninsula. bon snakes, is unknown. The cottonmouth Eleven other lizards are common residents usually brings to mind swamps and the edge in longleaf pine savannas (Table 2) or live in of aquatic habitats, but radiotelemetry studies adjacent habitats and are occasionally found have shown that some individuals also spend in the interface between longleaf pine habi- substantial amounts of their lives in uplands tats and evergreen shrub wetlands (coal skink), (Means 2005b). hardwood forests (five-lined skink, broadhead skink), or sand pine scrub (Florida scrub lizard). The eastern fence lizard is abundant Lizards in longleaf pine sandhills where turkey oaks Lizards are an impoverished group in the are small trees and at the edges of hardwood southeastern United States because they flour- hammocks (Crenshaw 1955). A relative, the 174 II. Ecology

Florida scrub lizard, is narrowly distributed in among all the lizards in longleaf pine savannas the Coastal Plain in the central Florida high- would make a worthwhile study. lands and peninsular coastal strand, where it lives primarily in sand pine scrub (Jackson 1973) but occasionally gets into adjacent lon- Turtles gleaf pine habitats (Carr 1940; Ashton and The eastern box turtle ranges widely through- Ashton 1985). The green anole may be the out the eastern United States and is found commonest lizard in the Southeast, occurring in many different habitat types from river in almost every habitat from cypress swamps swamps and other wetlands and hardwood to longleaf pine sandhills (Smith 1946). In forests to offshore barrier islands of the At- longleaf pine savannas, this arboreal lizard is lantic and Gulf coasts (Dodd 2001). It is abun- found on logs, dead trees, and where woody dant in longleaf pine savannas of all types but ground cover plants are abundant, such as saw probably is more common in flatwoods than palmetto, gallberry, scrub oaks, and others. sandhills. The fact that the eastern box turtle The six-lined racerunner, most abundant is one of the few vertebrates in longleaf pine in xeric longleaf pine habitats where ground savannas that suffer from considerable injury cover is sparse and bare patches of sandy and mortality from fire (Means and Campbell ground are exposed, is a fast-running, diur- 1982; Ernst et al. 1995) is probably an indica- nal species that is most active in the hottest tion that the eastern box turtle is really only part of the day (Smith 1946). It is completely a visitor in longleaf pine ecosystems, foraging terrestrial, preferring bare soil conditions that there commonly out of the many stringers of are best available in the year or two after swampy hardwoods along sluggish flatwoods ground fires in longleaf pine savanna (Mushin- streams, and out of the southern temperate sky 1985) as well as in sand pine scrub (Green- hardwood forests (Platt and Schwartz 1990) berg et al. 1994). The little ground skink is that are interspersed among clayhills and sand- another ecologically widespread lizard that is hills. Its great longevity and the abundance of found wherever there is some ground cover, wetland and hardwood forest habitats enable such as dead grass and leaf litter (Ashton and the survival of those segments of its Coastal Ashton 1985). In longleaf savannas, it is com- Plain populations that experience heavy mor- monest where leaf litter is built up under scrub tality from fire. oaks. Brooks (1967) estimated population den- On the other hand, the gopher tortoise is sity for the ground skink in a hardwood for- a strict longleaf pine savanna specialist that est at about 649 lizards per hectare, a figure avoids fire, summer heat, winter cold, and he thought was low. The southeastern five- predators by excavating burrows up to 14.5 m lined skink is abundant in longleaf pine sa- long and 3 m deep (Hallinan 1923; Hansen vannas and prefers down logs or snags where 1953; Auffenberg 1969). The gopher tortoise it hides under exfoliating dead bark and for- (Fig. 4) is one of the important keystone verte- ages on dead down or standing tree trunks and brates (Eisenberg 1983) in longleaf pine savan- in litter (Mushinsky 1992). Three glass lizards nas, because its burrows are long-lasting mi- besides the mimic glass lizard are occasionally crocaverns that are utilized by more than 300 found in longleaf pine savannas (Ashton and species of other vertebrates and invertebrates Ashton 1985). (Jackson and Milstrey 1989). It is the primary Many species of snakes, small mammals, and grazer in longleaf pine savannas (Landers and birds of prey feed on lizards, all of which are in- Speake 1980), eating a variety of broad-leaved sectivores. Although probably not contributing grasses, legumes, fleshy fruits, mushrooms, as much as amphibians to the overall biomass and wiregrass (Garner and Landers 1981). of the vertebrate predator–prey web in longleaf Auffenberg (1969) speculated that the gopher pine savannas, lizards may be more abundant, tortoise is important in the dispersal of seeds of and important, than is realized. The spatial, plants that it eats and that pathways radiating temporal, and dietary relationships compared away from its burrows are fertilized strips that 6. Vertebrate Faunal Diversity 175

FIGURE 4. Gopher tortoise, Gopherus polyphemus, at mouth of its burrow. The burrows of this keystone species, averaging about 8 m long and 2 m deep, are refugia for more than 300 species of other animals. A specialist in longleaf pine sandhills and clayhills. encourage the growth of these plants. Tortoise also move into the uplands during winter or burrows disrupt the ground cover, producing following dry-down of ponds, where they hi- local microhabitats that may be germination bernate or aestivate until weather and wa- sites for plants (Kaczor and Hartnett 1990; Her- ter levels are conducive to returning to their mann 1993; Platt 1999). Densities range to 20 ponds (Bennett et al. 1978; Burke and Gibbons gopher tortoises per hectare in longleaf pine– 1995). Another species that is a casual visitor turkey oak habitat but average 7.75 (Franz and in longleaf pine savannas for the purpose of Auffenberg 1978; Cox et al. 1987). nesting is the Florida softshell, Apalone ferox. Seven species of aquatic turtles commonly Fossil tortoises that might have played an im- live in the same small, temporary ponds portant role in longleaf pine savanna ecology utilized by amphibians in the Coastal Plain are discussed below. and nest in adjacent longleaf pine savannas (chicken turtle, Florida cooter, Florida redbelly turtle, yellowbelly slider, striped mud turtle, Birds eastern mud turtle, stinkpot). Most of these same turtles can be found in other water bod- Birds are the most vagile terrestrial vertebrates, ies, such as rivers and streams and swampy many of which engage in long-distance mi- lakes, so they are not specialists that live only grations from tropical winter habitats to habi- in a longleaf pine savanna landscape. How- tats far north of Coastal Plain longleaf pine sa- ever, they are highly mobile animals when out vannas to breed and raise young in summer of water and commonly migrate overland for (Stevenson and Anderson 1994). Migrants three reasons. First, individuals disperse over- pass through longleaf pine savannas on both land seeking water when ponds dry up; sec- trips, feeding as they progress. Because veg- ond, females of all species move into longleaf etative structural diversity has been impli- pine uplands to lay eggs in nests dug in friable cated in bird species richness (MacArthur upland soils; third, individuals of both sexes 1958; James 1971), one would not expect the 176 II. Ecology open-canopied, longleaf pine savannas to be cultural fields, and annually burned short- especially rich in resident birds, but the oppo- leaf/loblolly pine oldfields (Stoddard 1931). site is true. Engstrom (1993) found that bird It is a terrestrial species that feeds on seeds species richness of coniferous forests in Florida of all kinds, leafy vegetation, fruits, grasses, was not lower than that of deciduous forests. acorns, and insects (Ehrlich et al. 1988). In optimal longleaf pine habitat, the northern bobwhite can reach densities of up to five birds Longleaf Pine Specialists per hectare (Kellogg et al. 1972), but densities Only five species of birds can be considered as are usually lower than this. A variety of food specialists in longleaf pine savannas (Table 3). plants, insects for young birds, and adequate The northern bobwhite had few other open- cover in which to hide are the most important canopied savanna habitats available to it in the habitat requirements, both of which are elimi- presettlement Coastal Plain, so it was primar- nated during plant succession leading to hard- ily found in longleaf pine savannas in spite of woods in the absence of fire (Stoddard 1931). its common occurrence in present-day, open, The red-cockaded woodpecker is undoubt- grassy ruderal habitats such as pastures, agri- edly the bird species most specialized for living

, TABLE 3. Characteristic birds of longleaf pine savannas.a b Birds—residents 1. Aix sponsa Wood duck 2. Cathartes aura Turkey vulture 3. Buteo jamaicensis Red-tailed hawk 4. Buteo lineatus Red-shouldered hawk 5. Falco sparvarius American kestrel 6. Colinus virginianus Northern bobwhite s 7. Zenaida macroura Mourning dove 8. Bubo virginianus Great horned owl 9. Melanerpes carolinus Red-bellied woodpecker 10. Melanerpes erythrocephalus Red-headed woodpecker 11. Picoides pubescens Downy woodpecker 12. Picoides villosus Hairy woodpecker 13. Picoides borealis Red-cockaded woodpecker s 14. Colaptes auratus Northern ﬂicker 15. Dryocopus pileatus Pileated woodpecker 16. Campephilus principalis Ivory-billed woodpecker∗ 17. Cyanocitta cristata Blue jay 18. Corvus brachyrhynchos American crow 19. Parus carolinensis Carolina chickadee 20. Parus bicolor Tufted titmouse 21. Sitta carolinensis White-breasted nuthatch s 22. Sitta pusilla Brown-headed nuthatch s 23. Thryothorus ludovicianus Carolina wren 24. Sialia sialis Eastern bluebird 25. Dumetella carolinensis Gray catbird 26. Mimus polyglottos Northern mockingbird 27. Toxostoma rufum Brown thrasher 28. Lanius ludovicianus Loggerhead shrike 29. Dendroica pinus Pine warbler 30. Geothlypis trichas Common yellowthroat 31. Cardinalis cardinalis Northern cardinal 32. Pipilo erythrophthalmus Rufus-sided towhee 33. Aimophila aestivalis Bachman’s sparrow s 34. Sturnella magna Eastern meadowlark 35. Agelaius phoeniceus Red-winged blackbird 36. Quiscalus quiscula Common grackle 6. Vertebrate Faunal Diversity 177

TABLE 3. (Continued ) Birds—breeders 1. Coragyps atratus Black vulture 2. Meleagris gallopavo Wild turkey 3. Columbina passerina Common ground-dove 4. Coccyzus americanus Yellow-billed cuckoo 5. Strix varia Barred owl 6. Chaetura pealagica Chimney swift 7. Chordeiles minor Common nighthawk 8. Caprimulgus carolinensis Chuck-will’s-widow 9. Contopus virens Eastern wood-pewee 10. Myiarchus crinitus Great crested flycatcher 11. Tyrannus tyrannus Eastern kingbird 12. Progne subis Purple martin 13. Corvus ossifragus Fish crow 14. Polioptila caerulea Blue-gray gnatcatcher 15. Hylocichla mustelina Wood thrush 16. Vireo griseus White-eyed vireo 17. Vireo flavifrons Yellow-throated vireo 18. Vireo olivaceus Red-eyed vireo 19. Parula americana Northern parula 20. Dendroica discolor Prairie warbler 21. Dendroica dominica Yellow-throated warbler 22. Wilsonia citrina Hooded warbler 23. Icteria virens Yellow-breasted chat 24. Piranga rubra Summer tanager 25. Guiraca caerulea Blue grosbeak 26. Passerina cyanea Indigo bunting 27. Spizella pusilla Field sparrow 28. Molothrus ater Brown-headed cowbird 29. Icterus spurius Orchard oriole Birds—Winter Visitors 1. Sphyrapicus varius Yellow-bellied sapsucker 2. Ectopistes migratorius Passenger pigeon∗ 3. Sayornis phoebe Eastern phoebe 4. Tachycineta bicolor Tree swallow 5. Certhia americana Brown creeper 6. Sitta canadensis Red-breasted nuthatch 7. Troglodytes aedon House wren 8. Regulus satrapa Golden-crowned kinglet 9. Regulus calendula Ruby-crowned kinglet 10. Catharus guttatus Hermit thrush 11. Turdus migratorius American robin 12. Bombycilla cedrorum Cedar waxwing 13. Vireo solitarius Solitary vireo 14. Vermivora celata Orange-crowned warbler 15. Dendroica coronata Yellow-rumped warbler 16. Dendroica palmarum Palm warbler 17. Spizella passerina Chipping sparrow 18. Melospiza melodia Song sparrow 19. Melospiza georgiana Swamp sparrow 20. Junco hyemalis Dark-eyed junco 21. Zonotrichia albicollis White-throated sparrow 22. Carduelis pinus Pine siskin 23. Carduelis tristis American goldfinch

a Modiﬁed from Engstrom (1993). b s = endemic longleaf pine Savanna specialists; ∗=recently extinct. 178 II. Ecology

FIGURE 5. Red-cockaded woodpecker, Picoides borealis, is the only North American bird that excavates its nest cavity exclusively in living trees. The long-persisting cavities of this keystone species are homes to dozens of other animals in the long life of the longleaf pine tree in which the cavity is made. A specialist in longleaf pine savanna. in longleaf pine savannas (Conner et al. 2001) suppression resulting in midstory develop- (Fig. 5). The only woodpecker that excavates ment, and the elimination of old trees (over nesting and roosting cavities in a living tree, 100 years of age) have been implicated in it selects longleaf pine trees 90 years or older the decline of the red-cockaded woodpecker (Costa 1995), presumably because longleaf (Baker 1981; Anonymous 1990; U.S. Fish and pines above 90 years of age are more likely Wildlife Service 2003). to have heart rot disease which makes the ex- Bachman’s sparrow is a characteristic species cavation of nest cavities easier (Hooper et al. and permanent resident of longleaf pine savan- 1991). It forages mainly on the trunks of the nas, “dependent through the ages on the fre- longleaf pine, where its main food appears to quent grass fires that kept the ‘flatwoods’ open be arboreal ants, especially Crematogaster ash- and parklike, with the characteristic prairie- meadi (Hess and James 1998). Studies have type flora” (Stoddard 1978). It has survived shown that the red-cockaded woodpecker sur- reduction of its longleaf pine habitat by ac- vives better with little or no understory (Van commodating to ruderal situations in short Balen and Doerr 1978), which was highly sig- herbaceous vegetation of open woodlands, nificantly related to the ground cover compo- borders of cultivated fields, or even fence rows sition and the extent of natural pine regen- with other sparrows. Bachman’s sparrow nests eration, both of which are indirect indicators and forages in dense ground cover of open of local fire history (James et al. 1997). The pine forests. It will breed in clearcuts but red-cockaded woodpecker is another impor- prefers open mature stands of timber with low, tant keystone species in longleaf pine savan- thick ground cover (Dunning and Watts 1990; nas, because it makes cavities in living trees Engstrom 1993). that may persist for several hundred years In the middle of the geographic range of lon- and are used by many other animals over the gleaf pine in south Georgia, the white-breasted long life of the tree. Forest fragmentation, fire nuthatch prefers mixed pine and hardwoods 6. Vertebrate Faunal Diversity 179 and is most numerous in big timber, whereas bark-gleaning, sallying forth for flying in- the brown-headed nuthatch is most abundant sects, ground-litter searching, and pecking into in older stands of longleaf and annually burned chambers made by wood-boring beetle larvae. shortleaf/loblolly pine stands (Stoddard 1978). And some are consummate aerial predators Both nuthatches nest in cavities they exca- that prey on mice, rats, and other small ver- vate in dead snags, and the brown-headed tebrates. Extinct bird species of the late Pleis- nuthatch, especially, forages on the trunks and tocene that may have been characteristic of among the limbs of pine trees. longleaf pine savannas are discussed below.

Other Characteristic Bird Mammals Species The role of mammals in grassland ecosystems Birds commonly found in longleaf pine savan- is unique. No other living group of vertebrates nas have been grouped as residents, breeders, produces species with such large body mass. and winter visitors (Table 3) (Engstrom 1993). Amphibians, reptiles, and birds generally have All five bird specialists are residents that also body masses in the range of about 10 g to about breed in longleaf pine savannas, but 32 other 10 kg, but mammals push the upper size lim- species live in these habitats year-round. Of the its to between 300 and 6500 kg (Nowak 1999) 29 breeding species, 21 migrate mostly to the in the largest herbivores. Continentwide since neotropics, and 8 either shift to other habi- at least the Middle Eocene, grasslands have tats or migrate to south Florida in the win- attracted large grazing animals (Webb 1977). ter (Robertson and Kushlan 1974; Engstrom The largest herbivore in longleaf pine savannas 1993). Some, if not most, of the south-central today is the white-tailed deer (Odocoileus vir- Florida prairie fauna that are dependent on ginianus), second to the American bison (Bison dry prairie also appear to be dependent on re- bison) that became extinct in the Coastal Plain cently burned sites (i.e., Florida grasshopper in the early 1800s (Rostlund 1960; Humphrey sparrow) or prefer recently burned sites over 1992b). While many small and medium-sized fire-suppressed prairies (Walsh et al. 1995). mammals are commonly found in longleaf Tucker and Robinson (2003) studied the pine savannas, a strikingly obvious dearth of influence and frequency of prescribed burn- megafaunal mammals is evident (Layne 1974; ing on Henslow’s sparrows (Ammodramus Stout and Marion 1993). I discuss the extant henslowii) wintering on Gulf Coast pitcher mammals in this section and the fossil record plant bogs, communities that grade downslope of mammals and other vertebrates afterward. out of longleaf pine savannas and depend on fires sweeping into them from the uplands. Longleaf Pine Specialists They found that bogs burned during winter typically hosted Henslow’s sparrows for only Thirty-six mammals are found commonly in one winter, but bogs burned during the grow- longleaf pine savannas, but only three are spe- ing season hosted the species for at least three cialists in them (Table 4). The fox squirrel (Sciu- winters. rus niger) is an animal of mature open longleaf Birds perform many functions in longleaf pine flatwoods, sandhills, and clayhills (Low- pine ecosystems. Many are seed and fruit ery 1974; Brown 1997). It forages extensively eaters that disperse seeds that pass through on the ground and does not do well in brushy, the gut unscathed. The rank growth of black thick ground cover. Periodic fires that kill back cherry (Prunus serotina), sassafras (Sassafras al- encroaching hardwoods or keep scrub oaks bidum), or persimmon (Diospyros virginiana) suppressed are absolutely necessary for fox along fence rows confirms this. Other birds squirrel populations to thrive (Brown 1997). are insectivores that consume large quanti- Weigl et al. (1989) hypothesized that there ties of arthropods by means of foliage- and may be a coevolved interdependence between 180 II. Ecology

, TABLE 4. Characteristic mammals of longleaf pine savannas.a b Mammals 1. Didelphis virginiana Virginia opossum 2. Sorex longirostris Southeastern shrew 3. Blarina carolinensis Southern short-tailed shrew 4. Cryptotis parva Least shrew 5. Scalopus aquaticus Eastern mole 6. Pipistrellus subflavus Eastern pipistrelle 7. Myotis austroriparius Southeastern myotis 8. Lasiurus borealis Red bat 9. Lasiurus seminolis Seminole bat 10. Lasiurus intermedius Yellow bat 11. Nycticeius humeralis Evening bat 12. Dasypus novemcinctus Nine-banded armadillo 13. Sylvilagus floridanus Eastern cottontail 14. Sciurus carolinensis Gray squirrel 15. Sciurus niger Fox squirrel s 16. Glaucomys volans Southern flying squirrel 17. Geomys pinetus Southeastern pocket gopher s 18. Geomys bursarius Plains gopher 19. Sigmodon hispidus Hispid cotton rat 20. Reithrodontomys humulis Eastern harvest mouse 21. Reithrodontomys fulvescens Fulvous harvest mouse 22. Oryzomys palustris Marsh rice rat 23. Podomys floridanus Florida mouse s 24. Peromycus polionotus Oldfield mouse 25. Peromycus gossypinus Cotton mouse 26. Ochrotomys nutallii Golden mouse 27. Baiomys taylori Northern pygmy mouse 28. Microtus pinetorum Pine vole 29. Procyon lotor Raccoon 30. Mustela frenata Long-tailed weasel 31. Mephitis mephitis Striped skunk 32. Spilogale putorius Spotted skunk 33. Canis latrans Coyote 34. Vulpes vulpes Red fox 35. Urocyon cinereoargenteus Gray fox 36. Felis concolor Cougar 37. Lynx rufus Bobcat 38. Ursus americanus Black bear 39. Sus scrofa Feral pig 40. Odocoileus virginianus White-tailed deer

a Adapted from Engstrom (1993). b s= endemic longleaf pine savanna specialists. longleaf pine, the fox squirrel, and hypogeous roots of the plants while the fungus receives fungi. In the longleaf pine ecosystem, food ap- carbohydrates and a substrate on which to live. pears to be the most important factor influenc- Most of these fungi depend upon animals for ing populations of the large fox squirrel, espe- dispersal of their spores, which Weigl et al. cially the seeds of longleaf pine and at least (1989) found in the gut of every fox squirrel eight genera of hypogeous fungi. The fungi they examined. The fitness of all three part- form mutualistic mycorrhizal associations with ners seems to be enhanced by the relationship, longleaf pine and probably other species of but the degradation of the longleaf pine ecosys- plants in the longleaf ecosystem, increasing tem by plant succession, logging, and fragmen- the surface area and nutrient absorption of the tation has had a deleterious effect upon the 6. Vertebrate Faunal Diversity 181 fox squirrel (Weigl et al. 1989), which in turn rodent’s tunnels (Franz 2005). The southeast- could feed back negatively on the fungi and ern pocket gopher is so specialized for living in those plants with which the fungi form myc- longleaf pine savannas that Gates and Tanner orrhizal associations. (1988) predicted that if succession toward a The Florida mouse is narrowly restricted to hardwood hammock caused by lack of fire con- fire-maintained, xeric, upland vegetation on tinued unimpeded, southeastern pocket go- deep, well-drained sandy soils covered with pher populations would decline. Apparently, longleaf pine–turkey oak, sand pine scrub, declines have been underway for some time. south Florida slash pine–turkey oak, coastal Entomologists Peter Kovarik and Paul Skel- scrub, and drier pine flatwoods vegetations ley, studying arthropods associated with the (Layne 1992). It has a small range confined southeastern pocket gopher, were unable to to the upper one-half of the Florida peninsula, locate populations at many historical localities but populations follow the sandy ridges par- (Bailey 2001). alleling the coastlines to Sarasota County on the west and Dade County on the east (Layne Other Characteristic Mammal 1992; Brown 1997). The Florida mouse inhabits side passages it excavates in the bur- Species rows of the gopher tortoise, and it forages at Although only three mammals are special- night on acorns, pine seeds, palmetto berries, ists in longleaf pine ecosystems, some 37 mushrooms, insects, and arthropods found in other species are commonly found in them gopher tortoise burrows (Layne 1992; Layne (Table 4). Six of these are insectivorous bats and Jackson 1994). On the Ordway Preserve in that take their insect prey on the wing. Four Putnam County, FL, where the species is most (three shrews and a mole) are tiny fossorial common in high pine communities with the or semifossorial carnivores that feed on ter- gopher tortoise, home ranges were unaltered restrial invertebrates. The gray squirrel is a after prescribed burns, and no fire mortality visitor from hardwood forests, but the south- was noted (Jones 1995). ern flying squirrel reaches highest densities in The southeastern pocket gopher (Fig. 6) mature pine-oak (longleaf pine/turkey oak) is found exclusively in sandy soils of lon- woodlands and xeric oak hammocks, where gleaf pine savannas in sandhills and xeric it commonly utilizes snags and woodpecker flatwoods from central Alabama and Georgia cavities (Brown 1997). The eastern cotton- south to central Florida (Humphrey 1992a; tail lives in virtually all upland communi- Brown 1997). It has been able to accommo- ties except dense forest, feeding preferentially date to man-made vegetation on sandy soils, on legumes, grasses, and various broad-leaved such as golf courses, lawns, pastures, road- forbs (Brown 1997). Ten species of rats and side rights-of-way, and other open, grassy ar- mice are common in longleaf pine savannas, eas. Because of its subterranean habits, lit- especially white-footed mice (Peromyscus spp.) tle is known about the key behaviors of the and the cotton rat. The cotton rat can reach southeastern pocket gopher, but southeastern high densities (25–35/ha) at times and feeds pocket gophers retard eluviation by returning on leaves, stems, roots, and seeds of grasses, leached materials to ground surface; in one sedges, legumes, and other herbaceous plants; study, casting up 81,600 kg per hectare of bur- it can be quite destructive to agricultural crops row soil per year (Kalisz and Stone 1984). but it is also a staple food item for most of Their sandy push-up mounds create microsites the Coastal Plain’s mammal, reptile, and avian that may promote colonization and secondary predators (Brown 1997). succession of a rich herbaceous flora in longleaf The largest mammal in longleaf pine sa- pine savannas (Platt 1975, Hermann 1993, vannas is the black bear, an omnivore Simkin and Michener 2005). Pocket gophers that eats a large percentage of fruits, nuts, serve as the principal prey of the pine snake, berries, and other plant material (Lowery which spends about half of its life in the 1974; Brown 1997), but will take vertebrates 182 II. Ecology

FIGURE 6. The characteristic push-up mounds of the southeastern pocket gopher, Geomys pinetus, a specialist in longleaf pine savanna, especially sandhills and clayhills. Its burrows provide refuge for more than 60 species of arthropods. opportunistically (Rogers 1999). The largest 80 individuals) remains in southern Florida strict carnivore is the cougar, but it has nearly (Wood 2001). All the rest of the longleaf pine been extirpated from the entire eastern United savanna mammals are small, generalized car- States and only a small population (fewer than nivores (opossum, raccoon, long-tailed weasel, 6. Vertebrate Faunal Diversity 183 striped skunk, spotted skunk, red fox, gray fox, birds, but not to most of the amphibians and bobcat), and three are recent arrivals (feral pig, reptiles. nine-banded armadillo, coyote). At the end of the North American Pleis- The white-tailed deer is the only large, her- tocene, 8 families, 46 genera, and 191 species bivorous, native, ungulate mammal that has of mammals became extinct (Holman 1991). the potential to impact vegetation. Mainly a Two families and 19 genera of North American browser that feeds on twigs, leaves, and ten- birds became extinct, with a very large num- der shoots of many trees and shrubs, it also ber of extinct species reported (Holman 1991). is a grassland feeder spending time in hard- In sharp contrast, no families and no genera of wood forests as well as longleaf pine savannas either amphibians or reptiles became extinct (Barlow and Jones 1965). That there are so in the Pleistocene of North America, and only few large, megafaunal mammals utilizing the 12 taxa out of a total of 229 identiﬁed Pleis- vegetative largess of longleaf pine savannas is tocene taxa were unquestioned extinct species highly unusual (Layne 1974; Stout and Mar- (Holman 1995). A list of Pleistocene megafau- ion 1993) and warrants an explanation (see nal vertebrates is given in Table 5 and those the following section). that may have been characteristic species of longleaf pine savannas are noted. Fossil vertebrates can give us information What Can the Vertebrate about two important aspects of longleaf pine savannas: the possible antiquity of savan- Fossil Record Tell Us about nas and how the savanna vegetation may Longleaf Pine Savannas? have been affected by the megafauna. Numer- ous studies of vertebrate fossils, mostly from The fossil record indicates that grazing mam- Florida localities, show that longleaf pine sa- mals have been evolving in midcontinental vanna specialists and residents were present North American woodland savannas (as op- at least as far back as the early Pleistocene, posed to treeless prairies or steppes) since at about 2 million years ago. The Inglis IA fauna least the middle Eocene (Webb 1977). By mid- from the earliest Irvingtonian in the earliest Miocene, a close faunal continuity was estab- Pleistocene consisted of an essentially modern lished between the Gulf Coast and the Great herpetofauna with 21 of 26 snake species, 2 of Plains, extending far south into Mexico by the 4 lizard species, and the Florida worm lizard late Miocene (Graham 1965; Webb 1977). On- surviving in Florida today (Meylan 1982). set of drier conditions in the Pliocene brought The combined evidence of the herpetofauna, about prairie or steppe conditions in the Great other vertebrates, and the sedimentary con- Plains, but the effects of summer rains, pre- text suggest a mixed habitat of mature lon- sumably, maintained a broad belt of savan- gleaf pine with xeric hammock interspersed nas around the Gulf of Mexico (Webb 1977). (Meylan 1982). Xeric hammock would be ex- Savannas, having scattered trees and riparian pected in the depressions characteristic of a gallery forests along the margins of streams karst topography, similar to that which ex- and swampy wetlands, support a more diverse ists locally today. Slightly later in the early herbivore fauna including browsing horses, Pleistocene, Leisey IA must have been adja- tapirs, most of the peccaries, most of the pro- cent to an upland community, most likely high boscideans, and many browsing ruminants pine or xeric hammock, to account for the (Webb 1977) in addition to the grazing mam- large number of highly terrestrial turtles, in- mals. All of the modern extant vertebrates that cluding Gopherus, Terrapene, and two species of inhabit longleaf pine savannas were present Hesperotestudo (Meylan 1995). by the end of the Pleistocene (Kurten and The bed of Rock Springs Run in Orange Anderson 1980; Holman 1995; Emslie 1998). County, FL, produced one of the richest avian In the past 11,000 years, however, something fossil deposits known from North America happened to the large mammals and large (Wolfenden 1959). Of 1025 bird bones, about 184 II. Ecology

TABLE 5. Late Pleistocene (Rancholabrean) vertebrates that may have been characteristic in longleaf pine savannas.a,b Turtles 1. Geochelone crassiscutata Eastern large tortoise s 2. Geochelone incisa Eastern small tortoise s 3. Terrapene carolina putnami Giant box turtle ? Birds 1. Phoenicopterus Flamingo d 2. Ciconia Stork ? 3. Anabernicula Shell duck d 4. Breagyps Condor s 5. Teratornis Teratorn s 6. Cathartornis Teratorn s 7. Spizaetus Hawk-eagle s 8. Amplibuteo Eagle s 9. Wetmoregyps Walking-eagle s 10. Neophrontops Old World vulture s 11. Neogyps Old World vulture s 12. Milvago Caracara s 13. Dorypaltus Lapwing ? 14. Burhinus Thick-knee ? 15. Protocitta Jay ? 16. Henocitta Jay ? 17. Cremaster Hangnest ? 18. Pandanaris Cowbird ? 19. Pyelorhamphus Cowbird ? Mammals 1. Dasypus bellus Beautiful armadillo s 2. Kraiglievichia floridanus Florida pampathere s 3. Holmesina septentrionalis Northern pampathere s 4. Glyptotherium floridanum Simpson’s glyptodont s 5. Megalonyx jeffersoni Jefferson’s ground sloth b 6. Megalonyx wheatleyi Wheatley’s ground sloth b 7. Eremotherium rusconii Rusconi’s ground sloth b 8. Glossotherium harlani Harlan’s ground sloth b 9. Homo sapiens Human b 10. Borophagus diversidens Bone-eating dog b 11. Canis lepophagus Johnson’s coyote b 12. Canis lupus Gray wolf b 13. Canis dirus Dire wolf b 14. Canis rufus Red wolf b 15. Ursus americanus Black bear b 16. Tremarctos floridanus Florida cave bear b 17. Arcodus pristinus Lesser short-faced bear b 18. Felis onca Jaguar b 19. Felis concolor Puma, cougar (Florida panther) b 20. Felis atrox Lion b 21. Felis yagouaroundi Jaguarundi b 22. Felis rufus Bobcat b 23. Felis amnicola River cat d 24. Smilodon fatalis Sabretooth b 25. Smilodon gracilis Gracile sabretooth b 26. Homotherium serum Scimitar cat b 27. Chasmaporthetes ossifragus American hunting Hyena b 28. Castoroides ohioensis Giant beaver d 29. Hydrochoerus holmesi Holmes’s capybara d 30. Neocherus pinckneyi Pinckney’s capybara d 6. Vertebrate Faunal Diversity 185

TABLE 5. (Continued) 31. Cuvieronius tropicus Cuvier’s gomphothere b 32. Mammuthus columbi Columbian mammoth s 33. Mammut americanum American mastodon b 34. Tapirus veroensis Vero tapir b 35. Nannippus phlegon Gazelle-horse s 36. Equus simplicidens American zebra s 37. Equus fraternus Brother horse s 38. Equus giganteus Giant horse s 39. Platygonus bicalcaratus Cope’s peccary b 40. Platygonus compressus Flat-headed peccary b 41. Mylohyus ﬂoridanus Kinsey’s peccary b 42. Hemiauchenia blancoensis Blanco llama s 43. Hemiauchenia macrocephala Large-headed llama s 44. Paleolama miriﬁca Stout-legged llama s 45. Odocoileus virginianus White-tailed deer b 46. Capromeryx arizonensis Skinner’s pronghorn s 47. Ovibos sp. Muskox b 48. Bison latifrons Giant bison s 49. Bison bison American bison s

a Modified from Brodkorb (1964), Webb (1974), Steadman and Martin (1984), Young and Laerm (1993). b b = prairie grasslands inhabitant; d = aquatic; s = endemic longleaf pine savanna specialists. half were identifiable to species and included which fossil red-cockaded woodpeckers could 35 species; some, including bones of the pas- have lived. The red-cockaded woodpecker re- senger pigeon, cardinal, and red-cockaded quires longleaf pines at least 90 years old for woodpecker, indicated the nearby presence of the construction of cavities (Costa 1995), the longleaf pine forest. The spring lies immedi- presence of red heart disease (Jackson 1977), ately east of the Pamlico shoreline (MacNeill a specialized diet of ants that only live on lon- 1950), indicating that the bones were de- gleaf pine trunks (Hess and James 1998), and posited during Pamlico or post-Pamlico times intolerance for closed canopy or brushy con- about 180,000 to 120,000 years ago (Webb and ditions (Van Balen and Doerr 1978). The same Wilkins 1984). can be said for another longleaf pine savanna By the late Pleistocene Sangamon Inter- specialist, the gopher tortoise, which is com- glacial age (about 100,000 years ago), all fam- mon in earliest Pleistocene sites. In addition, ilies and genera of the Williston IIIA her- the blind cave arthropods that use the bur- petofauna represent extant taxa, and only rows of the gopher tortoise, and the tortoise 1 of 36 taxa is extinct at the species level tick that is endemic on the gopher tortoise, re- (Holman 1996). A low-energy aquatic envi- quired some time to evolve, helping infer the ronment is suggested by some of the species, antiquity of longleaf pine savannas. Some sem- but the bulk of the species were character- blance of a fire-maintained, longleaf pine sa- istic of a distinctively dry and sandy lon- vanna, therefore, must have been present in gleaf pine savanna including eastern spade- the Coastal Plain for at least 2 million years and foot, eastern small tortoise (the extinct form), had antecedents far back into the Miocene. gopher tortoise, Florida worm lizard, eastern If we can be reasonably certain that lon- indigo snake, southern hognose snake, eastern gleaf pine savannas existed throughout the hognose snake, coachwhip, pine snake, Florida Pleistocene, then we can explain why present- crowned snake, and eastern diamondback rat- day longleaf pine savannas are wanting in tlesnake (Holman 1996). megafaunal vertebrates that are to be expected The red-cockaded woodpecker is so special- in savannas having an abundance of grasses ized that it is difficult to imagine any kind of and forbs for grazing animals and woody plants pine forest other than longleaf pine savanna in for browsing animals. Longleaf pine savannas 186 II. Ecology

FIGURE 7. Pleistocene megafauna. Possibly 49 extinct species of large ungulates and carnivores, 19 species of extinct birds, and 3 giant turtles were inhabitants of longleaf pine savanna habitats prior to about 11,000 years ago. Here illustrated is the skull of one of the extinct large carnivores, the sabertooth (Smilodon fatalis). How longleaf pine savanna habitats were structured in the presence of these animals, and how the longleaf pine savanna vegetation has responded to their extinction, may never be known. did, in fact, have a large megafauna that the livestock digestive systems (Janzen 1984). fossil record tells us went extinct rather rapidly Some of the herbaceous plants of longleaf in an approximately 2000-year period from pine savannas may well have depended on about 13,000 to 11,000 years ago (Webb 1990) large mammals for seed dispersal to certain (Fig. 7). The present-day biota of longleaf kinds of high-quality germination sites (e.g., pine savannas evolved in concert with the edges of game trails, nitrogen-rich soils below megafauna, and we should wonder how many legume trees or dung piles, heavily insolated of their adaptations that we see, or do not sites, watercourse edges, gopher tortoise bur- see, are ghosts of evolution past (Janzen and row aprons, tree tip-up mounds), or for seed Martin 1982; Barlow 2000). Not only should removal from sites of easy harvest by small we ask how different must have been the veg- predators that concentrate on harvesting seeds etative characteristics of longleaf pine savan- from maturing fruits near ground level (Janzen nas during the late Pleistocene, but also how 1984). different must have been presettlement lon- Mastodon dung found preserved under wa- gleaf pine savannas from their naturally co- ter in the Aucilla River, FL, consisted of evolved conditions with the megafauna that cypress (Taxodium), wild grape (Vitis), button- were present only a few thousand years ear- bush (Cephalanthus occidentalis), willow (Salix), lier? pine (Pinus), pokeweed (Phytolacca americana), The vegetative characteristics of Pleistocene Mexican poppy (Argemone), and wild gourd longleaf pine savannas must have been greatly seeds (Cucurbit pepo) (Webb et al. 1992). Most affected by the megafauna. Many small herb of these plants indicate a browsing diet, which seeds survive the trip through cattle and other corresponds well with the bunodont dentition 6. Vertebrate Faunal Diversity 187 of the mastodon. The site of deposition of the the New World by the Spaniards, hogs have dung was the mouth of a cave that probably been observed to kill more than 17,000 two- served as a watering hole. Some of the seeds year-old, longleaf pine seedlings per hectare recovered are upland species that could not ex- at rates of more than 200 seedlings per day, ist near the watering hole. Mastodons proba- so one hog can obliterate a hectare of planted bly went on feeding forays, ranging far from pines in 2 days (Simberloff 1993). On the De the water source as the food supply was dimin- Soto National Forest, hogs greatly damaged ished in an ever-growing ring centered around seedlings 5–120 cm and saplings 1.5 – 4.5 m the watering hole during the driest months of high. Chapman (1943) and Bruce (1947) de- the year. scribed equally devastating damage by hogs in There were no less than three proboscideans experimental plots in Louisiana. It may well be in the southeastern United States Coastal Plain the case that longleaf pine ecosystems in the in the late Pleistocene, the Columbian mam- late Pleistocene were typically very sparsely moth, American mastodon, and Cuvier’s gom- treed, and that longleaf pine was only able phothere, whereas of the two living elephants, to become established when migratory herds only one species occurs today in either the missed finding regeneration for a few years. savannas of Africa or India. Like living ele- One could even argue that many of the adap- phants, mammoths, mastodons, and gom- tations of longleaf pine such as masting, grow- photheres could have debarked, stripped, and ing a deep, strong taproot, living to 500 years pushed over trees, thus creating a substan- of age, and having dense, strong wood were tial impact on the environment. Owen-Smith evolved to avoid predation by megaherbivores. (1987, 1988) pointed out that the feeding ac- One longleaf pine savanna resident verte- tivities of elephants quickly change wooded brate that had a strong impact on vegeta- savanna to open short-grass prairies domi- tion was humans, both directly and indirectly. nated by rapidly regenerating plants that pro- Probably immediately after arriving in the New vide food for smaller herbivores. Elimination World, humans directly impacted vegetation of elephants leads to massive environmen- by burning it to stampede game, remove hid- tal changes, and the extinction of megaherbi- ing places for small game, make walking and vores at the end of the Pleistocene may have hunting easier, and attract game to nutri- caused changes from grasslands to forests (Hol- tious new growth following fire (Pyne 1982). man 1995). Pine- and oak-dominated savan- Humans may have had an even greater im- nas were present in the Southeast during the pact on vegetation indirectly through preda- Wisconsin glaciation (Delcourt and Delcourt tion on the megafauna. When Europeans first 1987). Pine and oak pollen alternated in dom- arrived, Amerindians were still using stone im- inance in lake sediments over the past 40,000 plements. After the earliest and unique Clovis, years in north Florida (Watts 1969, 1971, 1975, Suwannee, and Simpson stone technology of 1980; Watts et al. 1992), with pine pollen be- Paleoindians (about 13,000 to 10,000 years coming dominant about 8000 years ago (Watts before present) that is thought to have been et al. 1992). Some have ascribed the ascen- used to bring down the megafauna in the dancy of pine pollen in lakes to climate change, southeastern United States (Goodyear 1999), but the largest shift to pines occurs shortly after spear points went through an evolution of the extinction of the megafauna. types. These ranged from Bolen Age (10,000 In the Pleistocene, there were ground sloths, to 9000 years before present) projectile points peccaries, and proboscideans that could have that were corner and side-notched to Ar- dug, rooted, or pulled up the large grass-stage chaic Period (about 9000 to 3200 years before longleaf pine seedlings and saplings to get at present) projectile points that were stemmed. the cortex of the main taproot. A modern ex- Then, woodland technology (about 3200 to ample verifies the palatability of longleaf pine 1000 years before present) was character- taproots and the likelihood that this scenario ized by basally notched stone spear points. is correct. Following their introduction into Finally, the Mississippian Culture (1000 to 188 II. Ecology

600 years before present) was characterized savannas. The demise of most of them proba- by temple mound complexes, farming, and bly followed the extinction of their megafaunal small stone points that were true arrow- prey. heads. The evolution of stone spear points In summary, longleaf pine savannas are from very well crafted large, lanceolate forms grasslands whose antecedents may go back at to small points suitable as arrow tips, proba- least into the early Miocene in Florida and bly reflected the change in human diet from probably earlier in other parts of the Coastal megafaunal animals to very small animals and Plain. Savanna vegetation has been present fishing. in some parts of the Coastal Plain (especially Some of the Coastal Plain megafauna sur- Florida where the fossil evidence is most abun- vived to about 11,000 years ago (Grayson dant) probably for much of the Cenozoic Era. 2001). Paleoindians, arriving in North America During climate changes in the Pleistocene, the about 13,000 years ago, lived contemporane- principal upland plant communities probably ously for about 2000 years with the Coastal expanded or contracted their local distribu- Plain megafauna and could have been the tions such as by moving farther upslope in the main reason for its extinction (Martin 1967). case of the southern temperate hardwood for- One way to understand how this might be pos- est or shrinking to ridge tops and deeper sands sible is to calculate how rapidly a small band in the case of longleaf pine savannas. Plant of, say, 20 humans (10 reproductive males and species composition may also have changed females) could have multiplied in the pres- somewhat with the arrival or departure of ence of the huge bounty of large, easy-to- northern species in or out of either of the kill, megafaunal animals that were completely two above-mentioned ecosystems, but it is dif- naive about the danger presented by this newly ficult to believe that the fire-controlling re- arrived predator (Flannery 2001). Assume a lationships between fire-sensitive hardwood simple doubling of the human population over communities and fire-tolerant longleaf pine- a generational period of 20 years. After 20 grasslands ever disappeared. A large vertebrate generations in 400 years, the human popula- fauna of at least 212 extant species plus an ex- tion could theoretically multiply to more than tinct component of up to possibly 49 megafau- 20 million people. There are five periods of nal mammals, 3 giant tortoises and turtle, and 400 years in the 2000 years during which the between 9 and 17 birds evolved with the veg- megafauna went extinct, so there was ample etation of longleaf pine savannas. Not only time for human populations to become large did the extinct fauna influence the character- enough to have had a significant hand in their istics and distribution of the longleaf pine sa- demise, if not having been the sole reason. The vanna plants in the late Pleistocene, but an first to go would have been the easiest ver- even greater effect might have been imposed tebrates to overpower, such as the two giant on these plants by means of ecological release tortoises and one giant box turtle (Table 5). following the demise of these animals. What These animals had no defense against a preda- those effects might have been, we can only tor that could turn a tortoise on its back, fash- guess, but they would have been substantial. ion stone and wooden tools to kill it, break Even our conception of what longleaf pine sa- open its protective shell, and cut out the edi- vanna was like 500 years ago may be dramat- ble parts (Clausen et al. 1979). Likewise, other ically different from its late Pleistocene aspect. large, slow-moving animals such as the giant Interglacial periods, such as the Holocene in beaver, Florida pampathere, northern pam- which we presently live, last only about 10,000 pathere, and Simpson’s glyptodont would have years whereas glacial periods with epiconti- been similarly easy targets. Many of the late nental ice sheets are about 100,000 years long. Pleistocene birds that went extinct were large The evolutionary setting for vertebrates of the carrion eaters or birds of prey (Table 5) that longleaf pine savanna was most likely more required open savannas and probably scav- typical during the long glacial periods than the enged or hunted the ancient longleaf pine warmer interglacials. 6. Vertebrate Faunal Diversity 189 Management striped newts breeding in his study pond em- igrated in excess of 500 m from the pond. He Considerations from the believed that a core of protected upland with Perspective of Vertebrates a radius of approximately 800 m would be required to protect the vast majority of individ- Value of Upland Habitat uals that bred in the pond. Burke and Gibbons Surrounding Temporary Ponds (1995) studied the distance from a South Car- olina pond that three species of turtles (mud As we have seen above, 35 amphibians that turtle, Florida cooter, pond slider) migrated are characteristic species in longleaf pine sa- into the uplands to lay eggs or hibernate. Their vannas breed in temporary ponds. Semlitsch data indicated that the turtles required a 275-m (1998) calculated that most of the salaman- zone of uplands around the pond to protect ders spend 86–99% of their lives in the uplands 100% of the nest and hibernation sites. Data and less than 15% in breeding ponds. A similar on the amount of upland habitat needed for analysis has not been conducted for anurans, the local survival of populations of species that but because of the long captive life spans for utilize temporary ponds are urgently needed many species such as the eastern spadefoot (12 for most longleaf pine savanna species. The up- years), gopher frog (9 years), narrowmouth land habitats are not buffer or riparian “zones” toad (6 years, 9 months), barking treefrog (10 but full-fledged “core” habitats that are as im- years), and others (Snider and Bowler 1992), portant to the vertebrates as the ponds them- the percentage of their lives spent in longleaf selves (Semlitsch and Jensen 2001; Semlitsch pine savannas is likely to be similarly high. Ten and Bodie 2003). turtles that live in these same ponds as adults Observations of long-distance dispersal in a all require upland habitat for nesting, hiber- few species—2 km for the gopher frog (Moler nation, and as a refuge when the pond goes and Franz 1988) and 1.7 km for the flatwoods dry (Burke and Gibbons 1995). The ecologi- salamander (Ashton and Ashton 2005)—does cal quality of the upland habitat surrounding not necessarily mean that the appropriate up- small isolated water bodies, therefore, is as im- land habitat should be a zone of 2 km around portant for vertebrates that utilize ponds as is a breeding pond, but it does illustrate another the quality of the aquatic habitat to their lar- need of longleaf pine savanna amphibians and vae and aquatic stages (Burke and Gibbons of all animals that utilize such ponds: con- 1995; Marsh and Trenham 2001; Semlitsch nected corridors for dispersal among ponds and Bodie 2003; Means and Means 2005). Suc- (Fig. 8). Metapopulation dynamics are espe- cessful management of temporary pond ver- cially important for pond breeders because at tebrates requires restoring or maintaining the a crucial juncture in their life cycles, and peri- natural vegetative quality of their upland habi- odically, adults must make long journeys out- tats as well as the ecological integrity of their side of their normal home ranges to breed breeding ponds (Means and Means 2005). Un- in small habitats that are discontinuously dis- fortunately, for most species, few data exist tributed in the landscape. Moreover, the criti- on how much adjacent upland habitat is nec- cal requirement of these pond environments— essary and sufficient for species to maintain presence of water—is highly ephemeral, mak- minimum viable populations. ing amphibian and turtle populations much For only six salamanders for which he could more vulnerable to local extinction than if they find reliable data (including the mole and tiger lived their entire lives only in a contiguous lon- salamanders of longleaf pine savannas), Seml- gleaf pine savanna like mammals, other rep- itsch (1998) calculated that the terrestrial habi- tiles, and birds. The ability to disperse long tat that would likely encompass 95% of the distances over land, therefore, is a manda- population would extend 164.3 m (534 feet) tory requirement for pond-breeding amphib- from the wetland margins of ponds. John- ians and pond-inhabiting turtles, which allows son (2003) estimated that at least 16% of the them the possibility of finding other breeding 190 II. Ecology

FIGURE 8. Temporary pond in longleaf pine sandhills in Leon County, FL. Throughout the range of longleaf pine, small ponds devoid of fish are the breeding grounds or homes of about 35 species of salamanders, frogs, turtles, and snakes that depend upon these ponds almost exclusively for their existence. ponds or water to live in and for restocking water bodies, but the water bodies them- locally extinct populations. selves are in peril (Semlitsch and Bodie 1998; Unfortunately, small isolated wetlands that Semlitsch 2000). were once regulated by the U.S. Army Corps of Engineers, were removed in 2001 from fed- Value of Dead Trees (Snags and eral jurisdiction by the U.S. Supreme Court decision in the case Solid Waste Agency of North- Down Logs) ern Cook County v. U.S. Army Corps of Engineers. Dead standing trees (snags) and down logs This bodes very badly for these extremely im- are highly valuable to vertebrates in forest portant habitats that contribute much to the ecosystems and particularly so in the sparsely biodiversity of the southeastern United States. treed longleaf pine savannas (Baker 1974; In South Carolina, for instance, the elimi- Maser et al. 1979; Thomas et al. 1979; Miller nation or alteration of more than 90% of and Marion 1995). Numerous studies of ter- Coastal Plain Carolina bay wetlands (Bennett restrial vertebrates in longleaf pine savannas and Nelson 1991) has reduced essential habitat mention the use of snags by lizards (Goin for the black swamp snake (Seminatrix pygaea), and Goin 1951), frogs (Dodd 1996; Boughton Florida green water snake (Nerodia floridana), et al. 2000), snakes (Brode and Allison 1958; and chicken turtle (Deirochelys reticularia), all of Franz 1995; Tennant and Bartlett 2000 for the whose distribution patterns are restricted pri- scarlet kingsnake), birds (Miller and Marion marily to seasonal wetlands (Buhlmann 1995; 1995), and mammals. For instance, nursery Dorcas et al. 1998; Gibbons et al. 2000). Not and roosting colonies of up to 20 evening bats only is it imperative to manage the longleaf were found under loose bark of a lightning- pine savanna uplands surrounding these small struck pine (Baker 1974). Bark only stays in 6. Vertebrate Faunal Diversity 191 a loose condition for a short time, making it a yet many of the vertebrates that live in lon- very temporary habitat for the bats. If roost- gleaf pine savannas are fossorial, living all or ing sites are a limiting factor for the evening part of their time underground. This is espe- bat, loose pine bark from lightning-struck trees cially true of the salamanders, whose moist may play a very important role in the local skin would desiccate in the heat of midsum- presence of the species, even though it might mer sun. All of them live underground and be usable as a roost for one season or less very little is known about their adult lives. (Baker 1974). Many frogs hide in burrows or cavities dur- In an experimental study of the use of snags ing the day and emerge to forage at night. The by cavity-nesting birds, numbers and species eastern spadefoot, southern toad, and gopher of birds were consistently higher on longleaf frog all do this as well as do some treefrogs pine sites than in slash pine plantations, even such as the barking treefrog and ornate cho- though longleaf sites had fewer snags (Miller rus frog. Except for a few lizards that are ar- and Marion 1995). They demonstrated clearly boreal, all of the reptiles are either active bur- that, despite higher snag densities, pine plan- rowers (gopher tortoise, pine snake, hognose tations provided marginal habitat for cavity- snakes, red-tailed skink) or live in the burrows nesting birds compared with longleaf pine ar- of other animals and other underground cav- eas. They concluded that retention of longleaf ities. Many small mammals are consummate pine will support more species and numbers of burrowers (oldfield mouse, cotton rat, south- birds than pine plantations. eastern pocket gopher, eastern mole, nine- Unfortunately, the remaining longleaf pine banded armadillo, pine vole, and others) and stands are almost all second-growth forests make extensive tunnel systems in longleaf pine that germinated after the decimation of old- savannas that are utilized by fossorial animals growth longleaf pine from the 1880s through which are not themselves burrowers. All of the 1930s (Frost 1993). These forests, for the these animals periodically need to retreat un- most part, are uniformly aged and contain few derground from temperature extremes, preda- or no snags because they consist of relatively tors, and fire. Probably the most overlooked youthful trees (longleaf pine can live to more refuge for longleaf pine savanna vertebrates is than 500 years; Wahlenberg 1946). A very im- the subterranean base (butt, stump) of dead portant management consideration in longleaf longleaf pine trees and their associated rotting pine savannas is to make sure that lightning- roots (Means 2005b). killed trees and trees that die from other causes Longleaf pine is the only tree in the south- are left for wildlife use and not cut by salvage eastern United States that grows a shaftlike, operations (Miller and Marion 1995). massive taproot up to 5 m deep. Its lateral roots, growing only 25 cm below ground surface, can reach out to 22 m from the base Value of Stumpholes or Tree of the tree (Heyward 1933). As long as the tree is alive, its roots are tough and woody, Bases consisting, like the bole of the tree, of a cen- Subterranean cavities (burrows, rootholes, tral core of heartwood that is heavily impreg- stumpholes) may be important to more verte- nated with oleoresin and surrounded by sap- brates than any other physical characteristic of wood with much less oleoresin. When the tree longleaf pine savannas. Burrows of the gopher dies, the dense, oleoresin-impregnated heart- tortoise, for instance, are known to provide wood resists decomposition for many years, temporary or even permanent refuge to over but the sapwood surrounding it rots quickly, 300 species of other animals (Jackson and Mil- creating a soft, moist substrate into which an- strey 1989) and burrows of the southeastern imals can easily dig or force their way (Means pocket gopher to more than 60 other species 2005b) (Fig. 9). Swelling as they grow, the (Hubbell and Goff 1939). Not all vertebrates roots leave behind cavities filled with soft can create their own burrows, however, and and decaying wood after they die, but the 192 II. Ecology

FIGURE 9. The long-persisting, rotting taproots of longleaf pine are a vitally important microhabitat for terrestrial vertebrates, most of which are fossorial. hard, oleoresin-impregnated heartwood re- underground for days after the ground fire has mains to keep the cavities partially filled and passed. The heartwood is so dense that it rarely protected. During the frequent fires in longleaf burns completely, but it may ignite repeat- pine savannas, the resinous heartwood may edly in subsequent fires, keeping the under- burn and the punky sapwood may smoulder ground cavity open. Numerous dead tree bases 6. Vertebrate Faunal Diversity 193 were scattered throughout the presettlement step into stump extraction holes (Frost 1993). longleaf pine savannas, offering hideaways for The harvest of resinous stumps may have been fossorial animals around their rotting bases an important contributing factor in the decline and root passageways (Means 2005b). of many pineland vertebrates in the Southeast Kauffeld (1957, 1969) recorded the sub- (Moler 1992; Means 2005b). terranean use of tree bases by 11 species In addition to underground refugia cre- of longleaf pine savanna vertebrates on a ated by decaying tree bases and roots, bur- large game plantation in South Carolina rows are made by many animals besides the (eastern box turtle, corn snake, yellow rat gopher tortoise, southeastern pocket gopher, snake, common kingsnake, eastern hognose and, recently, armadillo. Some are excavated snake, copperhead, timber rattlesnake, east- by earthworms (Diplocardia spp.), scarab bee- ern diamondback rattlesnake, eastern cotton- tles (Geotrupes spp.), cicada larvae (Magicicada tail, cotton rat, Savannah sparrow). In the spp.), mole crickets (Gryllotalpidae), ants (e.g., pulpy wood of decaying pine stumps or in Pogonomyrmex spp., Dolichoderus spp., Solenop- tunnels formed by decaying pine roots in sis spp.), wolf spiders (Lycosidae), eastern Richmond County, GA, Neill (1948) found mole (Scalopus aquaticus), pine vole (Microtus the slimy salamander, southern toad, mud pinetorum), and oldfield mouse (Peromyscus snake, black racer, corn snake, black rat snake, polionotus), to name a few. No doubt all common kingsnake, and copperhead. During a the burrows provide refuge for many of the radiotelemetry study of the eastern diamond- small vertebrates that do not dig, such as the back rattlesnake and cottonmouth in northern flatwoods salamander, common and striped Florida, Means (1986, 2005b) compared the newts, oak toad, eastern narrowmouth toad, use by rattlesnakes of gopher tortoise burrows scarlet snake, pine woods snake, short-tailed versus stumpholes (tree bases) and found that snake, crowned snakes, coral snake, glass these snakes occupied stumpholes much more lizards, and mole skink. A few studies have frequently than gopher tortoise burrows. Inci- examined the ground surface disturbances of dentally, he observed nine other species using burrowing animals in longleaf savanna (Kalisz stumpholes including the common kingsnake, and Stone 1984, Hermann 1993, Simkin and cotton rat, coachwhip, common garter snake, Michener 2005), but no studies have exam- black racer, northern bobwhite quail, east- ined the subterranean extent of burrows nor ern box turtle, gray rat snake, and opossum. how important burrows are to the fossorial Richter et al. (2001) radiotracked 12 craw- fauna. Moreover, if subterranean burrows and fish frogs emigrating from a breeding pond other cavities are limiting to the population in southern Mississippi for 24 to 88 days and survival of many longleaf pine savanna fos- noted that most of the frogs used stumpholes sorial vertebrates, then disturbances to the or root mounds during the period and all were soil and subterranean cavities from mechani- in stumpholes at the end of the tracking study. cal activities such as plowing, harrowing, bed- One might think that resinous stumpholes ding, roller chopping, stump extraction, or would be abundant after the cutting of the even conversion to pasture may be signifi- virgin longleaf pine forests at the turn of the cantly harmful. Many of the recently reported twentieth century. Unfortunately for fossorial declines in fossorial amphibians (Means et al. vertebrates, most of the old stumpholes have 1996, Franz and Smith 1999) and reptiles slowly disappeared after the naval stores in- (Gibbons et al. 2000; Tuberville et al. 2000; dustry discovered, in the early 1900s, that the Krysko and Smith 2005) might be at least resinous stumps could be pulled or dozed from partly explained by range-wide depletion of the ground and cooked to extract turpentine, undisturbed longleaf pine savanna soil and rosin, and pine oil (Wahlenberg 1946; Dyer its microcavities. Studies of the importance of and Sicilia 1990). Today, one has to be careful subterranean cavities to longleaf pine savanna while walking in longleaf pine savannas not to vertebrates are urgently needed. 194 II. Ecology The Importance of Fire to killed, but prescribed burning is beneficial to most herpetofauna by perpetuating their habi- Wildlife in Longleaf Pine tat (Means and Campbell 1982). Savannas and Adjacent Prescribed burning in a Florida sandhill habitat increased diversity and abundance of Ecosystems amphibians and reptiles over control plots that had not been burned for 17 years, and Fire is responsible worldwide, if not for the some fire periodicities were better than oth- origin, then for the perpetuation of pine as- ers for maintaining high diversity (Mushinsky sociations known as savannas (Mirov 1967). 1985). The six-lined racerunner, which likes Pines may have evolved their special relation- open herbaceous vegetation and bare ground, ships with fire early in the evolutionary history was more abundant on sites burned annually of the group going back into the Jurassic, 150 (Mushinsky 1985), but the southeastern five- million years ago (Mirov 1967). Longleaf pine, lined skink, which uses thick litter for forag- as a species, probably does not go nearly that ing and shelter, was more abundant on 5- to far back in time, but it may have evolved from 7-year burn cycles (Mushinsky 1992). On the ancestors that were part of grassland/fire habi- other hand, fire periodicity had no effect on tats throughout the later half of the Cretaceous the peninsula crowned snake (Tantilla r.relicta), and all of the Cenozoic. Longleaf pine savannas a longleaf pine savanna specialist (Mushin- are pyrogenic communities that require peri- sky and Witz 1993). As expected, vegeta- odic fires in order to exist through time (Means tion structure is affected by periodicity of fire 1996; Platt 1999). In the absence of fire, lon- (Mushinsky 1986, Mushinsky and Gibson gleaf pine savannas succeed to southern mixed 1991), so arboreal and ground-dwelling lizards hardwood forest (Platt and Schwartz 1990; will be affected differently. Both are adapted Ware et al. 1993; Platt 1999), a completely to living in longleaf pine savannas, so dur- different ecosystem, but fires associated with ing frequent fires, the six-lined racerunner summer lightning probably have been charac- experiences a population increase while the teristic of the climate of the Coastal Plain of the southeastern five-lined skink declines, and southeastern United States at least since the vice versa when the fire periodicity is length- Miocene, 25 million years ago. Longleaf pine ened. So long as the habitat does not suc- savanna plants are so completely adapted to ceed into a hardwood forest, both species per- the periodic effects of fire, and fire has been so sist. Prescribed burning was recommended as a continuously present in the Coastal Plain, that much more beneficial management tool than longleaf pine savanna is considered a climax other techniques in maintaining optimal go- vegetative type (Platt 1999). pher tortoise habitat (Cox et al. 1987). Density Few studies of animal response to fire have of gopher tortoise populations is closely related been conducted, but Means and Campbell to biomass of herbaceous food plants, which (1982) reviewed the effects of prescribed burn- are promoted by frequent fires (Landers and ing on amphibians and reptiles and concluded Speake 1980). that for longleaf pine savanna vertebrates, few Changes in the vertebrate community and individuals are killed during fires, but some vegetation can be dramatic on annually species experience limited mortality because burned pinelands that are managed to re- they are casual visitors from hardwood com- semble longleaf pine savannas. Shortleaf and munities, such as the eastern box turtle (Ernst loblolly pines grow up on abandoned agricul- et al. 1995), or are aggregated under unusu- tural land in rich clayhills soils of the Coastal ally dense, flammable litter, such as the east- Plain—soils that previously supported longleaf ern glass lizard (Babbitt and Babbitt 1951). Un- pine savanna. If these oldfield successional der certain conditions of heavy fuel loads, strip vegetations are prescribe burned annually af- burning, or when a snake is in its shed cy- ter the pines reach about 10 years of age and cle and cannot see, a few individuals can be are not so vulnerable to fire damage, a pine 6. Vertebrate Faunal Diversity 195 forest with a herbaceous ground cover mixed brates. Fires in the pre-European Coastal Plain with fire-pruned hardwood sprouts (Ware et landscape probably were ignited by lightning al. 1993) can be maintained. Following 15 mostly in the uplands of longleaf pine savan- years of fire exclusion on an 8.6-ha plot of nas and swept downslope into adjacent veg- one such oldfield pine forest that had been an- etations (Means 1996). Many of these veg- nually burned for over 70 years in northern etations, such as hillside seepage bogs, wet Florida, only 11 of 43 bird species were en- flats, and Atlantic white cedar (Chamaecyparis countered every year of the study. Most finches thyoides) stands, are wetland communities in and brush-nesting species that were common which fires may not be easily ignited, but they at the beginning of the study no longer oc- nevertheless depend upon fires for their own curred on the study area, whereas several characteristic nature (Folkerts 1982; Cerulean species associated with mesic hardwood forest and Engstrom 1995; Peet this volume). This conditions increased in abundance (Engstrom is especially true of seepage bogs that can be et al. 1984). The hardwood response in such invaded by evergreen shrubs from downslope habitats is vigorous, because hardwoods estab- swamps in creek and stream bottoms (Folkerts lish underground root systems that continue 1982; Drewa et al. 2002a,b). to grow and store nutrients in spite of hav- Means and Longden (1976) noticed that ing their annual stem growth killed. The same breeding choruses of male pine barrens tree- hardwood species seem to have a more difficult frogs were more robust in landscapes in which time becoming established in native wiregrass herbaceous seepage bogs were free from ground cover, possibly from root competition woody shrubs that grew farther downslope in with the native herbs. wetter soils. Means and Moler (1979) hypoth- Spatial variation in fire temperatures ap- esized that the shallow puddles of larval breed- pears to have much less effect on vegetation ing sites of the pine barrens treefrog were dried patterns and dynamics of vegetation in fre- up by the increased evapotranspiration result- quently burned savannas than does season of ing from the invasion of woody vegetation burn (Platt et al. 1988). Open-canopied lon- into seepage bogs. These same shallow seepage gleaf pine savannas characterized by a ground rills and puddles have recently been discov- cover of grasses and forbs appear to be a result ered as the breeding habitat of a new species primarily of lighting-initiated fires that occur of the dwarf salamander complex (Means and repeatedly and frequently during the spring Jensen unpublished). In the long-term (5-plus (May–June) of the year (Platt et al. 1988). years) absence of fire, seepage bogs dry up Unfortunately, no experimental research has and become shaded by dense shrub growth. been done on the effects on vertebrates of Heliophilic vegetation gives way to peaty soils the season of burn. On the other hand, many covered with deep hardwood leaf litter and no studies have shown that vertebrates respond exposed seepage water, all of which are unsuit- positively to frequency of burn. For exam- able for the larvae of the pine barrens treefrog ple, for the endangered red-cockaded wood- and the new species of the dwarf salamander pecker, a longleaf pine specialist whose ecol- complex. ogy is best understood (Jackson 1994; Conner Two other examples of forest vegetations et al. 2001), experimental research has shown that respond to fires coming into them from that social groups living with high percent- longleaf pine savannas are Atlantic white cedar ages of wiregrass and low percentages of gall- forests (Ward and Clewell 1989; Frost 1995) berry in the ground cover of their territories, and eastern and southern redcedar (Junipe- and with larger areas of natural pine regener- rus virginiana, J. silicicola). These species are ation, have more adults and more neighbor- sensitive to fire, but they are also sensitive ing groups, and produce more young than do to shade from hardwoods. They most often other groups (James et al. 1997). occur in wetland ecotones or steep upper Longleaf pine savannas are vitally impor- slopes between longleaf pine savannas and tant to adjacent communities and their verte- southern temperate hardwood forests where 196 II. Ecology

fires periodically burn back the competing regionwide burning bans) it is not possible to hardwoods and provide a zone for even-aged burn in May or June, prescribed fire at other stands to grow (Monk 1968; Kucera 1981; seasons is better than no fire at all (Cox et al. Wright and Bailey 1982). 1987). Kirkman (1995) reviewed the impacts of fire and hydrological regimes on vegetation in depression marshes (temporary ponds). Fire during drought promotes species richness of Problems Associated with grasses and sedges. Winter fire followed by Pine Plantation Silviculture inundation can lead to shifts in the dominance of emergent vegetation, and long-term The largest known breeding migration of the fire suppression can result in the replacement flatwoods salamander (Ambystoma cingulatum) of herbaceous wetland vegetation with hard- was monitored over a 22-year period follow- wood species (Kirkman 1995). Hardwood litter ing its discovery in 1970 in Liberty County, FL. can significantly affect the occurrence of fires Nightly migrations of 200–300 adults across in wetlands and, from the point of view of ver- a 4.3-km stretch of paved highway in 1970– tebrates, the mix of amphibian species that uti- 1972 had dwindled to less than one individual lize those wetlands (Pechmann et al. 2001). per night in 1990–1992. Bedding and planting Recently, Schurbon and Fauth (2003) re- slash pine on the wet flat into which the flat- ported a decrease of frogs and salamanders woods salamanders were migrating to breed in plots burned on shorter versus longer ro- may have interfered with migration, success- tations. They argued that amphibians would ful hatching, larval life, and feeding and find- benefit from longer prescribed burn frequen- ing suitable cover postmetamorphosis (Means cies of 5 to 7 years. Means et al. (2004) pointed et al. 1996). out many weaknesses in the Schurbon and Gopher tortoises abandoned their burrows Fauth (2003) study and maintained that no in mature slash pine plantations at an average one has yet demonstrated unequivocally any rate of 22%/year (Aresco and Guyer 1999a) explicit amphibian response to fire; therefore, and grew more slowly on slash pine planta- managers should not refrain from burning lon- tions than they did in any other published gleaf pine savannas every 1 to 3 years, other- study and were estimated to require at least wise hardwood encroachment will eliminate 20 years to attain sexual maturity (Aresco and the habitat altogether in time. Guyer 1999b). Intensive soil disturbance as- Until we know more about the details of sociated with site preparation and conversion the effects of fire on longleaf pine savannas to pine plantations in the 1970s destroyed and their vertebrates, the best fire manage- much of the native ground cover. Slow growth, ment strategy probably should be to use pre- which resulted in delayed maturity, was at- scribed fires frequently (1 to 3 years) and in the tributed to poor forage quality of sparse ground season in which lightning-ignited fires burned cover vegetation, especially legume and non- naturally (May–June). Frequent burning is re- legume forbs (Aresco and Guyer 1999b). quired on most longleaf pine sites because of Breeding bird density, species richness, di- their recent long history (more than 80 years) versity (H ), and biomass were highest in the of fire suppression. Early growing season fires longleaf pine forest and differed (P < 0.05) are also important because they suppress the from those found in all age-classes (1, 10, 24, naturally woody components of longleaf pine 40 years) of slash pine plantations (Repen- savannas better than winter fires (Platt et al. ning and Labisky 1985). Although slash pine 1988), and they stimulate flowering of the plantations in northern Florida do not provide grasses and some forbs (Platt et al. 1988, 1991; habitat that will maintain the breeding bird Streng et al. 1993) that otherwise do not read- community of the natural longleaf pine forest, ily reproduce sexually. On the other hand, if older plantations (more than 40 years) do pro- for some reason (severe drought that forces vide habitat for a wintering bird community 6. Vertebrate Faunal Diversity 197 that is reasonably similar to that of the natural tion is sometimes quite subtle and often just longleaf pine forest (Repenning and Labisky as dangerous. It is not so obvious that pine 1985). plantations, for instance, create large holes White et al. (1975) compared habitat differ- in the population landscape of many longleaf ences 9 years after the planting of slash pine pine savanna vertebrates. The red-cockaded among three replicated levels of site prepara- woodpecker neither colonizes nor survives in tion ranging from clearcut, burn, and plant planted pine stands of any commercial age, (low intensity) to clearcut, burn, KG, single even though both habitats have an abundance harrow, bed, and plant (high intensity). They of pines (Costa 1995). Likewise, small pop- found that where grasses and forbs were least ulations occupying a large area with a few productive (high site preparation intensity), patches of unconnected longleaf pine savanna ground arthropods, small mammals, herbi- scattered about are likely to have inbreed- vores, birds, and insectivores were least abun- ing problems, restricted genetic diversity, and dant. They concluded that growth and devel- lack of gene flow between patches, depend- opment of slash pine overstories was favored ing upon the landscape matrix in which the by intensive site preparation at the expense of patches are embedded. Habitat fragmentation understory wildlife habitat. produces the same result as habitat reduction: It is well documented that the wildlife habi- breaking up into smaller areas. The species– tat values of pine plantations are inferior to area relationship, a well-documented principle those of natural pine stands (Harris et al. 1974; of ecology, says that each fragment, at equilib- Umber and Harris 1974; Kautz 1984; Repen- rium, will have fewer species than the original ning and Labisky 1985; McComb et al. 1986). (Simberloff 1993). Hence, the net result is that a large percentage The metapopulation dynamics of verte- of remaining pine forests in the Coastal Plain brate specialists in longleaf pine savannas dif- now provide poor-quality habitat for many fer for different groups. Those vertebrates, formerly abundant species of wildlife (Kautz mostly amphibians, whose life cycles depend 1993). Throughout the Coastal Plain, there is on breeding ponds must live in a landscape that no natural ecological analog of the modern not only provides a breeding pond for a local pine plantation (Means 2005a). Because the population and a sufficient upland habitat for main objective in pine plantations is to max- the adults, but also alternative ponds in which imize cellulose production, planted pines that to breed if one pond is destroyed, or when the can tolerate close stocking and canopy closure hydroperiod fails or is altered. More than for (slash, loblolly, sand) have largely replaced na- many of the reptiles, mammals, and birds that tive longleaf pine savannas. After 10–15 years live in the same longleaf pine savanna habi- of growth, plantations have very little or no tats during all life stages, the vertebrates that ground cover and are not savannas, but in- depend upon ponds need proximity of other stead, densely stocked forests where most of ponds from which to recruit new propagules if the photosynthesis takes place in the canopy the local population goes extinct. If ponds are and virtually none at ground level. One should too far apart or isolated by roads, agricultural not expect, therefore, that vertebrates adapted fields, older silviculture stands, clearcuts, or ur- to open-canopied savannas with a rich, herba- banization, populations cannot be restored. ceous ground cover would be suited for life in And yet, for the intact remnants of lon- deep shade and pine detritus. gleaf pine savannas, almost no data exist on the effects of roads on the smaller species of amphibians, reptiles, and mammals that may The Problem of Habitat be loathe to cross such a hot, coverless, and Fragmentation hostile habitat—even dirt roads. In a Maine study of road effects on amphibian move- Habitat reduction poses obvious problems for ments, forest roads served as a partial filter vertebrate populations, but habitat fragmenta- to the movements of some amphibian species, 198 II. Ecology notably salamanders, including a longleaf pine Declining Species savanna species, the eastern newt (DeMay- nadier and Hunter 2000). Fossorial vertebrates Between 1513, when Florida was first sighted such as the eastern mole, shrews, eastern by Ponce de Leon, and the mid-1900s, six pocket gopher, ground skink, Florida worm indigenous vertebrates disappeared from the lizard, mole skink, southern hognose snake, Coastal Plain at the hands of European man. mole kingsnake, and others might be inhib- The American bison (Bison bison bison) was ited from crawling on the surface of bare hard extirpated in the late 1700s or early 1800s ground, pavement, or exposing themselves to as a result of wanton slaughter by early set- birds of prey and other predators. tlers (Humphrey 1992b). The Florida red wolf Means (1999) found fewer than expected (Canis rufus floridanus) disappeared in the early mole salamanders immigrating into a pond 1900s (Robson 1992). The Carolina parakeet next to a heavily trafficked federal high- (Conuropsis carolinensis) was virtually extinct way through longleaf pine sandhills in north by 1900, exterminated by man as agricultural Florida and Smith et al. (2005) discussed the pests, for food, and for sport (Hardy 1978a). direct and indirect impacts highways have The passenger pigeon was driven to extinc- on amphibians and reptiles, generally, using tion by 1914, the victim of mass slaughter for the case study of U.S. Highway 441 through food and sport (Hardy 1978b). And the ivory- Payne’s Prairie in north central Florida as an billed woodpecker, last sighted in Georgia in example. In the only study of its kind, Rudolph 1955 (Means 2004b) and in Florida in 1969 et al. (1999) trapped large snakes at incremen- (Kautz 1993), met its demise due to the log- tal distances away from roads through longleaf ging of mature stands of lowland hardwoods pine savannas on the Angelina National Forest and, some think, to the removal of old-growth in east Texas. Their data strongly suggest that longleaf pines in the early part of the twen- populations of large snakes (e.g., pine snake tieth century (Tanner 1942). Not the least of and timber rattlesnake) are reduced by 50% the extinctions was indigenous man, who was or more to a distance of 450 m from roads of completely gone from Florida between 1700 moderate use (2400 vehicles per day). and 1800 (Tebeau 1971) and largely removed What were thought to have been good pop- from the southeastern United States by 1839 ulations of keystone species in longleaf pine (Perdue and Green 1995). savannas were found to be declining in spite of Among the 17 species of amphibians that are being on “large” reserves. McCoy and Mushin- specialists in longleaf pine savannas, only the sky (1992) found that a population of the flatwoods salamander is federally threatened, gopher tortoise had a 33% decline in active but the striped newt is under review (Linda burrows over a decade on the J.N. “Ding” LaClaire personal communication) and the go- Darling National Wildlife Refuge in Florida. pher and crawfish frogs are threatened (Godley Similarly, in the largest remaining population 1992) or imperiled on various state lists (Bailey of the red-cockaded woodpecker on Florida’s and Means 2004). All 17 species have declined Apalachicola National Forest, which is split when one takes into account that their primary into two roughly equivalent ranger districts, habitat, longleaf pine savannas, have been re- James (1991) discovered that the birds on the duced to less than 3% of their original extent eastern district were declining and that the (Ware et al. 1993). The best review of declin- river separating the two districts may be a bar- ing amphibians and threats to their popula- rier to gene flow in the species. Moreover, tions in the southeastern United States is by the fact that artificially made cavities are read- Dodd (1997), and additional information is in ily occupied on the western district, which Lannoo (2005). was deemed to have a healthy and stable red- Most of the longleaf pine savanna spe- cockaded woodpecker population of about 500 cialist snakes are in decline and in need of colonies, shows that cavity limitation was se- studies of their basic terrestrial biology. The vere even there (James et al. 1997). 6. Vertebrate Faunal Diversity 199 eastern Indigo snake is a federally threatened mechanical disturbance, except that S. invicta species (U.S. Fish and Wildlife Service 1978). was abundant in the littoral zones around the The short-tailed snake is listed as threatened margins of ponds, which are naturally dis- in Florida (Campbell and Moler 1992). The turbed by rising and falling pond water levels. Florida pine snake is a Species of Special Con- If the tendency for fire ants to stay out of undis- cern in Florida (Franz 1992), imperiled in Al- turbed native ground cover is widespread, then abama (Means 2004a), and should rank at least fire-ant predation on vertebrate eggs, if it were as a species of special concern in Georgia and in fact a problem, would not have much of an South Carolina. The southern hognose snake impact in remnant patches of the native lon- has declined severely throughout its range (Tu- gleaf pine savannas having undisturbed soils. berville et al. 2000), and the historic range This would give land managers all the more of the eastern diamondback rattlesnake has reason to protect native herbaceous vegetation shrunk, with populations at the extremes of from mechanical disturbance and to maintain its range extinct in Louisiana and threatened it with appropriate prescribed fires. in North Carolina (Martin and Means 2000), One of the longleaf pine specialist birds, and experiencing severe age-class truncation the red-cockaded woodpecker, has been feder- in Florida, Georgia, and Alabama (Means in ally threatened since 1970. In spite of receiv- preparation). Gibbons et al. (2000) likened the ing more research and management attention global decline of reptiles to that of amphibians, than any other longleaf pine savanna verte- citing several examples in the southeastern brate, it remains on the decline in part of its United States from small ponds surrounded by largest surviving enclave (James 1991), and longleaf pine savannas (black swamp snake, current management there for sawtimber is Florida green water snake, chicken turtle) to not sustainable (James et al. 2003). The north- residents of longleaf pine ecosystems (gopher ern bobwhite quail, whose 1920s population tortoise, indigo snake, eastern diamondback decline in the Southeast was responsible for rattlesnake, eastern box turtle, timber rat- stimulating the research that discovered the tlesnake, southern hognose snake). Except for importance of fire in longleaf pine savannas very limited data for a few species, very lit- (Stoddard 1931, 1962), has endured a de- tle is known about how much acreage of good cline of more than 65% over the last 20 years habitat is required to support minimal breed- throughout its range (Brennan 1991). Among ing populations of any of them. the factors of the decline is degradation and Fire ants have been proposed as a possible shrinking of habitat due to hardwood invasion threat to ground-nesting and terrestrial egg- of unburned pinelands, the very same cause laying vertebrates (Mount 1981). The fire ant of the decline in the 1920s. The white-breasted Solenopsis geminata was once reported from a and brown-headed nuthatches and Bachman’s wide variety of habitats but seems to have sparrow are not yet listed species, but their been partly displaced by another species of fire longleaf pine savanna habitat continues to de- ant, S. invicta, which was accidentally intro- crease. duced into Mobile, AL, from Brazil about 1940 All three species of mammals that are lon- (Wilson and Brown 1958; Buren 1972). In a gleaf pine specialists have declined. The Florida study of the use of flatwoods versus sandhill mouse, having the smallest range and being as- savannas on the Apalachicola National Forest, sociated with the gopher tortoise, whose pop- FL, Tschinkel (1987) found that both species ulations are dwindling, is considered a threat- preferred sites whose soil had been mechani- ened species in Florida, the state in which it cally disturbed, such as paved road shoulders, is endemic (Layne 1992). The principal race graded roadsides, and cleared and replanted of the fox squirrel in Florida is threatened flatwoods (S. invicta) or sandhills (S. gemi- (Kantola 1992), and the species has declined nata). The presence of both species in native and become extirpated widely over its entire ground cover of flatwoods and sandhills was range (Weigl et al. 1989). Goff’s gopher, a greatly reduced in comparison to sites having race of the southeastern pocket gopher that 200 II. Ecology was endemic on a sandy ridge in east-central Because rights of the private landowner are Florida, is thought to be extinct (Humphrey so strong, it is difficult to imagine how bio- 1992a; Brown 1997). All over the range of the diversity might be governmentally regulated southeastern pocket gopher, however, popula- on private lands, so the main hope for con- tions that were abundant and obvious in the servation of longleaf pine savanna biodiver- 1960s and 1970s have vanished (D. B. Means sity is on publicly owned lands. Unfortunately, personal observation). publicly owned lands are not distributed very The drastic loss of longleaf pine savannas has well in the Coastal Plain. Some states have had an even more severe impact on plants. very few. The federal lands offer the largest op- Hardin and White (1989) listed 191 rare plant portunity, but these, too, are inequitably dis- taxa that occur in ecosystems in the south- tributed throughout the Coastal Plain. Enscon- eastern United States (Alabama, Florida, Geor- cement of longleaf pine savannas on publicly gia, Mississippi, North Carolina, South Car- owned lands is not enough to ensure their sur- olina) where wiregrass (Aristida stricta)isan vival, however. Longleaf pine savannas need important component. Using the Nature Con- to be managed properly, but changing poli- servancy’s Natural Heritage Program method- tics with respect to resource utilization (log- ology, 122 taxa were considered endangered ging, recreation), air pollution (smoke from or threatened throughout their ranges. Seven prescribed burning), and encroaching develop- taxa have been proposed or listed as endan- ment (homes, municipalities, in-holdings) can gered by the U.S. Fish and Wildlife Service, present big challenges to management, espe- and 61 taxa are listed as endangered or threat- cially prescribed burning. ened by rare plant laws in three states. Hardin Single-species management for a target and White (1989) estimated that 66 rare wire- species such as the red-cockaded woodpecker grass associates are local endemics, one of the is probably not a good overall management higher numbers reported for a regional ecosys- strategy in any ecosystem. Considering how tem type in the United States. many longleaf pine savanna species are rare or threatened, the best management should be a multiple species program that focuses mainly on restoring and maintaining longleaf pine sa- Conclusions vannas to the most natural conditions possible. Prescribed fires on short rotations (1 to 3 The total number of resident vertebrates in years) in the early lightning season (May and longleaf pine savannas, 212 species including June) are most likely to produce an ecosystem 38 species that are specialists occurring exclu- mosaic suitable for all the native species. As re- sively or primarily in longleaf pine savannas, search produces new information about how is greater than for any other habitat type in the presettlement longleaf pine savannas really the Coastal Plain of the southeastern United were structured, it should be integrated into States and one of the largest vertebrate faunas the overall management program. A prime ex- in temperate North America. Such high species ample comes from recent red-cockaded wood- richness should be expected, given the antiq- pecker research on the Apalachicola National uity of this type of ecosystem and that longleaf Forest (ANF), the largest national forest in pine savannas once accounted for more than the eastern United States. James et al. (2001, 60% of a landscape stretching from southeast- 2003) found that present stocking of longleaf ern Virginia to east Texas. Unfortunately, the pines may be too dense and that not enough native longleaf pine savannas in which resi- longleaf pine reproduction is taking place. In dent and specialist vertebrates evolved and to other words, current management for sawtim- which, at least the specialists, are best adapted, ber on the ANF is not sustainable in this, the have shrunk to less than 3% of their former ex- largest population of this endangered species. pansive range, with the remnants highly frag- They recommended more vigorous burning mented and isolated. and monitoring of the condition of the ground 6. Vertebrate Faunal Diversity 201 cover. When it is less than 30% herbaceous, Anonymous. 1990. Summary report, Scientific they recommend a form of single tree selection summit on the Red-cockaded Woodpecker. Un- that emphasizes thinning from below in com- published report, Southeast Negotiation Net- bination with minigroup selection (harvest of work, Georgia Institute of Technology, Atlanta. patches of trees up to a radius equal to the Aresco, M.J., and Guyer, C. 1999a. Burrow aban- height of canopy trees). donment by gopher tortoises in slash pine plantations of the Conecuh National Forest. J Wildl If we are to keep any semblance of the fau- Manage 63(1):26–35. nal diversity that the original longleaf pine sa- Aresco, M.J., and Guyer, C. 1999b. Growth of vannas bequeathed us, we must recognize how the tortoise Gopherus polyphemus in slash pine valuable are the few remaining tracts and in- plantations of southcentral Alabama. Herpetolog- sure that they are properly studied and man- ica 55(4):499–506. aged to prevent further losses of the ecosys- Ashton, R.E., Jr. 1992. Flatwoods salamander, Am- tems and the vertebrate species they contain. bystoma cingulatum (Cope). In Rare and Endangered Whatever other values are gained from the Biota of Florida, Volume III, Amphibians and Reptiles, preservation of the small percentage of reed. P.E. Moler, pp. 39–43. Gainesville: University maining longleaf pine savannas and their fau- Press of Florida. nal diversity, these remnants are, at the very Ashton, R.E., Jr., and Ashton, P.S. 1981. Handbook of Reptiles and Amphibians of Florida. Part One: The least, exceedingly valuable repositories where Snakes. Miami, FL: Windward Publishing Co. most of the knowledge of the evolution and Ashton, R.E., Jr., and Ashton, P.S. 1985. Hand- adaptation of the constituent species is stored. book of Reptiles and Amphibians of Florida. Part Two: Learning why the red-cockaded woodpecker is Lizards, Turtles and Crocodilians. Miami, FL: Wind- a threatened species, for instance, would not ward Publishing Co. be possible in ruderal habitats in which it did Ashton, R.E., Jr., and Ashton, P.S.2005. Natural his- not evolve. To lose this storehouse of valuable tory and status of the flatwoods salamander, Am- information would be a loss to humanity far bystoma cingulatum (Cope) in Florida. In Status and greater than loss of its simple parts. Conservation of Florida Amphibians and Reptiles, eds. W. Meshaka and K. Babbitt, pp 62–73. Malabar, FL: Krieger Press. Auffenberg, W. 1969. Tortoise Behavior and Survival. Acknowledgments Chicago: Rand McNally. Auffenberg, W., and Franz, R. 1982. The status I owe many thanks to the late Edwin V. Ko- and distribution of the gopher tortoise (Gopherus marek, Sr. for empowering me to study lon- polyphemus). In North American Tortoises: Conserva- gleaf pine savanna vertebrates and fire ecology. tion and Ecology, ed. R.B. Bury, pp. 95–126. U.S. William J. Platt III provided intellectual stim- Fish and Wildlife Service Research Report 12. ulation and W. Wilson Baker abundant field Babbitt, L.H., and Babbitt, C.H. 1951. A herpetolog- knowledge and the love of fieldwork. Over the ical study of burned-over areas in Dade County, years some of my most helpful companions in Florida. Copeia 1951:79. the field were James B. Atkinson, Robert L. Bailey, M.A. 2001. The pocket gopher project. Crawford, Guy H. Means, and Ryan C. Means. The tortoise burrow. Newsl Gopher Tortoise Counc 21(1):7. I am especially grateful to my colleagues, Bailey, M.A., and Means, D.B. 2004. Gopher frog, W. Wilson Baker, Ken Dodd, Kevin Enge, Rana capito, and Mississippi gopher frog, Rana Harley Means, and Ryan Means, who allocated sevosa.InVertebrate and Selected Invertebrate Wildlife their valuable time to review this manuscript. of Alabama, ed. R.E. Mirarchi, pp. 15–17. Auburn, AL: Proceedings of the Second Nongame Wildlife References Conference. Baker, W.W. 1974. Longevity of lightning-struck Allen, E.R. 1968. The increase of rattlesnakes, Cro- trees and notes on wildlife use. Proc Tall Timbers talus adamanteus, in Florida. American Association Fire Ecol Conf 13:497–504. of Zoological Parks and Aquariums, 15 May 1968. Baker, W.W. 1981. The distribution, status, and fu- Pittsburgh, PA. ture of the red-cockaded woodpecker in Georgia. 202 II. Ecology

In Proceedings of the Nongame and Endangered Brown, R.B., Stone, E.L., and Carlisle, V.W. 1990. Wildlife Symposium, eds. R.R. Odum and J.W. Soils. In Ecosystems of Florida, eds. R. Myers and Guthrie, pp. 82–87. Georgia Department of Natu- J. Ewel, pp. 35–69. Orlando: University of Central ral Resources, Game and Fish Division, Technical Florida Press. Bulletin WL5. Bruce, D. 1947. Thirty-two years of annual burning Barlow, C. 2000. The Ghosts of Evolution: Nonsensical in longleaf pine. J For 45:809–814. Fruit, Missing Partners, and Other Ecological Anachro- Buhlmann, K.A. 1995. Habitat use, terrestrial nisms. New York: Basic Books. movements, and conservation of the turtle, Die- Barlow, R.F., and Jones, F.K., Jr. 1965. The white- rochelys reticularia in Virginia. J Herpetol 29:173– tailed deer in Florida. Florida Game and Fresh 181. Water Fish Commission Technical Bulletin 9:1– Buren, W.F. 1972. Revisionary studies on the tax- 239. onomy of the imported ﬁre ants. J Ga Entomol Soc Beane, J.C., Thorp, T.J., and Jackson, D.A. 1998. 7:1–27. Heterodon simus (southern hognose snake) diet. Burke, V.J., and Gibbons, J.W. 1995. Terrestrial Herpetol Rev 29:44–45. buffer zones and wetland conservation: A case Belson, M.S. 2000. Drymarchon corais couperi (east- study of freshwater turtles in a Carolina Bay. Con- ern indigo snake) and Micrurus fulvius fulvius serv Biol 9(6):1365–1369. (eastern coral snake). Predator-prey. Herpetol Rev Burton, T.M., and Likens, G.E. 1975. Salamander 31:105. populations and biomass in the Hubbard Brook Bennett, D.H., Gibbons, J.W., and Franson, J.C. Experimental Forest, New Hampshire. Copeia 1978. Terrestrial activity in aquatic turtles. Ecol- 1975:541–546. ogy 51:738–740. Caldwell, J.P. 1987. Demography and life history of Bennett, S.H., and Nelson, J.B. 1991. Distribution two species of chorus frogs (Anura: Hylidae) in and status of Carolina bays in South Carolina. South Carolina. Copeia 1987:114–127. Nongame and Heritage Trust Publication No. 1, Campbell, H.W., and Christman, S.P. 1982. The SC Wildlife and Marine Resources Department, herpetological components of Florida sandhill Columbia. and sand pine scrub associations. In Herpetologi- Boughton, R.G., Staiger, J., and Franz, R. 2000. cal Communities, ed. N.J. Scott, pp. 163–171. USDI Use of PVC pipe refugia as a sampling tech- Wildlife Research Report 13. nique for hylid treefrogs. Am Midl Nat 144:168– Campbell, H.W., and Moler, P.E. 1992. Threatened 177. short-tailed snake. In Rare and Endangered Biota of Brennan, L.A. 1991. How can we reverse the north- Florida, Volume III, Amphibians and Reptiles, ed. P. ern bobwhite population decline? wildl Soc Bull Moler, pp. 150–153. Gainesville: University Press 19:544–555. of Florida. Bridges, E.L., and Orzell, S.L. 1989. Longleaf pine Carr, A.F., Jr. 1940. A contribution to the herpetol- communities of the west Gulf Coastal Plain. Nat ogy of Florida. Univ Fla Publ Biol Sci Ser 3(1):1– Areas J 9(4):246–263. 118. Brimley, C.S. 1923. The dwarf salamander at Cerulean, S.I. 1991. The Preservation 2000 report. Raleigh, NC. Copeia 120:81–83. Florida’s natural areas—what have we got to lose? Brode, W.E., and Allison, P. 1958. Burrowing snakes The Nature Conservancy, Winter Park, FL. of the panhandle counties of Mississippi. Herpeto- Cerulean, S.I., and Engstrom, R.T. 1995. Fire in wet- logica 14(1):37–40. lands: A management perspective. Proc Tall Tim- Brodkorb, P. 1964. Catalogue of fossil birds: Part 2 bers Fire Ecol Conf 19. (Anseriformes through Galliformes). Bull Fla State Chapman, H.H. 1943. A 27-year record of annual Mus Biol Sci 8(3):195–335. burning versus protection of longleaf pine repro- Brooks, G.R., Jr. 1967. Population ecology of the duction. J For 41:71–72. ground skink, Lygosoma laterale (Say). Ecol Monogr Christensen, N.L. 1988. Vegetation of the southeast- 37:71–87. ern Coastal Plain. In North American Terrestrial Veg- Brown, E.E. 1979. Some snake food records from etation, eds. M. G. Barbour and W.D. Billings, pp. the Carolinas. Brimleyana 1:113–124. 317–363. London: Cambridge University Press. Brown, L.E., and Means, D.B. 1984. Fossorial be- Christman, S.P., and Means, D.B. 1992. Rare havior and ecology of the chorus frog Pseudacris striped newt. In Rare and Endangered Biota of ornata. Amphibia-Reptilia 5:261–273. Florida, Volume III, Amphibians and Reptiles, ed. Brown, L.N. 1997. Mammals of Florida. Miami, FL: P.E. Moler, pp. 62–65. Gainesville: University Windward Publishing Co. Press of Florida. 6. Vertebrate Faunal Diversity 203

Clark, R.F. 1949. Snakes of the hill parishes of Dodd C.K., Jr. 1995b. Reptiles and amphibians in Louisiana. J Tenn Acad Sci 24:244–261. the endangered longleaf pine ecosystem. In Our Clausen, C.J., Cohen, A.D., Emiliani, C., Holman, Living Resources: A Report to the Nation on Distribu- J.A., and Stipp, J.J. 1979. Little Salt Spring, tion, Abundance, and Health of U.S. Plants, Animals, Florida: A unique underwater site. Science 203: and Ecosystems, eds. E.T. LaRoe, G.S. Farris, C.E. 609–614. Puckett, P.D. Doran, and M.J. Mac, pp. 129–131. Conant, R., and Collins, J.T. 1998. A Field Guide Washington, DC: National Biological Service. to Reptiles and Amphibians: Eastern and Central Dodd, C.K., Jr. 1996. Use of terrestrial habitats North America, 3rd edition, expanded. Boston: by amphibians in the sandhill uplands of north- Houghton Mifﬂin. central Florida. Alytes 14:42–52. Conner, R.N., Rudolph, D.C., and Walters, J.R. Dodd, C.K., Jr. 1997. Imperiled amphibians: A his- 2001. The Red-cockaded Woodpecker: Surviving in torical perspective. In Aquatic Fauna in Peril: The a Fire-Maintained Ecosystem. Austin: University of Southeastern Perspective, Special Publication 1, eds. Texas Press. G.W. Benz and D.E. Collins, pp. 165–200. South- Costa, R. 1995. Biological opinion on U. S. Forest east Aquatic Research Institute. Decatur, GA: Service environmental impact statement for the Lenz Design and Communications. management of the red-cockaded woodpecker Dodd, C.K., Jr. 2001. North American Box Turtles: A and its habitat on national forests in the southern Natural History. Norman: University of Oklahoma region. In USDA Forest Service, Final Environmental Press. Impact Statement, Vol. II, pp. 1–192. Management Dodd, C.K., Jr., and Franz, R. 1995. Seasonal abun- Bulletin R8-MB–73. dance and habitat use of selected snakes trapped Cox, J., Inkley, D., and Kautz, R. 1987. Ecology and in xeric and mesic communities of north-central habitat protection needs of gopher tortoise (Go- Florida. Bull Fla Mus Nat Hist 38, Part I(2):43–67. pherus polyphemus) populations found on lands Dodd, C.K., Jr., Johnson, S.A., and Means, D.B. slated for large-scale development in Florida. 2005. Notophthalmus perstriatus.InAmphibian de- Florida Game and Fresh Water Fish Commis- clines, the conservation status of United States species, sion, Nongame Wildlife Program Technical Report ed. M. Lannoo. Berkeley: University of California No. 4. Press. Crenshaw, J.W., Jr. 1955. The life history of the Dorcas, M.E., Gibbons, J.W., and Dowling, H.G. southern spiny lizard, Sceloporus undulatus undu- 1998. Seminatrix Cope, black swamp snake. Cat- latus Latreille. Am Midl Nat 54(2):257–298. alogue of American Amphibians and Reptiles Dalrymple, G.H., Bernardino, F.S., Jr., Steiner, T.M., 679.1–679.5. and Nodell, R.J. 1991. Patterns of species di- Drewa, P.B., Platt, W.J., and Moser, E.B. 2002a. versity of snake community assemblages, with Community structure along elevation gradients data on two Everglades snake assemblages. Copeia in headwater regions of longleaf pine savannas. 1991:517–521. Plant Ecol 160:61–78. Deckert, R.F. 1918. A list of reptiles from Jack- Drewa, P.B., Platt, W.J., and Moser, E.B. 2002b. Fire sonville, Florida. Copeia 1918(54):30–33. effects on resprouting of shrubs in headwaters Delcourt, P.A., and Delcourt, H.R. 1987. Long-Term of southeastern longleaf pine savannas. Ecology Forest Dynamics of the Temperate Zone: A Case Study 83(3):755–767. of Late-Quaternary Forests in Eastern North America. Duellman, W.E., and Trueb, L. 1986. Biology of New York: Springer-Verlag. Amphibians. New York: McGraw–Hill Book Co. DeMaynadier, P.G.,and Hunter, M.L., Jr. 2000. Road Dundee, H.A., and Rossman, D.A. 1989. The effects on amphibian movements in a forested Amphibians and Reptiles of Louisiana. Baton Rouge: landscape. Nat Areas J 20(1):56–65. Louisiana State University Press. Diemer, J.E., and Speake, D.W. 1983. The distribu- Dunning, J.B., and Watts, B.D. 1990. Regional dif- tion of the eastern indigo snake, Drymarchon corais ferences in habitat occupancy by Bachman’s spar- couperi, in Georgia. J Herpetol 17:256–264. rows. Auk 107:463–472. Dodd, C.K., Jr. 1992. Biological diversity of a tem- Dyer, D., and Sicilia, D.B. 1990. Labors of a Mod- porary pond herpetofauna in north Florida sand- ern Hercules: The Evolution of a Chemical Company. hills. Biodivers Conserv 1:125–142. Boston: Harvard Business School Press. Dodd, C.K., Jr. 1995a. The ecology of a sandhills Echternacht, A.C., and Harris, L.D. 1993. The fauna population of the eastern narrow-mouthed toad, and wildlife of the southeastern United States. In Gastrophryne carolinensis, during a drought. Bull Biodiversity of the Southeastern United States: Low- Fla Mus Nat Hist Biol Sci 38, Pt. I(1):11–41. land Terrestrial Communities, eds. W.H. Martin, S.G. 204 II. Ecology

Boyce, and A.C. Echternacht, pp. 81–116. New Franz, R. 1986. Florida pine snakes and gopher frogs York: John Wiley & Sons. as commensals of gopher tortoise burrows. In Ehrlich, P.R., Dobkin, D.S., and Wheye, D. 1988. The Gopher Tortoise and Its Community, eds. D.R. The Birder’s Handbook: A Field Guide to the Natural Jackson and R.J. Bryant, pp. 16–20. Proceed- History of North American Birds. New York: Simon ings 5th Annual Meeting of the Gopher Tortoise & Schuster. Council. Eisenberg, J. 1983. The gopher tortoise as a key- Franz, R. 1991. Remember the drought? Fla Wildl stone species. Proceedings of the Annual Meeting of 45(6):10–12. the Gopher Tortoise Council 4:1–4. Franz, R. 1992. Florida pine snake. In Rare and En- Emslie, S.D. 1998. Avian community, climate, and dangered Biota of Florida, Volume III, Amphibians and sea-level changes in the Plio-Pleistocene of the Reptiles, ed. P.E. Moler, pp. 254–258. Gainesville: Florida Peninsula. Ornithol Monogra 50:1–113. University Press of Florida. Enge, K.M. 1997. A standardized protocol for drift- Franz, R. 1995. Habitat use, movements, and home fence surveys. Florida Game and Fresh Water Fish range in two species of ratsnakes (genus Elaphe) Commission Technical Report No. 14. in a north Florida sand hill. Florida Game and Enge, K.M. 2002. Herpetofaunal drift-fence survey Fresh Water Fish Commission, Nongame Wildlife of two seepage bogs in Okaloosa County, Florida. Program Project Report, Tallahassee, FL. Fla Sci 65:189–203. Franz, R. 2005. Up close and personal: A glimpse Enge, K.M., and Sullivan, J.O. 2000. Seasonal ac- into the life of pine snakes in a north Florida sand- tivity of the scarlet snake, Cemophora coccinea,in hill. In Status and Conservation of Florida Amphibians Florida. Herpetol Rev 31:82–84. and Reptiles, eds. W. Meshaka and K. Babbitt, pp. Enge, K.M., and Wood, K.N. 2002. A pedestrian 120–131. Melbourne, FL: Krieger Press. road survey of an upland snake community in Franz, R., and Auffenberg, W. 1978. The gopher Florida. Southeast Nat 1:365–380. tortoise: A declining species. In Proceedings of the Enge, K.M., and Wood, K.N. 2003. A pedestrian Rare and Endangered Wildlife Symposium, eds. R. road survey of the southern hognose snake Odom and L. Landers, Technical Bulletin WL 4, (Heterodon simus) in Hernando County, Florida. pp. 61–63. Georgia Department of Natural Re- Fla Sci 66:189–203. sources, Game and Fish Division, Social Circle. Engstrom, R.T. 1993. Characteristic mammals and Franz, R., and Smith, L.L. 1999. Distribution and birds of longleaf pine forests. Proc Tall Timbers Fire status of the striped newt and Florida gopher frog Ecol Conf 18:127–138. in peninsular Florida. Final Report, Florida Fish Engstrom, R.T., Crawford, R.L., and Baker, W.W. and Wildlife Conservation Commission, Tallahas- 1984. Breeding bird populations in relation to see. changing forest structure following ﬁre exclusion: Frost, C.C. 1993. Four centuries of changing land- A 15-year study. Wilson Bull 96:437–450. scape patterns in the longleaf pine ecosystem. Proc Ernst, C.H., and Ernst, E.M. 2003. Snakes of the United Tall Timbers Fire Ecol Conf 18:17–43. States and Canada. Washington, DC: Smithsonian Frost, C.C. 1995. Presettlement ﬁre regimes in Books. Southeastern marshes, peatlands, and swamps. Ernst, C.H., Boucher, T.P., Sekscienski, S.W., and Proc Tall Timbers Fire Ecol Conf 19:39–60. Wilgenbusch, J.C. 1995. Fire ecology of the Garner, J.A., and Landers, J.L. 1981. Food and habi- Florida box turtle, Terrapene carolina bauri. Herpetol tat of the gopher tortoise in southwestern Geor- Rev 26(4):185–187. gia. Proc Annu Conf Southeast Assoc Fish Wildl Agenc Flannery, T. 2001. The Eternal Frontier, an Ecological 35:120–134. History of North America and Its Peoples. Melbourne, Gates, C.A., and Tanner, G.W. 1988. Effects of pre- Australia: Text Publishing. scribed burning on herbaceous vegetation and Folkerts, G.W. 1982. The Gulf coast pitcher plant pocket gophers (Geomys pinetus) in a sandhill com- bogs. Am Sci 70:260–267. munity. Fla Sci 51(3/4):129–139. Folkerts, G.W. 1997. Citronelle ponds: Little-known Gibbons, J.W., Scott, D.E., Ryan, T.J., Buhlmann, wetlands of the Central Gulf Coastal Plain, USA. K.A., Tuberville, T.D., Metts, B.S., Greene, J.L., Nat Areas J 17(1):6–16. Mills, T., Leiden, Y., Poppy, S., and Wynne, C.T. Ford, N.B., Cobb, V., and Stout, J. 1991. Species 2000. The global decline of reptiles, dej´ a` vu am- diversity and seasonal abundance of snakes in phibians. Bioscience 50(8):653–666. a mixed pine–hardwood forest in eastern Texas. Godley, J.S. 1992. Threatened gopher frog. In Southwest Nat 36:171–177. Rare and Endangered Biota of Florida, Volume III, 6. Vertebrate Faunal Diversity 205

Amphibians and Reptiles, ed. P.E. Moler, pp. 15–19. of Ambystoma cingulatum and Ambystoma mabeei. Gainesville: University Press of Florida. Herpetologica 30:156–160. Goin, C.J. 1947. A note on the food of Heterodon Hardy, J.W. 1978a. Recently extinct Carolina para- simus. Copeia 1947:275. keet. In Rare and Endangered Biota of Florida, ed. Goin, C.J. 1950. A study of the salamander Am- H.W. Kale, p. 120. Gainesville: University Press of bystoma cingulatum, with the description of a new Florida. subspecies. Ann Carnegie Mus 31:299–321. Hardy, J.W.1978b. Extinct passenger pigeon. In Rare Goin, O.B., and Goin, C.J. 1951. Notes on the natu- and Endangered Biota of Florida. Vol. 2. Birds, ed. ral history of the lizard Eumeces laticeps, in Florida. H.W. Kale, pp. 120–121. Gainesville: University Q J Fla Acad Sci 14:29–33. Press of Florida. Goodyear, A.C. 1999. The early Holocene occupa- Harris, L.D., White, L.D., Johnston, J.E., and tion of the southeastern United States: A geoar- Milchunas, D.G. 1974. Impact of forest planta- chaeological summary. In Ice Age Peoples of North tions on north Florida wildlife and habitat. Proc America: Environments, Origins, and Adaptations, Annu Conf Southeast Assoc Game Fish Comm 28: 659– eds. R. Bonnichsen and K.L. Turnmire, pp. 432– 667. 481. Corvallis: Oregon State University Press. Harrison, J.R. 1973. Observations on the life history Graham, A. 1965. Origin and evolution of the biota and ecology of Eurycea quadridigitata (Holbrook). of southeastern North America: Evidence from HISS News J 1:57–58. the fossil plant record. Evolution 18:571–585. Harrison, J.R., and Guttman, S.I. 2003. A new Grayson, D.K. 2001. The archaeological record of species of Eurycea (Caudata: Plethodontidae) from human impacts on animal populations. J World North and South Carolina. Southeast Nat 2(2):159– Prehist 15:1–68. 178. Greenberg, C.H. 2001. Spatio-temporal dynamics of Hermann, S.M. 1993. Small-scale disturbances in pond use and recruitment in Florida gopher frogs longleaf pine forest. In The Longleaf Pine Ecosys- (Rana capito aesopus). J Herpetol 35:74–85. tem: Ecology, Restoration, and Management, ed. S.M. Greenberg, C.H., Neary, D.G., and Harris, L.D. 1994. Hermann, pp. 265–274. Proceedings of the Tall Effect of high-intensity wildﬁre and silvicultural Timbers Fire Ecology Conference, Tallahassee, FL. treatments on reptile communities in sand-pine Hess, C.A., and James, F.C. 1998. Diet of the red- scrub. Conserv Biol 8:1047–1057. cockaded woodpecker in the Apalachicola Na- Guyer, C., and Bailey, M.A. 1993. Amphibians and tional Forest. J Wildl Manage 62:509–517. reptiles of longleaf pine communities. Proc Tall Heyward, F. 1933. The root system of longleaf pine Timbers Fire Ecol Conf 18:139–158. on the deep sands of western Florida. Ecology Hallinan, T. 1923. Observations made in Duval 14(2):136–148. County, northern Florida, on the gopher tortoise Holman, J.A. 1991. North American Pleistocene (Gopherus polyphemus). Copeia 1923:11–20. herpetological stability and its impact on the inter- Hansen, K.L. 1953. The burrow of the gopher tor- pretation of modern herpetofaunas: An overview. toise. Q J Fla Acad Sci 26:353–360. Ill State Mus Sci Pap 23:227–235. Hansen, K.L. 1958. Breeding pattern of the eastern Holman, J.A. 1995. Pleistocene Amphibians and Rep- spadefoot toad. Herpetologica 14:57–67. tiles in North America. New York: Oxford University Harcombe, P.A., Glitzenstein, J.S., Knox, R.G., Press. Orzell, S.L., and Bridges, E.L. 1993. Vegetation Holman, J.A. 1996. The large Pleistocene (Sanga- of the longleaf pine region of the west Gulf monian) herpetofauna of the Williston IIIA site, Coastal Plain. Proc Tall Timbers Fire Ecol Conf 18:83– north-central Florida. Herpetol Nat Hist 4(1):35– 104. 47. Hardin, E.D., and White, D.L. 1989. Rare vascu- Hooper, R.G., Lennartz, M.R., and Muse, H.D. lar plant taxa associated with wiregrass (Aristida 1991. Heart rot and cavity tree selection by red- stricta) in the southeastern United States. Nat Areas cockaded woodpeckers. J Wildl Manage 55:323– J 9(4):234–245. 327. Hardy, J.D., Jr. 1969. Reproductive activity, growth, Hubbell, T.H., and Goff, C.C. 1939. Florida pocket and movements of Ambystoma mabeei Bishop in gopher burrows and their arthropod inhabitants. North Carolina. Bull Md Herpetol Soc 5(2):65–76. Proc Fla Acad Sci 4:127–177. Hardy, J.D., and Olmon, J. 1974. Restriction of Humphrey, S.R. 1992a. Recently extinct Goff’s the range of the frosted salamander, Ambystoma pocket gopher, Geomys pinetis gofﬁ.InRare and cingulatum, based on a comparison of the larvae Endangered Biota of Florida, Vol. 1, Mammals, ed. 206 II. Ecology

S.R. Humphrey, pp. 11–18. Gainesville: Univer- James, F.C., Hess, C.A., Kicklighter, B.C., and Thum, sity Press of Florida. R.A. 2001. Ecosystem management and the niche Humphrey, S.R. 1992b. Recently extirpated Plains gestalt of the red-cockaded woodpecker in lon- bison, Bison bison bison.InRare and Endan- gleaf pine forests. Ecol Appl 11:854–870. gered Biota of Florida, Vol. 1, Mammals, ed. S.R. James, F.C., Richards, P.M., Hess, C.A., McCluney, Humphrey, pp. 47–53. Gainesville: University K.E., Walters, E.L., and Schrader, M.S. 2003. Press of Florida. Sustainable forestry for the red-cockaded wood- Jackson, D.L., and Franz, R. 1981. Ecology of the pecker’s ecosystem. In Red-Cockaded Woodpecker: eastern coral snake (Micrurus fulvius) in northern Road to Recovery, eds. R. Costa and S.J. Daniels. peninsular Florida. Herpetologica 37:213–228. Blain, WA: Hancock House Publishers. Jackson, D.L., and Milstrey, E.G. 1989. The fauna of Janzen, D.H. 1984. Dispersal of small seeds by big gopher tortoise burrows. In Proceedings of the Go- herbivores: Foliage is the fruit. Am Nat 123:338– pher Tortoise Relocation Symposium, eds. J.E. Diemer, 353. D.R. Jackson, J.L. Landers, J.N. Layne, and D.A. Janzen, D.H., and Martin, P.S. 1982. Neotropical Woods, pp. 86–98. Florida Game and Fresh Wa- anachronisms: The fruits the gomphotheres ate. ter Fish Commission Nongame Wildlife Program Science 215:19–27. Technical Report 5. Tallahassee, FL. Jensen, J.B. 1996. Heterodon simus (southern hog- Jackson, J.A. 1970. Predation of a black rat snake nose snake). Hatchling size. Herpetol Rev 27:25. on yellow-shafted flicker nestlings. Wilson Bull Johnson, S.A. 2002. Life history of the striped newt 82:329–330. at a north-central Florida breeding pond. Southeast Jackson, J.A. 1974. Gray rat snakes versus red- Nat 1(4):381–402. cockaded woodpeckers: Predator–prey adapta- Johnson, S.A. 2003. Orientation and migration dis- tions. Auk 9(2):342–347. tances of a pond-breeding salamander (Notoph- Jackson, J.A. 1977. Notes on the behavior of the thalmus perstriatus, Salamandridae). Alytes 21:28– gray rat snake (Elaphe obsoleta spiloides). J Miss Acad 47. Sci 22:94–96. Johnson, T.G. 1988. Forest statistics for southeast Jackson, J.A. 1978. Predation by a gray rat snake on Georgia, 1988. USDA Forest Service Resource red-cockaded woodpecker nestlings. Bird-Banding Bulletin SE-104. 49:187–188. Jones, C.A. 1995. Habitat use and home ranges of Jackson, J.A. 1988. The southeastern pine forest Podomys floridanus on the Ordway Preserve, Put- ecosystem and its birds: Past, present, and future. nam County, Florida. Bull Fla Mus Nat Hist 38, Pt. In Bird Conservation, Vol. 3, ed. J.A. Jackson, pp. II(7):195–209. 119–159. Madison: University of Wisconsin. Kaczor, S.A., and Hartnett, D.C. 1990. Gopher tor- Jackson, J.A. 1994. Red-cockaded woodpecker toise (Gopherus polyphemus) effects on soils and (Picoides borealis). In The Birds of North America, vegetation in a Florida sandhill community. Am No. 85, eds. A. Poole and F. Gill. Philadelphia: The Midl Nat 123:100–111. Academy of Natural Sciences, Washington, DC: Kalisz, P.J., and Stone, E.L. 1984. Soil mixing The American Ornithologist’s Union. by scarab beetles and pocket gophers in north- Jackson, J.A., and Dakin, O.H. 1982. An encounter central Florida. Soil Sci Soc Am J 48:169–172. between a nesting barn owl and a gray rat snake. Kantola, A.T. 1992. Threatened Sherman’s fox Raptor Res 16:60–61. squirrel. In Rare and Endangered Biota of Florida. Jackson, J.F. 1973. Distribution and population phe- Volume 1. Mammals, ed. S.R. Humphrey, pp. 234– netics of the Florida scrub lizard, Sceloporus woodi. 241. Gainesville: University Press of Florida. Copeia 1973:746–761. Kauffeld, C. 1957. Snakes and Snake Hunting. Garden James, F.C. 1971. Ordinations of habitat relation- City, NY: Hanover House. ships among breeding birds. Wilson Bull 83:215– Kauffeld, C. 1969. Snakes, the Keeper and the Kept. 236. Garden City, NY: Doubleday & Co. James, F.C. 1991. Signs of trouble in the largest re- Kautz, R. 1984. Criteria for evaluaZing impacts of maining population of red-cockaded woodpeck- development on wildlife habitats. Proc Annu Conf ers. Auk 108:419–423. Southeast Assoc Fish Wildl agenc 38:121–136. James, F.C., Hess, C.A., and Kufrin, D. 1997. Kautz, R. 1993. Trends in Florida wildlife habitat Species-centered environmental analysis: 1936–1987. Fla Scie 56(1):7–24. Indirect effects of fire history on red-cockaded Keegan, H.L. 1944. Indigo snakes feeding upon poi- woodpeckers. Ecol Appl 7(1):118–129. sonous snakes. Copeia 1944:59. 6. Vertebrate Faunal Diversity 207

Kellogg, F.E., Doster, G.L., and Komarek, E.V. 1972. MacNeill, F.S. 1950. Pleistocene shorelines in The one quail per acre myth. Natl Quail Symp 1:15– Florida and Georgia. US Geol Surv Prof Pap 221- 20. F, pp. 95–107. Kiester, A.R. 1971. Species density of North Ameri- Marsh, D.M., and Trenham, P.C. 2001. Metapopula- can amphibians and reptiles. Syst Zool 20:127–137. tion dynamics and amphibian conservation. Con- Kirkman, L.K. 1995. Impacts of fire and hydrologi- serv Biol 15:40–49. cal regimes on vegetation in depression wetlands Martin, P.S. 1967. Prehistoric overkill. In Pleistocene of southeastern USA. Proc Tall Timbers Fire Ecol Conf Extinctions, eds. P.S. Martin and H.E. Wright, Jr., 19:10–20. pp. 75–120. New Haven, CT: Yale University Krysko, K.L. 2001. Ecology, conservation, and Press. morphological and molecular systematics of the Martin, W.H., and Boyce, S.G. 1993. Introduction: kingsnake, Lampropeltis getula (Serpentes: Col- The Southeastern setting. In Biodiversity of the ubridae). Dissertation, University of Florida, Southeastern United States, eds. W.H. Martin, S.G. Gainesville. Boyce, and A.C. Echternacht, pp. 1–46. New York: Krysko, K.L., and Smith, D.J. 2005. The decline and John Wiley & Sons. extirpation of the kingsnake (Lampropeltis getula) Martin, W.H., and Means, D.B. 2000. Geographic in Florida. In Status and Conservation of Florida Am- distribution and habitat relationships of the east- phibians and Reptiles, eds. W. Meshaka and K. Bab- ern diamondback rattlesnake, Crotalus adaman- bitt, pp. 132–141. Gainesville: University Press of teus. Herpetol Nat Hist 7(1):9–35. Florida. Martof, B.S. 1968. Ambystoma cingulatum (Cope). Kucera, C.L. 1981. Grasslands and fire. In Fire Catalogue of American Amphibians and Reptiles Regimes and Ecosystem Properties: Proceedings of 57.1–57.2. the Conference; 1978 December 11–15, eds. H.A. Martof, B.S., Palmer, W.M., Bailey, J.R., Harrison, Mooney, T.M. Bonnicksen, and N.L. Christensen, J.R., III, and Dermid, J. 1980. Amphibians and Rep- pp. 90–111. Technical Report WO-26. Washing- tiles of the Carolinas and Virginia. Chapel Hill: Uni- ton, DC: USDA Forest Service. versity of North Carolina Press. Kurten, B., and Anderson, E. 1980. Pleistocene Mam- Maser, C., Anderson, R.G., Cromack, K., Jr., mals of North America. New York: Columbia Uni- Williams, J.T., and Martin, R.E. 1979. Dead and versity Press. down woody material. In Wildlife Habitats in Man- Landers, J.L., and Speake, D.W. 1980. Manage- aged Forests, the Blue Mountains of Oregon and Wash- ment needs of sandhill reptiles in southern Geor- ington, ed. J.W. Thomas, pp. 78–79. USDA Forest gia. Proceedings of the Annual Conference. Annu Service, Agriculture Handbook 553, Washington, Conf Southeast Assoc Fish Wildl Agenc 34:515– DC. 529. May, P.G., Farrell, T.M., Heulett, S.T., Pilgrim, M.A., Lannoo, M. 2005. Amphibian Declines: The Conserva- Bishop, L.A., Spence, D.J., Rabatsky, A.M., Camp- tion Status of United States Species. Berkeley: Univer- bell, M.G., Aycrigg, A.D., and Richardson, W.E., sity Press of California. II 1996. Seasonal abundance and activity of a Layne, J.N. 1974. The land mammals of south rattlesnake (Sistrurus miliarius barbouri) in central Florida. In Environments of South Florida: Present Florida. Copeia 1996:389–401. and Past, Memoir 2, ed. P.J. Gleason, pp. 386–413. May, P.G., Heulett, S.T., Farrell, T.M., and Pilgrim, Miami, FL: Miami Geological Society. M.A. 1997. Live fast, love hard, and die young: Layne, J.N. 1992. Threatened Florida mouse. In Rare The ecology of pigmy rattlesnakes. Reptile and Am- and Endangered Biota of Florida, Vol. 1, Mammals, ed. phibian Magazine Jan./Feb. 1997:36–49. S.R. Humphrey, pp. 250–264. Gainesville: Uni- McComb, W.C., Bonney, S.A., Sheffield, R.M., and versity Press of Florida. Cost, N.D. 1986. Den tree characteristics and Layne, J.N., and Jackson, R.J. 1994. Burrow use by abundance in Florida and South Carolina. J Wildl the Florida mouse (Podomys floridanus) in south- Manage 50(4):584–591. central Florida. Am Midl Nat 131:17–23. McCoy, E.D., and Mushinsky, H.R. 1992. Studying a Lowery G.H., Jr. 1974. The Mammals of Louisiana and species in decline: Changes in populations of the Its Adjacent Waters. Baton Rouge: Louisiana State gopher tortoise on federal lands in Florida. Fla Sci University Press. 55:116–124. MacArthur, R.H. 1958. Population ecology of some McPherson, G. 1997. Ecology and Management of warblers of northeastern coniferous forest. Ecology North American Savannas. Tuscon: University of 39:599–619. Arizona Press. 208 II. Ecology

Means, D.B. 1986. Life history and ecology of Means, D.B., and Krysko, K.L. 2001. Biogeog- the eastern diamondback rattlesnake (Crotalus raphy and pattern variation in the kingsnake, adamanteus). Final Report of Project No. GFC-84- Lampropeltis getula, across the Apalachicola Re- 013. Florida Game and Fresh Water Fish Commis- gion of Florida. Contemporary Herpetology sion, Tallahassee. 2001(5). http://www.calacademy.org/research/ Means, D.B. 1996. Longleaf pine forest, going, go- herpetology/ch/ch/2001/5/index.htm ing. . . . In Eastern Old-Growth Forest: Prospects for Means, D.B., and Longden, C.J. 1976. Aspects of Rediscovery and Recovery, ed. M.B. Davis, pp. 210– the biology and zoogeography of the pine barrens 229. Washington, DC: Island Press. treefrog (Hyla andersonii) in northern Florida. Her- Means, D.B. 1999. The effects of highway mortality petologica 32(2):117–130. on four species of amphibians at a small, tempo- Means, D.B., and Means, R.C. 1998. Red Hills sur- rary pond in northern Florida. In Proceedings of vey for breeding pond habitat of the flatwoods the Third International Conference on Wildlife Ecol- salamander (Ambystoma cingulatum), gopher frog ogy and Transportation, September 13–16, 1999, eds. (Rana capito), and striped newt (Notophthalmus G. Evink, P. Garrett, and D. Zeigler, pp. 125–128. perstriatus) in the Tallahassee Red Hills of Leon, Missoula, MT. Gadsden, and Jefferson counties, Florida, and the Means, D.B. 2000. Southeastern U. S. Coastal Plain Tifton Uplands of Thomas and Grady counties, habitats of the Plethodontidae: The importance Georgia. Final report to the U.S. Fish and Wildlife of relief, ravines, and seepage. In The Biology Service, Jackson, MS, for Order No. 43910-5- of Plethodontid Salamanders, eds. R.C. Bruce, R.J. 0091. Jaeger, and L.D. Houck, pp. 287–302. New York: Means, D.B., and Means, R.C. 2005. Effects of sand Plenum press. pine silviculture on pond-breeding amphibians Means, D.B. 2001. Reducing impacts on rare ver- in the Woodville Karst Plain of north Florida. In tebrates that require small isolated water bod- Amphibians and Reptiles: Status and Conservation in ies along U.S. Highway 319. Final Report to the Florida, eds. W. Meshaka and K. Babbitt, pp. 56– Florida Department of Transportation. 61. Malabar, FL: Krieger Press. Means, D.B. 2004a. Florida pine snake, Pituophis Means, D.B., and Moler, P.E. 1979. The pine bar- melanoleucus mugitus (Barbour). In Alabama rens treefrog: Fire, seepage bogs, and manage- Wildlife, Volume 3, Imperiled Amphibians, Reptiles, ment implications. In Proceedings of the Rare and Birds, and Mammals, eds. R.E. Mirarchi, M.A. Endangered Wildlife Symposium, August 3–4, Athens, Bailey, T.M. Haggerty, and T.L. Best, pp. 69–70. GA, eds. R.R. Odum and L. Landers, pp. 77–83. Tuscaloosa: University of Alabama Press. Georgia Department of Natural Resources, Game Means, D.B. 2004b. Ghosts of the Red Hills. In Be- and Fish Division Technical Bulletin WL-4. tween Two Rivers: Stories from the Red Hills to the Gulf, Means, D.B., Palis, J.G., and Baggett, M. 1996. eds. S. Cerulean, J. Ray, and L. Newton, pp. 89– Effects of slash pine silviculture on a Florida 99. Tallahassee, FL: Heart of the Earth and Red population of flatwoods salamander. Conserv Biol Hills Writer’s Project. 10(2):426–437. Means, D.B. 2005a. Pine silviculture. In Amphib- Means, D.B., Dodd, C.K., Jr., Johnson, S.A., and ian Declines: The Conservation Status of United States Palis, J.G. 2004. Amphibians and fire in longleaf Species, ed. M. Lanoo, pp. 139–145. Berkeley: Uni- pine ecosystems. Conserv Biol 18(3):1149–1153. versity Press of California. Meylan, P.A. 1982. The squamate reptiles of the In- Means, D.B. 2005b. The value of dead tree bases glis 1A fauna (Irvingtonian: Citrus Co., Florida). and stumpholes as habitat for wildlife. In Amphib- Bull Fla State Mus Biol Sci 27(3):1–85. ians and Reptiles: Status and Conservation in Florida, Meylan, P.A. 1995. Pleistocene amphibians and eds. W. Meshaka and K. Babbitt, pp. 74–78. Mel- reptiles from the Leisey Shell Pit, Hillsborough bourne, FL: Krieger Press. County, Florida. Bull Fla State Mus Nat Hist 37, Pt. Means, D.B. In preparation. Herpetophilia. (Submit- I(9):273–297. ted). Miller, S.H., and Marion, W.R. 1995. Natural and Means, D.B., and Campbell, H.W. 1982. Effects of created snags and cavity-nesting birds in north prescribed burning on amphibians and reptiles. In Florida pine forests. Report for Project GFC-84- Prescribed Fire and Wildlife in Southern Forests. Pro- 020, Florida Game and Fresh Water Fish Com- ceedings of a Symposium, ed. G.W. Wood, pp. 89–97. mission, Tallahassee. Belle W. Baruch Forest Science Institute, Clemson Mirov, N.T. 1967. The genus Pinus. New York: University, Clemson, SC. Ronald Press Co. 6. Vertebrate Faunal Diversity 209

Mitchell, J.C., and Hedges, S.B. 1980. Ambystoma inexpectatus, on a periodically burned sandhill in mabeei Bishop (Caudata: Ambystomatidae): An Florida. Herpetologica 48:307–312. addition to the fauna of Virginia. Brimleyana Mushinsky, H.R., and Gibson, D.J. 1991. The influ- 3:119–121. ence of fire periodicity on habitat structure. In Moler, P.E. 1985a. Distribution of the eastern in- Habitat Complexity: The Physical Arrangement of Ob- digo snake, Drymarchon corais couperi, in Florida. jects in Space, eds. S.S. Bell, E.D. McCoy, and H.R. Herpetol Rev 16(2):37–38. Mushinsky, pp. 237–259. New York: Chapman & Moler, P.E. 1985b. Home range and seasonal activ- Hall. ity of the eastern indigo snake, Drymarchon corais Mushinsky, H.R., and Witz, B.W. 1993. Notes on couperi, in northern Florida. Final Performance the peninsula crowned snake, Tantilla relicta,in Report, Study No. E-1-06, III-A-5, Florida Game periodically burned habitat. J Herpetol 27:468– and Fresh Water Fish Commission, Tallahassee. 470. Moler, P.E. 1992. Eastern indigo snake Drymar- Neill, W.T. 1948. Hibernation of amphibians and chon corais couperi.InRare and Endangered Biota of reptiles in Richmond County, Georgia. Herpetolog- Florida, Volume III, Amphibians and Reptiles, ed. P.E. ica 4:107–114. Moler, pp. 181–186. Gainesville: University Press Neill, W.T. 1951a. The eyes of the worm lizard, of Florida. and notes on the habits of the species. Copeia Moler, P.E., and Franz, R. 1988. Wildlife val- 1951:177–178. ues of small, isolated wetlands in the South- Neill, W.T. 1951b. Notes of the role of crawfishes eastern coastal plain. In Proceedings of the Third in the ecology of reptiles, amphibians, and fishes. Nongame and Endangered Wildlife Symposium, eds. Ecology 32:764–766. R.R. Odum, K.A. Riddleberger, and J.C. Ozier, Neill, W.T. 1952. Burrowing habits of Hyla gratiosa. pp. 234–241. Georgia Department of Natural Re- Copeia 1952:196. sources, Social Circle. Neill, W.T. 1957. Some misconceptions regarding Monk, C.D. 1968. Successional and environmen- the eastern coral snake, Micrurus fulvius. Herpeto- tal relationships of the forest vegetation of north logica 13:111–118. central Florida. Am Midl Nat 79(2):441–457. Neill, W.T. 1958. The varied calls of the barking Mosimann, J.E., and Rabb, G.B. 1948. The salaman- treefrog, Hyla gratiosa LeConte. Copeia 1958:44– der Ambystoma mabeei in South Carolina. Copeia 46. 1948:304. Nelson, D.H., and Gibbons, J.W. 1972. Ecology, Mount, R.H. 1963. The natural history of the red- abundance, and seasonal activity of the scarlet tailed skink, Eumeces egregius Baird. Am Midl Nat snake, Cemophora coccinea. Copeia 1972:582–584. 70:356–385. Nowak, R.M. 1999. Walker’s Mammals of the World, Mount, R.H. 1975. Reptiles and Amphibians of Al- 6th ed. Baltimore: Johns Hopkins University abama. Alabama Agricultural Experiment Station, Press. Auburn University, Auburn. Olson, M.S., and Platt, W.J. 1995. Effects of habitats Mount, R.H. 1981. The red imported fire ant, and growing season fires on resprouting of shrubs Solenopsis invicta (Hymenoptera: Formicidae), as in longleaf pine savannas. Vegetatio 119:101–118. a possible serious predator on some native south- Owen-Smith, N. 1987. Pleistocene extinctions: eastern vertebrates: Direct observations and sub- The pivotal role of megaherbivores. Paleobiology jective impressions. J Ala Acad Sci 52(2):71–78. 13:351–362. Mushinsky, H.R. 1984. Observations of the feeding Owen-Smith, R.N. 1988. Megaherbivores: The Influ- habits of the short-tailed snake, Stilosoma extenua- ence of Very Large Body Size on Ecology. London: tum, in captivity. Herpetol Rev 15:67–68. Cambridge University Press. Mushinsky, H.R. 1985. Fire and the Florida sandhill Palis, J.G. 1996. Element stewardship abstract: Flat- herpetofaunal community: with special attention woods salamander (Ambystoma cingulatum Cope). to responses of Cnemidophorus sexlineatus. Herpeto- Nat Areas J 16:49–54. logica 41:333–342. Palis, J.G. 1998. Breeding biology of the gopher frog, Mushinsky, H.R. 1986. Fire, vegetation structure Rana capito, in western Florida. J Herpetol 32:217– and herpetological communities. In Studies in 223. Herpetology, ed. Z. Rocek, pp. 383–387. Prague, Palis, J.G., and Means, D.B. 2005. Ambystoma cingu- Czechoslovakia: Charles University. latum.InAmphibian declines, the conservation status Mushinsky, H.R. 1992. Natural history and abun- of United States species, ed. M. Lannoo. Berkeley: dance of southeastern five-lined skinks, Eumeces University of California Press. 210 II. Ecology

Palmer, W.M. 1987. A new species of glass lizard Prouty, W.F. 1952. Carolina bays and their origin. (Anguidae: Ophisaurus) from the southeastern Bull Geol Soc Am 63:167–224. United States. Herpetologica 43(4):415–423. Punzo, F. 1995. An analysis of feeding in the Palmer, W.M., and Braswell, A.L. 1995. Reptiles of oak toad, Bufo quercicus (Holbrook), (Anura, North Carolina. Chapel Hill: University of North Bufonidae). Fla Sci 58(1):16–20. Carolina Press. Pyne, S.J. 1982. Fire in America, a Cultural History of Palmer, W.M., and Tregembo, G. 1970. Notes on Wildland and Rural Fire. Princeton, NJ: Princeton the natural history of the scarlet snake Cemophora University Press. coccinea copei Jan in North Carolina. Herpetologica Repenning, R.W., and Labisky, R.F. 1985. Effects of 26:300–302. even-age timber management on bird communi- Pearson, P.G.1955. Population ecology of the spade- ties of the longleaf pine forest in northern Florida. foot toad, Scaphiopus h. holbrooki (Harlan). Ecol J Wildl Manage 49(4):1088–1098. Monogr 25:233–267. Reynolds, J.H. 1980. A mark-recapture study of the Pechmann, J.H.K., Estes, R.A., Scott, D.E., and Gib- scarlet snake, Cemophora coccinea, in a coastal plain bons, J.W. 2001. Amphibian colonization and use sandhill community. Unpublished master’s thesis, of ponds created for trial mitigation of wetland North Carolina State University, Raleigh. loss. Wetlands 21(1):93–111. Richter, S.C. 1998. The demography and reproduc- Peet, R.K., and Allard, D.J. 1993. Longleaf pine tive biology of gopher frogs, Rana capito, in Missis- vegetation of the southern Atlantic and eastern sippi. Unpublished master’s thesis, Southeastern Gulf Coast regions. Proc Tall Timbers Fire Ecol Conf Louisiana University, Hammond. 18:45–81. Richter, S.C., Young, J.E., Seigel, R.A., and John- Perdue, T., and Green, M.D., eds. 1995. The Chero- son, G.N. 2001. Post-breeding movements of the kee Removal: A Brief History with Documents. Boston: dark gopher frog, Rana sevosa Goin and Netting: Bedford Books of St. Martin’s Press. Implications for conservation and management. Petranka, J.W. 1998. Salamanders of the United States J Herpetol 35:316–321. and Canada. Washington, DC: Smithsonian Insti- Robertson, W.B., Jr., and Kushlan, J.A. 1974. The tute Press. southern Florida avifauna. In Environments of Pianka, E.R. 1973. The structure of lizard commu- South Florida: Present and Past, ed. P.J. Gleason, nities. Annu Rev Ecol Syst 4:53–74. pp. 414–452. Miami Geological Society Memoir Pianka, E.R. 1975. Niche relations of desert lizards. 2, Miami, FL. In Ecology and Evolution of Communities, eds. M. Robson, M.S. 1992. Recently extinct Florida red Cody and J. Diamond, pp. 292–314. Cambridge: wolf. In Rare and Endangered Biota of Florida, Vol- Harvard University Press. ume I, Mammals, ed. S.R. Humphrey, pp. 29–34. Platt, W.J. 1975. The colonization and formation Gainesville: University Press of Florida. of equilibrium plant species associations on bad- Rogers, L.L. 1999. Black bear. In The Smithsonian ger disturbances in a tall-grass prairie. Ecol Monogr Book of North American Mammals, eds. D. Wil- 45:285–305. son and S. Ruff, pp. 157–160. Washington, DC: Platt, W.J. 1999. Southeastern pine savannas. In Sa- Smithsonian Institute. vannas, Barrens, and Rock Outcrop Plant Communities Rostlund, E. 1960. The geographic range of his- of North America, eds. R.C. Anderson, J.S. Fralish, toric bison in the Southeast. Ann Assoc Am Geogr and J.M. Baskin, pp. 23–51. London: Cambridge 50:395–407. University Press. Rudolph, D.C., Burgdorf, S.J., Conner, R.N., and Platt, W.J., and Schwartz, M.W. 1990. Temperate Schaefer, R.R. 1999. Preliminary evaluation of the hardwood forests. In Ecosystems of Florida, eds. R. impact of roads and associated vehicular traffic on Myers and J. Ewel, pp. 194–229. Orlando: Uni- snake populations in east Texas. In Proceedings of versity of Central Florida Press. the Third International Conference on Wildlife Ecol- Platt, W.J., Evans, G.W., and Davis, M.M. 1988. ogy and Transportation, September 13–16, 1999, eds. Effects of fire season on flowering of forbs and G. Evink, P. Garrett, and D. Zeigler, pp. 129–136. shrubs in longleaf pine forests. Oecologia 76:353– Missoula, MT. 363. Savage, H. 1982. The Mysterious Carolina Bays. Platt, W.J., Glitzenstein, J.S., and Streng, D.R. 1991. Columbia: University of South Carolina Press. Evaluating pyrogenicity and its effects on vegeta- Scholes, R.J., and Archer, S.R. 1997. Tree–grass in- tion in longleaf pine savannas. Proc Tall Timbers teractions in savannas. Annu Rev Ecol Syst 28:517– Fire Ecol Conf 17:143–161. 544. 6. Vertebrate Faunal Diversity 211

Schurbon, J.M., and Fauth, J.E. 2003. Effects of Smallwood, pp. 217–248. New York: Academic prescribed burning on amphibian diversity in a Press. southeastern U. S. national forest. Conserv Biol Sharitz, R.R., and Gibbons, J.W. 1982. The ecology 17:1338–1349. of southeastern shrub bogs (pocosins) and Car- Semlitsch, R.D. 1981. Terrestrial activity and sum- olina Bays: A community proﬁle. U. S. Fish and mer home range of the mole salamander (Am- Wildlife Service, FWS/OBS-82/04. Washington, bystoma talpoideum). Can J Zool 59:315–322. DC. Semlitsch, R.D. 1983a. Structure and dynamics of Shoop, C.R. 1960. The breeding habits of the mole two breeding populations of the eastern tiger sala- salamander, Ambystoma talpoideum (Holbrook), in mander, Ambystoma tigrinum. Copeia 1983:608– southeastern Louisiana. Tulane Stud Zool 8:65–82. 618. Simberloff, D. 1993. Species-area and fragmenta- Semlitsch, R.D. 1983b. Burrowing ability and be- tion effects on old-growth forests: Prospects for havior of salamanders of the genus Ambystoma. longleaf pine communities. Proc Tall Timbers Fire Can J Zool 61:616–620. Ecol Conf 18:1–13. Semlitsch, R.D. 1998. Biological delineation of ter- Simpkin, S.M., and Michener, W.K. 2005. Faunal restrial buffer zones for pond-breeding salaman- soil disturbance regime of a longleaf pine ecosys- ders. Conserv Biol 12(5):1113–1119. tem. Southeast Nat 4(1):133–152. Semlitsch, R.D. 2000. Size does matter: The value of Smith, H.M. 1946. Handbook of Lizards. Ithaca, NY: small isolated wetlands. National Wetlands Newslet- Cornell University Press. ter January–February:5–6, 13. Smith, L.L., Smith, K.G., Barichievich, W.J., Dodd, Semlitsch, R.D., and Bodie, J.R. 1998. Are C.K., Jr., and Sorensen, K. 2005. Roads and small, isolated wetlands expendable? Conserv Biol Florida’s herpetofauna: A review and mitigation 12(5):1129–1133. case study. In Status and conservation of Florida am- Semlitsch, R.D., and Bodie, J.R. 2003. Biological cri- phibians and reptiles, eds. W. Meshaka and K. Bab- teria for buffer zones around wetlands and ripar- bitt. Malabar, FL: Krieger Press. ian habitats for amphibians and reptiles. Conserv Snider, A.T., and Bowler, J.K. 1992. Longevity of Biol 176(5):1219–1228. reptiles and amphibians in North American col- Semlitsch, R.D., and Jensen, J.J. 2001. Core habi- lections, 2nd ed. Society for the Study of Amphib- tat, not buffer zone. National Wetlands Newsletter ians and Reptiles Herpetological Circular No. 21. 23(4):5–6, 11. Snyder, J.R., Herndon, A., and Robertson, W.B. Semlitsch, R.D., and McMillan, M.A. 1980. Breed- 1990. South Florida rockland ecosystems: Trop- ing migrations, population size structure, and ical hammocks and pinelands. In Ecosystems of reproduction of the dwarf salamander, Eurycea Florida, eds. R. Myers and J. Ewel, pp. 230–274. quadridigitata, in South Carolina. Brimleyana Orlando: University of Central Florida Press. 3:97–105. Steadman, D.W., and Martin, P.S. 1984. Extinction Semlitsch, R.D., and Moran, G.B. 1984. Ecology of of birds in the late Pleistocene of North Amer- the redbelly snake (Storeria occipitomaculata) us- ica. In Quaternary Extinctions, a Prehistoric Revolu- ing mesic habitats in South Carolina. Am Midl Nat tion, eds. P.S. Martin and R.G. Klein, pp. 466–477. 111:33–40. Tuscon: University of Arizona Press. Semlitsch, R.D., Brown, K.L., and Caldwell, J.P. Stevenson, D.J., Dyer, K.J., and Willis-Stevenson, 1981. Habitat utilization, seasonal activity, and B.A. 2003. Survey and monitoring of the eastern population size structure of the southeastern indigo snake in Georgia. Southeast Nat 2(3):393– crowned snake Tantilla coronata. Herpetologica 408. 37:40–46. Stevenson, H.M., and Anderson, B.H. 1994. The Semlitsch, R.D., Gibbons, J.W., and Tuberville, T.D. Birdlife of Florida. Gainesville: University Press of 1995. Timing of reproduction and metamor- Florida. phosis in the Carolina gopher frog (Rana capito Stickel, W.H., and Cope, J.B. 1947. The home ranges capito) in South Carolina. J Herpetol 29(4):612– and wanderings of snakes. Copeia 1947:127–136. 614. Stoddard, H.L. 1931. The Bobwhite Quail: Its Habits, Semlitsch, R.D., Scott, D.E., Pechmann, J.H.K., and Preservation and Increase. New York: Charles Scrib- Gibbons, J.W. 1996. Structure and dynamics of an ner’s Sons. amphibian community, evidence from a 16-year Stoddard, H.L. 1962. Use of ﬁre in pine forests study of a natural pond. In Long-Term Studies of and game lands of the deep Southeast. Proc Tall Vertebrate Communities, eds. M.L. Cody and J.A. Timbers Fire Ecol 1:31–42. 212 II. Ecology

Stoddard, H.L. 1978. Birds of Grady County, Geor- sandhills. Proc Annu Conf Southeast Assoc Game Fish gia. Bull Tall Timbers Res Stn 21:1–175. Comm 28:686–693. Stout, I.J., and Marion, W.R. 1993. Pine flatwoods U.S. Fish and Wildlife Service. 1978. Part 17— and xeric pine forests of the southern (lower) Endangered and threatened wildlife and plants; Coastal Plain. In Biodiversity of the Southeastern listing of the eastern indigo snake as a threatened United States: Lowland Terrestrial Communities, eds. species. Fed Regist 43(21):4026–4028. W.H. Martin, S.G. Boyce, and A.C. Echternacht, U.S. Fish and Wildlife Service. 2003. Recovery plan pp. 373–446. New York: John Wiley & Sons. for the red-cockaded woodpecker, Picoides bore- Streng, D.R., Glitzenstein, J.S., and Platt, W.J. 1993. alis; 2nd revision. US Fish and Wildlife Service, Evaluating the season of burn in longleaf pine Atlanta, GA. forests: A critical literature review and some re- Van Balen, J.B., and Doerr, P.D. 1978. The relation- sults from an ongoing long-term study. Proc Tall ship of understory vegetation to red-cockaded Timbers Fire Ecol Conf 18:2278–263. woodpecker activity. Proc Annu Conf Southeast Assoc Tanner, J.S. 1942. The ivory-billed woodpecker. Re- Fish Wildl Agenc 32:82–92. search Report No. 1, National Audubon Society. Varner, J.M., and Kush, J.S. 2001. Old-growth Tebeau, C.W. 1971. A History of Florida. Coral Gables, longleaf pine forests—filling in the blanks. FL: University of Miami Press. Proceedings Third Longleaf Alliance Regional Tennant, A. 1997. A Field Guide to the Snakes of Florida. Conference, Longleaf Alliance Report 5:204– Houston, TX: Gulf Publishing Co. 208. Tennant, A., and Bartlett, A.D. 2000. Snakes of North Vogl, R.J. 1973. Fire in the Southeastern grasslands. America Eastern and Central Regions. Houston, TX: Proc Tall Timbers Fire Ecol Conf 12:175–198. Gulf Publishing Co. Wahlenberg, W.G. 1946. Longleaf Pine: Its Use, Ecol- Test, F.C. 1893. The “gopher frog.” Science 22:75. ogy, Regeneration, Protection, Growth, and Manage- Thomas, J.W., Anderson, R.G., Maser, C., and Bull, ment. Washington, DC: Charles Lathrop Pack E.L. 1979. Snags. In Wildlife Habitats in Managed Forestry Foundation. Forests, the Blue Mountains of Oregon and Washington, Walker, J. 1993. Rare vascular plant taxa associated ed. J.W. Thomas, pp. 60–77. USDA Forest Service, with the longleaf pine ecosystems: Patterns in tax- Agriculture Handbook 553, Washington, DC. onomy and ecology. Proc Tall Timbers Fire Ecol Conf Timmerman, W.W. 1995. Home range, habitat use, 18:105–125. and behavior of the eastern diamondback rat- Walker, J., and Peet, R.K. 1983. Composition tlesnake (Crotalus adamanteus) on the Ordway Pre- and species diversity of pine–wiregrass savannas serve. Bull Fla Mus Nat Hist 328, Pt. I(5):127– of the Green Swamp, North Carolina. Vegetatio 158. 55:163–179. Travis, J. 1992. Status undetermined eastern tiger Walsh, P.B, Darrow, D.H., and Dyass, J.G. 1995. salamander Ambystoma tigrinum tigrinum (Green). Habitat selection by Florida grasshopper sparrows In Rare and Endangered Biota of Florida, Volume III, in response to fire. Proc Annu Conf Southeast Assoc Amphibians and Reptiles, ed. P.E. Moler, pp. 70–76. Fish Wildl Agenc 49:342–349. Gainesville: University Press of Florida. Ward, D.B., and Clewell, A.F. 1989. Atlantic white Tschinkel, W.R. 1987. Distribution of the fire ants cedar (Chamaecyparis thyoides) in the southern Solenopsis invicta and S. geminata (Hymenoptera: states. Fla Sci 51(1):8–47. Formicidae) in northern Florida in relation to Ware, S., Frost, C., and Doerr, P.D. 1993. South- habitat and disturbance. Ann Entomol Soc Am ern mixed hardwood forest: The former longleaf 81:76–81. pine forest. In Biodiversity of the Southeastern United Tuberville, T.D., Bodie, J.R., Jensen, J.B., LaClaire, States, eds. W.H. Martin, S.G. Boyce, and A.C. L.V., and Gibbons, J.W. 2000. Apparent decline of Echternacht, pp. 447–493. New York: John Wi- the southern hognose snake (Heterodon simus). J ley & Sons. Elisha Mitchell Sci Soc 116(1):19–40. Watts, W.A. 1969. A pollen diagram from Mud Lake, Tucker, J.W., Jr., and Robinson, W.D. 2003. Influ- Marion County, north-central Florida. Geol Soc Am ence of season and frequency of fire on Henslow’s Bull 80:631–642. sparrows (Ammodramus henslowii) wintering on Watts, W.A. 1971. Postglacial and interglacial veg- Gulf Coast pitcher plant bogs. Auk 120(1):96– etation history of southern Georgia and central 106. Florida. Ecology 52(4):670–690. Umber, R.W., and Harris, L.D. 1974. Effects of inten- Watts, W.A. 1975. A late Quaternary record of veg- sive forestry on succession and wildlife in Florida etation from Lake Annie, south-central Florida. 6. Vertebrate Faunal Diversity 213

Geology 3:344–346. White, L.D., Harris, L.D., Johnston, J.E., and Watts, W.A. 1980. The late Quaternary vegetation Milchunas, D.G. 1975. Impact of site preparation history of the southeastern United States. Annu on flatwoods wildlife habitat. Proc Annu Conf Rev Ecol Syst 11:387–409. Southeast Assoc Game Fish Comm 29:347–353. Watts, W.A., Hansen, B.C.S., and Grimm, E.C. 1992. Wilson, E.O., and Brown, W.L. 1958. Recent Camel Lake: A 40,000-yr record of vegetational changes in the introduced population of the fire and forest history from northwest Florida. Ecology ant, Solenopsis saevissima (Fr. Smith). Evolution 73(3):1056–1066. 12:211–218. Webb, S.D. 1974. Chronology of Florida Pleistocene Wilson, V. 1951. Some notes on a captive scarlet mammals. In Pleistocene Mammals of Florida, ed. snake. Herpetologica 7:172. S.D. Webb, pp. 5–31. Gainesville: University Press Wolfe, S.H., Reidenauer, J.A., and Means, D.B. of Florida. 1988. An ecological characterization of the Webb, S.D. 1977. A history of savanna vertebrates Florida panhandle. US Fish Wildl Serv Biol Rep in the New World. Part I: North America. Annu 88(12):1–277. Rev Ecol Syst 8:355–380. Wolfenden, G.E. 1959. A Pleistocene avifauna from Webb, S.D. 1990. Historical biogeography. In Ecosys- Rock Springs, Florida. Wilson Bull 71(2):183– tems of Florida, eds. R. Myers and J. Ewel, pp. 70– 187. 100. Orlando: University of Central Florida Press. Wood, D.A. 2001. Florida’s Fragile Wildlife: Conserva- Webb, S.D., and Wilkins, K.T. 1984. Historical bio- tion and Management. Gainesville: University Press geography of Florida Pleistocene mammals. In of Florida. Contributions in Quaternary Vertebrate Paleontology, Wright, A.H. 1932. Life-histories of the Frogs of Okefi- Special Publication No. 8, eds. H.H. Genoways nokee Swamp, Georgia. New York: Macmillan Pub- and M.R. Dawson. Pittsburgh, PA. lishing Co. Webb, S.D., Dunbar, J., and Newsom, L. 1992. Wright, A.H., and Wright, A.A. 1949. Handbook of Mastodon digesta from north Florida. Curr Res Frogs and Toads of the United States and Canada. Pleistocene 1992:114–116. Ithaca, NY: Comstock Publishing Associates. Weigl, P.D., Steele, M.A., Sherman, L.J., Ha, J.C., Wright, A.H., and Wright, A.A. 1957. Handbook of and Sharpe, T.L. 1989. The ecology of the fox Snakes of the United States and Canada, 2 volumes. squirrel (Sciurus niger) in North Carolina: Implica- Ithaca, NY: Comstock Publishing Associates. tions for survival in the Southeast. Bull Tall Tim- Wright, H.A., and Bailey, W. 1982. Fire Ecology: bers Res Stn 24:1–93. United States and Southern Canada. New York: John Werner, P.A., ed. 1991. Savanna Ecology and Man- Wiley & Sons. agement: Australian Perspectives and Intercontinen- Young, T.S., and Laerm, J. 1993. A late Pleistocene tal Comparisons. Oxford, UK: Blackwell Scientific vertebrate assemblage from the St. Marks River, Publications. Wakulla County, Florida. Brimleyana 18:15–57. Section III Silviculture

215 Chapter 7 Uneven-Aged Silviculture of Longleaf Pine

James M. Guldin

Introduction creased the ecological value of the stands that remain. The scattered tracts of remnant un- The use of uneven-aged silviculture has in- managed longleaf pine stands have high emo- creased markedly in the past 20 years. This is tional and physical appeal. Managed stands of especially true in the southern United States, longleaf pine, especially those in which pre- where the use of clearcutting and planting is scribed fire has been regularly applied, pro- often viewed as a practice whose emphasis on vide exceptional values for endangered species fiber production results in unacceptable con- such as the red-cockaded woodpecker (Pi- sequences for other values, such as those that coides borealis Vieillot), game species such as benefit from maintenance of continuous for- bobwhite quail (Colinus virginianus L.), and est cover over time. Public lands in general, fire-dependent species such as wiregrass (pri- and national forest lands in particular, have marily Aristida beyrichiana Trin. & Rupr.). Then become the focal point for the replacement too, many foresters fondly recall the memory of clearcutting and planting with even-aged of the exceptional quality of lumber that ma- and uneven-aged reproduction cutting meth- ture stands of longleaf pine were, and are, ca- ods that rely on natural regeneration, and that pable of producing. can better achieve management goals that are The perception exists that many of these defined by residual stand structure and condi- values can be provided by management of tion rather than by harvested volume. longleaf pine especially through the use of Land managers in the southern United uneven-aged silviculture. In public debates, States are keenly interested in a renaissance this may be supported by little other than for longleaf pine (Pinus palustris Mill.) (Landers the layperson’s view that uneven-aged silvi- et al. 1995; Barnett 1999). Of the four south- culture is the opposite of clearcutting, and ern pines, longleaf pine has experienced the thus innately has something to recommend greatest percentage loss in forest area, from it. Foresters have been a bit more reluctant 37 million hectares prior to European colo- to wholly embrace the application of uneven- nization to approximately 2.2% of that cur- aged silviculture in longleaf pine, citing among rently (Frost this volume). That scarcity has in- other reasons the intolerance to shade of the

r James M. Guldin Arkansas Forestry Sciences Laboratory, Southern Research Station, USDA Forest Service, Monticello, Arkansas 71656.

217 218 III. Silviculture species, the difficulty in obtaining natural re- silvicultural prescriptions can be successful in generation, and the cost. longleaf pine stands (Farrar and Boyer 1991; These factors may in part explain why the Farrar 1996; Moser et al. 2002). There is also focus of discussion regarding uneven-aged sil- considerable debate about the implications of viculture in southern pines is especially promi- habitat quality for red-cockaded woodpeckers nent on public lands such as national and state in uneven-aged longleaf pine stands (e.g., En- forests, and private lands managed as game gstrom et al. 1996; Rudolph and Conner 1996). plantations. These ownership entities share a In view of the current situation, the oppor- number of attributes, including a diversity of tunity to develop and refine the application ownership objectives, and the capability, di- of uneven-aged silviculture in longleaf pine is rectly or indirectly, to subsidize timber pro- timely. duction with other resource values. For ex- This review of the selection method and of ample, the National Forests of Florida have the principles that underlie its application for made a commitment to manage longleaf pine longleaf pine is based on another southern using both even-aged and uneven-aged sys- pine species in which experience has been suc- tems. While this is admirable from the per- cessful over a long period of time—specifically, spective of using a diversity of reproduction naturally regenerated stands of loblolly pine cutting methods to meet a diversity of forest with a minor and varying proportion of short- management objectives, the proposed scale of leaf pine in the upper West Gulf Coastal the practice may outstrip the research that sup- Plain in southern Arkansas. Following that ports widespread application. overview, thoughts about the application and There is no reason to suspect that the prin- modification of the method to longleaf pine ciples of uneven-aged silviculture cannot be will be discussed in detail. successfully adapted to longleaf pine stands in the lower Atlantic and Gulf Coastal Plains. The method has been successfully applied over Definitions and Concepts time in other southern pines—most notably, mixed loblolly (P.taeda L.)–shortleaf (P.echinata The goal of any silvicultural system is to ad- Mill.) pine stands in the upper West Gulf vantageously utilize the resources in a given Coastal Plain (Baker et al. 1996). A review of stand for social benefit through the emulation that history of success and failure will be of of natural processes of succession and distur- value in providing perspective regarding the bance. Helms (1998) defines a silvicultural sys- application of the method in longleaf pine. tem as a planned series of treatments for tend- Uneven-aged silviculture has been success- ing, harvesting, and reestablishing a stand, and ful in different forest types (Guldin 1996). Suc- a regeneration method as a cutting procedure cess with the method depends on the ability to by which a new age class is created within obtain regeneration of the desired species, and the stand. Smith (1986) also makes this dis- to have that regeneration develop into mer- tinction, using the term “reproduction cutting chantable size classes. Conversely, failures with method” instead of “regeneration method.” uneven-aged silviculture are typically associ- These terms, “regeneration method” and “sil- ated with an inability to obtain desired regen- vicultural system,” are commonly misapplied eration (Guldin 1996; Guldin and Baker 1998). in two ways. The first is that they are often There is good reason to expect that the details mistakenly used interchangeably. The former of regeneration establishment and develop- refers to the short period of time during which ment under an uneven-aged system in longleaf a new age cohort of regeneration of the desired pine stands will be difficult, if the experience species is obtained, and the latter refers to the associated with the development of the shel- entire program of treatments for the life of the terwood method in longleaf pine (Croker and stand. The second is that they are often mis- Boyer 1975) is any indication. There is anec- takenly applied to a forest rather than to the dotal evidence to suggest that uneven-aged individual stand, which is their intended scope 7. Uneven-Aged Silviculture 219

(Smith 1986; Helms 1998). The confusion is system depend upon the origin of the regener- in part because the treatment prescriptions in ation, its age distribution, and its spatial distri- the silvicultural system are closely related to bution over the stand. stand structure at any given point in time, and The choice of regeneration method also has stand structure is established primarily using an inordinate inﬂuence on the overall course the reproduction cutting method at the point of silvicultural treatments and the way those of stand or cohort establishment. treatments are imposed in the stand (Fig. 1). The choice of regeneration method deter- In an even-aged stand, the sequence of sil- mines the scale of disturbance that foresters vicultural practices depends upon stand age. can imitate (Brockway et al. this volume). The new stand is obtained using a regenera- Even-aged regeneration methods such as tion method that results in a new cohort of re- clearcutting, the seed tree method, or the shel- generation across the entire stand. Subsequent terwood method are designed to emulate vary- treatments such as site preparation in advance ing intensities of stand-replacing disturbance of the new cohort, release of that cohort, and events, resulting in a new cohort of regen- intermediate treatments such as thinning also eration across the entire stand. Unevenaged occur across the entire stand. Each treatment is regeneration methods such as the selection imposed in a manner that is correlated with the method are designed to emulate a small-scale age of the new cohort of regeneration. Eventu- within-stand disturbance event, resulting in a ally, when the stand reaches maturity, a new new age class only in that subset of the stand reproduction cutting method is implemented where the practice was imposed. In either at the rotation age r, which gives rise to a sub- even-aged or uneven-aged methods, the main sequent stand managed with a subsequent sil- indication of success in the execution of a re- vicultural system. generation method is whether a new stand Conversely, in an uneven-aged stand, there of trees is successfully obtained to replace the is no rotation age r. Instead, treatments are trees that were removed during the harvest. based on a cutting cycle c , the basic inter- Thus, the reproduction cutting method is the val of stand entry, which varies from 5 to 20 primary element of the silvicultural system in years. Cutting cycle harvests are imposed in that the actions that comprise the silvicultural each stand every c years. However, the trees

Intermediate treatments Regeneration cutting RegenerationSite preparation cuttingRelease treatments (e.g., thinning) in next rotation

FIGURE 1. Chronosequential application of individual practices of a silvicultural system in even- aged stand. During the rotation age r, treatments are applied Stand biomass across the entire stand to meet silvicultural objectives that are related to tree age. Rotation length r 220 III. Silviculture

FIGURE 2. Concurrent applica- Regeneration cutting tion of individual practices of an uneven-aged silvicultural system Site preparation during a cutting cycle harvest in a balanced uneven-aged stand. Release treatments Treatments are applied to subunits of the stand depending on Intermediate treatments conditions within each subunit. (e.g., thinning) Each cutting cycle harvest will support similar treatments. Stand biomass

C1 C2 C3 − Cutting cycles C1 C3 removed during a cutting cycle harvest are selection method, used to regenerate and taken for different reasons—regeneration cut- maintain a multiaged structure by removing ting, release, or intermediate treatment such as trees either singly, or in small groups or strips thinning (Fig. 2). As a result, the different sil- (Helms 1998). By definition, an unevenaged vicultural practices that occur chronosequen- stand has at least three age classes (Smith 1986; tially in an even-aged stand and that cover its Helms 1998). In practice, age is not measured; entire area are conducted at the same time in stands are managed by controlling either the an uneven-aged stand, but only over a portion volume that is harvested, the diameter distri- of the stand area. This complicates the prac- bution in reference to a target distribution, or tical implementation of the method. Uneven- the area within the stand that is cut. aged systems are often thought to be less inten- In an uneven-aged stand, growing space is sive than even-aged systems, and that may be subdivided among trees of all size classes, from true in relation to, say, the degree of site expo- regeneration through the largest overstory sure during regeneration cutting or the capital trees. A starting point to determine the appro- outlay required to establish a new stand. But priate basal area to maintain in an uneven- uneven-aged systems require more attention aged stand is to apply the basal area found in on the part of the forester, and can be more a mature even-aged stand of the same species inefficient to conduct in some ways because that is marginally fully stocked or slightly un- the scale of operation is at the scale of subunits derstocked. A stand in that condition will have within a stand rather than across the entire a slight amount of growing space available in stand. the understory for the regeneration establishment but will be marginal for regeneration development. If that basal area is translated to an uneven-aged stand, the heterogeneous condi- The Selection Method—An tions that typify an uneven-aged stand will re- Overview sult in fully stocked clusters of overstory trees in some areas, and other areas that are suf- Uneven-aged silviculture is implemented us- ficiently understocked such that regeneration ing a reproduction cutting method called the development can occur. This concept gives rise 7. Uneven-Aged Silviculture 221 to two generally accepted methods by which form, condition, or other attribute, from the the selection method is implemented—one us- competition provided by the tree being re- ing very small openings and the other using moved. Removal of a large-diameter tree, such larger openings. as one at or larger than the maximum diameter desired for retention, is intended to create a canopy opening within which regeneration Single-Tree Selection Method is to become established and to develop. Singletree selection imitates the smallest scale of disturbance, such as when a single tree Group Selection Method falls or dies while standing in the woods. Possible causes of such a disturbance are in- In the group selection method, trees are re- sects, lightning, disease, windthrow, or some moved and new age classes are established in other agent. In unmanaged stands, this pro- small groups (Smith 1986; Helms 1998). Group cess results in gap-phase dynamics indicative selection imitates a small-scale natural distur- of late-successional conditions (White 1979; bance that kills a small group or cluster of trees Runkle 1982; Pickett and White 1985). When within a stand; examples include mortality of a large tree in an uneven-aged stand is re- trees from a blowup of a surface fire or an infes- moved, the growing space used by that tree tation of pine beetles. In theory, the nature of is made available to adjacent trees, to smaller the disturbance should promote a suitable and trees in the midstory, and to regeneration in receptive seedbed for seedling establishment, the understory. In the smallest gaps, the gap and the size of the gap that is created by the may close before the regeneration can accede disturbance is large enough to promote regen- into the main canopy, and the regeneration eration development. The seed source for that may then persist without further growth or new cohort can be advance growth of regener- may even succumb to suppression. The occur- ation in place prior to creation of the opening, rence of multiple gaps (where a nearby tree seed from trees that border the gap or stored in succumbs and creates a second nearby open- the soil beneath the gap, or as seedfall dissemi- ing, either concurrently with or soon after the nated from the trees being removed within the first), or expansion of existing gaps (where gap at the time of their removal. The size of the gap-bordering trees fall) can tip this ecological gap affects the species composition of that new balance in favor of regeneration survival and cohort. Larger gaps favor species of greater in- development. tolerance to shade, while smaller gaps favor In the single-tree selection method, indi- shade-tolerant species. vidual trees of all size classes are removed The group selection method is applied or more or less uniformly across the stand (Smith suggested when there is a desire to use an 1986; Helms 1998). This is typically conducted unevenaged silvicultural system in a stand, by first identifying the trees that are to re- yet still regenerate shade-intolerant species; main in the stand. The trees to remove then hence the interest in the method relative to become obvious because their removal pro- the southern pines. The maximum size for motes the continued growth and develop- group openings depends on one’s interpre- ment of the trees that are to remain. The tation of ecological literature as modified by diameter of the tree being removed is di- prevailing forest management guidelines. It rectly related both to the silvicultural objec- is generally agreed that the upper ecologi- tive for its removal, and to the retention of cal size limit for tolerant species within a cir- stocking levels by size class deemed desir- cular group selection opening is one whose able across the residual diameter distribution. radius equals the height of the surrounding Small-diameter trees, such as pulpwood or trees in the stand (Helms 1998). In trees with small sawlogs, are removed according to clas- a height of 25 m, the ecological upper limit sical thinning rationale—to free an immature would be on the order of 0.2 ha. Larger open- neighboring tree, presumably one with better ings would then be suggested for intolerant 222 III. Silviculture species such as longleaf pine. The suggested contracting treatments, since the precise loca- maximum group opening size on national for- tion of each group can be specified. est lands within the Southern Region of the USDA Forest Service is 0.8 ha, which should allow for acceptable regeneration establishment Selective Cutting and development of any of the southern pines. The term “selective cutting” is often used In practice, some trees are typically removed to describe harvesting activity that resembles from the stand matrix between the groups as uneven-aged regeneration cutting in that trees well, so as to promote stand health and accept- remain on the site. But a strict definition of able basal area levels between group openings, the term is “select some trees to cut and cut and to prepare trees as seed producers in sub- them”; it has no commonly accepted pro- sequent group openings. fessional meaning except a derogatory one. The decision to locate a group opening The term does not refer to, and is not syn- within the stand is usually done to alleviate un- onymous with, the practice of uneven-aged derstocking, rectify excessive stem density, or harvests using the selection method. All too take advantage of existing regeneration. If part often, stands harvested using selective cutting of the stand is understocked, creating an open- are high-graded—harvest is uncontrolled, the ing gives the forester an opportunity to regen- best trees are cut, and the poorest remain. erate that group, and thus to restore that part Perhaps the most important silvicultural dis- of the stand back to full stocking. At the other tinction between the selection method and se- extreme, if there is a place within the stand lective cutting is that under the latter, no provi- where the trees are all in a surplus size class on sion is made for establishment or development the marking tally, creating an opening harvests of desired species of regeneration. This is a trap those trees and helps the forester more quickly into which improperly applied harvests under achieve the desired residual stand density. Fi- the selection method can fall. nally, if a species is difficult to regenerate naturally, one should not fail to create openings within the stand where advance regeneration is found. Regulation of Group selection has a number of administra- Uneven-Aged Stands tive advantages over single-tree selection that contribute to its popularity. Most group open- Regardless of whether single-tree selection or ings serve as points of concentration for logging group selection is implemented, the methods operations in the immediate area; logs are fre- for ensuring the regulation of growth and har- quently decked in the openings, and haul roads vest are the same. Two methods of regula- typically run from one group to another. As a tion have historically been associated with the result, group openings are often heavily scar- selection method—regulation of the sawtim- ified. This is an advantage in promoting pine ber volume in the stand through harvest of regeneration, which requires exposed mineral growth, and regulation of the structure of the soil for optimal seed germination and estab- stand through conformance with a target di- lishment. Moreover, the group opening is the ameter distribution. With group selection, a only part of the stand in which regeneration third method enters the picture, in which regu- is expected. Site preparation or release treat- lation is based on proportion of harvested area ments can thus be restricted to the groups, during each cutting cycle. which is an advantage in that less area must These methods have varying degrees to be treated and the specifications for treatment which they conform to the origins of uneven- can be made clearer than when a compara- aged silviculture (Guldin 1996). The selection ble treatment is prescribed in a single-tree se- method, the most recently developed of the lection stand. The advent of geographic posi- regeneration methods historically, traces its tioning system technology adds to the ease of origins to the Dauerwald in Germany, which 7. Uneven-Aged Silviculture 223 evolved as a counterpoint to area-based regu- prepare a stand and stock table that quantifies lation methods. Under the Dauerwald, allow- the stem density and sawlog volume before able cut was based on stand volume growth; a harvest by diameter class per unit area. This volume equal to roughly 15 times the annual requires application of an appropriate local growth was retained on the site (Troup 1928, sawlog volume table, so that the sawlog vol- 1952). Other European work at the turn of ume by diameter class can be included in the the century used a negative exponential alge- table. braic relationship among the number of trees Second, the compound growth rate of the by diameter class to quantify the reverse J- stand must be determined. This is usually done shaped curve that approximates an uneven- by averaging the growth rate for a number of aged stand (deLiocourt 1898), and that ap- trees of varying size in the stand, using 10- proach was further developed in North Amer- year radial increment measured from incre- ica in modern approaches to forest management cores, and calculated per tree using the ment (Meyer et al. 1961). Little historic evi- formula dence exists to support an area-based regulation method in the evolution of uneven-aged V0/Vp G = exp − 1 silvicultural systems. On the other hand, Smith n (1986) points out that regulation method should be independent of silvicultural system; where under such logic, any of the following systems G = growth rate (%) might legitimately be applied to regulation of V = previous tree volume uneven-aged stands. p V0 = present tree volume n = growth interval (years) Regulation of Volume Alternatively, experience shows that in man- The most successful application of the selec- aged stands, one can use an appropriate com- tion method in southern pines has been at pound growth percentage for sawlog volume the Crossett Experimental Forest (EF) in Ash- increment. Values between 6% and 8% are ley County, AR, in the mixed loblolly–shortleaf typical in unmanaged and managed loblolly– pine forest type of the upper West Gulf Coastal shortleaf pine stands, respectively, in south Plain. The Crossett EF, managed by the South- Arkansas. An appropriate range for longleaf ern Research Station of the USDA Forest Ser- pine would be based on local experience. vice, was established in 1934 and is still ac- Either approach then requires the forester to tive. It supports several long-term research determine an after-cut volume for the planned studies and demonstrations, notably the Good harvest. The after-cut volume is the cumula- and Poor Farm Forestry Forty Demonstration tive volume to which the current stand must Stands and the Methods of Cutting Study. be reduced. That level is set by predicting the The uneven-aged stands of loblolly– future cutting cycle length, by selecting the fu- shortleaf pine at the Crossett EF were regulated ture stand volume sought at that time, and using the volume control-guiding diameter then by calculating the volume to which the limit (VCGDL) method (Reynolds 1959, 1969; current stand must be reduced so that it will Baker et al. 1996; Guldin 1996, 2002). In this grow to the intended future volume at the ap- method, sawlog volume and volume growth propriate rate of growth. are used to calculate harvests, and trees are Third, the allowable cut is calculated as the marked based on whether their individual difference between the before-cut volume and growth rates are sufficient to allow them to the planned after-cut volume. This leads di- maintain acceptable sawlog volume growth. rectly to the calculation of the guiding diam- Implementation of the VCGDL method has eter limit (GDL), which is that diameter class four broad steps. First, a current inventory of that meets the allowable cut if all trees in larger the stand is taken. That inventory is used to diameter classes are cut. Usually, part of the 224 III. Silviculture

GDL class must also be cut to exactly match also requires that large high-value trees above the allowable cut. the GDL have a compelling record of growth Finally, the field crews are given the GDL to be retained. class and the percentage of the GDL class to cut. However, there are several limitations of Markers are instructed to retain trees above the the VCGDL method. The approach does not GDL if they are growing acceptably, and then to provide any evidence to the forester about mark an equivalent volume to that retained in the growth of trees below the sawtimber size diameter classes smaller than the GDL. Mark- class. Foresters must judge in some other way ing crews can only do this efficiently by memo- whether regeneration is being established, and rizing the appropriate local volume table. The whether sub-sawtimber size classes are devel- crews then must keep a running tally, either oping at an acceptable rate. That judgment is mentally or on a notepad, of cumulative vol- typically based on experience rather than ob- ume retained above the GDL and that removed jective standards, and that can be a limitation. below the GDL. At the end of the marking, the Second, the method requires a high degree of volume marked below the GDL should balance experience on the part of the field crews who that retained above the GDL (Fig. 3). are marking the stand, especially in regard to The VCGDL method has a number of ad- estimation of the volume of trees above and vantages. It requires the field crew to examine below the GDL that are being retained and sawlog component of the stand from the per- marked, respectively, such that the cumula- spective of trees that should be retained. As a tive volume tally balances when the marking is result, crews can balance whether to cut and completed. Finally, because decisions are made leave trees across a range of diameter classes. It in the field about retaining trees above the GDL

60 Before cut GDL target 50 After marking 40

Trees per hectare Trees 20

0 12.5 17.5 22.5 27.5 32.5 37.5 42.5 47.5 52.5 57.5 62.5 67.5 72.5

DBH class (5-cm midpoint)

FIGURE 3. The volume control-guiding diameter limit (VCGDL) regulation approach conceptually applied in the 1978 before-cut inventory from the Good Farm Forestry Forty demonstration stand, an uneven-aged loblolly–shortleaf pine stand on the Crossett Experimental Forest in southern Arkansas. Before-cut, GDL target, and after-cut diameter distributions are drawn as curves rather than histograms. The GDL target reflects the allowable cut in sawtimber cubic volume based on 6% growth rate. The after-cut diameter distribution illustrates one possible outcome resulting from retaining trees above the guiding diameter limit, and removing trees below the limit such that the volume of the stand is retained at the guiding level. 7. Uneven-Aged Silviculture 225 and removing trees below the GDL, the use to any desired residual basal area B per unit of computer models to predict stand develop- area in the stand. Specification of B, D, and q ment in advance of harvest is difficult. thus constitutes a unique solution of diameter The VCGDL regulation approach evolved distribution. This approach, called the BDq ap- as a means to regulate single-tree selection proach, is used to generate the target balanced stands. However, modifying the approach to diameter distribution against which the exist- regulate a stand being managed using group ing stand structure is compared. The method selection is relatively straightforward. All trees was developed by Leak (1964), and its prac- within the group would be marked, and added tical implementation was described in detail to the marking tally. Trees smaller than the by Marquis (1978). Modifications for uneven- GDL within a group would have their volume aged stands of loblolly–shortleaf pines in the added to the below-GDL cumulative volume West Gulf region were described in Baker et tally. That would lead to retaining an equiva- al. (1996), and for southern pines generally by lent volume of trees at or above the GDL in the Farrar (1996). matrix between groups. Simply stated, selection of the target BDq parameters allows the forester to calculate a unique hypothetical target diameter distribu- Regulation of Stand Structure tion. This is typically prepared on a unit area The regulation of stand structure is based on (per-hectare) basis. The diameter distribution the notion that the diameter distribution of a of the before-cut stand, prepared from a pre- balanced uneven-aged stand has an ideal the- harvest inventory, is then compared to that oretical relationship, which can be compared target. In an ideal case, the before-cut stand with the actual stand structure for generating will contain a surplus of trees in every diame- an after-cut residual stand (and indirectly, a ter class compared with the target; that sur- marking tally) that carries the existing stand plus then becomes the marking tally. How- closer to the theoretical ideal. Several ap- ever, far more common is the situation in proaches can be developed to quantify this which some diameter classes in the prehar- ideal theoretical stand, such as use of stand vest stand will contain a surplus of trees com- density index (Long 1998) or leaf area in- pared to the target, and others will contain a dex (O’Hara 1996). But the most common in deficit of trees. Here, the basal area of those southern pines is based on the assumption of deficits must be calculated, and that basal area a constant ratio q in the number of trees in deficit must be accounted for by retaining more adjacent size classes, according to the simple trees than called for in those diameter classes formula that have a surplus relative to the target stand (Fig. 4). Ultimately, deficits in a given diameter tn q = class will be corrected through ingrowth from t + (n i) smaller diameter classes over time. where When the final tally of trees to cut is determined for each diameter class, the proportion q = q ratio of trees to cut by diameter class is calculated. t = number of trees per unit area in the nth n That information—number of trees to cut, and diameter class percentage, by diameter class—is given to the t + = number of trees per unit area in the (n i) field crews, who use that information as they next larger class of class width i mark the stand. Field crews will find it easier Thus, q is dependent on diameter class width, to base their marking on the proportion rather and the use of a given q ratio must include ref- than the absolute number, since it is easier to erence to the class width i. If the maximum think about removing a set percentage of a diameter class D of trees to retain in the stand given diameter class rather than an absolute is known, one can use q to construct a neg- number of trees in a given diameter class per ative exponential relationship that can be fit unit area. That also allows the reinforcement 226 III. Silviculture

70 Before cut 60 BDq target 50 After cut

Trees per hectare Trees 20

0 12.5 17.5 22.5 27.5 32.5 37.5 42.5 47.5 52.5 57.5 62.5 67.5 72.5

DBH class (5-cm midpoint)

FIGURE 4. Regulation of stand structure using the basal area–diameter–q ratio (BDq) method conceptually applied in the 1978 before-cut inventory from the Good Farm Forestry Forty demonstration stand, an uneven-aged loblolly–shortleaf pine stand on the Crossett Experimental Forest in southern Arkansas. Before-cut, BDq target, and after-cut diameter distributions are drawn as curves rather than histograms. The BDq target reflects a B = 14 m2 ha−1, D = 57.5 cm, and q (5-cm classes) of 1.44. The after-cut diameter distribution illustrates the compensation by basal area according to the q ratio; the basal area in deficit diameter classes is retained in surplus diameter classes such that the target basal area is retained. to be given that the poorest percentage of trees stand, and finally to determine if the projected in the diameter class should be marked for har- harvest is operable. vest, and the best trees retained. The number Structural regulation has the advantage of of size classes the field crews must work with objectivity. When generating a target diam- can be reduced if broader product classes are eter distribution, target diameter class data used. For instance, a fivefold product classifica- can be calculated for submerchantable diam- tion that includes small pulpwood, large pulp- eter classes as well, which can provide guid- wood, small sawtimber, medium sawtimber, ance about whether cutting-cycle harvests are and large sawtimber would be very convenient providing acceptable regeneration establish- for field crews to apply. ment and development through the submer- If some method other than the BDq ap- chantable component and acceptable recruit- proach is used to generate the target structure, ment into the merchantable component of the the process for implementation still is most ef- stand. This depends on the assumption that the ficient if conducted as described above. Sup- mathematical relationship used to characterize pose a target diameter distribution is generated the stand structure is biologically meaningful using a power function, for example, rather at the smallest size classes, and this may not than the negative-exponential BDq approach. be the case for the negative exponential rela- Once the target diameter distribution is ob- tionship upon which q is based (Baker et al. tained, there is still need to compare the ex- 1996). isting stand to that target, generate a mark- The main disadvantage is that the process ing guide, and to compensate for diameter for calculating the marking tally is cumber- class deficits between the preharvest stand and some, especially in cases in which deficit di- the target, so as not to overcut the preharvest ameter classes are adjusted according to the 7. Uneven-Aged Silviculture 227 mathematical relationship used in the initial sizes less than two tree heights in radius, such target calculation. Spreadsheet programs are that the entire group opening is under the eco- available to assist this calculation (Baker et al. logical influence of the gap-bordering trees, 1996). A second disadvantage is that the ap- would also provide ecological distinctions with propriate B, D, and q parameters for applica- patch clearcutting. tion to longleaf pine have yet to be identified through research or practice. Adaptive Experience in Area-Based Regulation Southern Pines Regulation of stands managed by group selec- The long-term studies and demonstrations tion has been advocated using the area-based at the Crossett EF in south Arkansas pro- regulation concept borrowed from even-aged vide keys to the successful implementation of forest management (e.g., McConnell 2002). the method in mixed loblolly–shortleaf pine Under this simple device, the initial decision stands. That background is the best source of is to establish a rotation age, r, for the trees in experience with the method in the South and the uneven-aged stand, essentially the age at serves as a point of departure for considering which trees for the species under management the application to other forest types. are typically harvested in a comparable even- aged context. The area a of the stand is then divided by the rotation age r, and the quotient Regeneration represents the proportion of the stand area a At the stand level, the first indicator of long- to be cut annually. That is converted to a,an term forest sustainability is whether adequate area to be cut in a given cutting cycle harvest regeneration is obtained when a regeneration by multiplying the annual percentage of area cutting is made in a stand. This is especially true to cut times the length of the cutting cycle, ac- with the selection method, which requires a cording to the simple formula delicate balance between the stocking and development of the merchantable component of A = (a/r)c c the residual stand versus the stocking and de- where velopment of seedlings and saplings. It applies also in situations where conversion from even- A = area to be cut in a cutting-cycle harvest c aged condition to uneven-aged condition is a = stand area imposed. It is critical to obtain regeneration af- r = rotation age and ter a cutting cycle harvest in any conversion, c = planned length of the cutting cycle (years) transition, or initial steps in implementation The problem in using area-based regulation of either the single-tree or group selection with group selection is more theoretical than method. practical. It is difficult to distinguish be- Abundant seed crops are an excellent at- tween group selection and patch clearcutting, tribute on which to rely when prescribing a the small-opening even-aged variant of the reproduction cutting method that depends on clearcutting method that is also regulated using natural regeneration. Long-term data on seed- this approach. The best way to draw a distinc- fall in loblolly–shortleaf pine stands on the tion between the area-based regulation of the Crossett EF show that, on average, natural re- group selection method versus patch clearcut- generation is adequate or better four years in ting is through applications that increase the five, and rarely do seed failures occur in two within-stand heterogeneity of structure. Ex- consecutive years (Cain and Shelton 2001). amples include varying the area cut in any one This prolific seedfall is one of the reasons un- cutting cycle, varying group size and shape, or derlying the successful application of either placing openings in a pattern that is not ge- even-aged or uneven-aged reproduction cut- ometric or predictable. Use of group opening ting methods that rely on natural regeneration 228 III. Silviculture in this forest type and region (Baker and Mur- phy 1982; Zeide and Sharer 2000; Cain and Shelton 2001). Conversely, irregular seed crops can reduce the probability of success in obtaining natural regeneration under the selection method. This is important both for initial period of conversion to uneven-aged structure, and in maintenance of that structure. Experience at the Cros- sett EF suggests that failure to secure a new age class following a given cutting cycle harvest was not in itself an impediment to maintaining desirable uneven-aged structure, but missing two age classes in consecutive cutting cycle harvests is to be avoided (Reynolds 1969). If a given cutting cycle harvest fails to secure a new age cohort of regeneration, supplemental site preparation efforts should be conducted at the next cutting cycle harvest to ensure that regeneration is obtained (Guldin and Baker 1998).

Rehabilitation of Understocked Conditions The stands on the Crossett EF originated as cutover understocked stands, and were managed FIGURE 5. A loblolly pine sapling on the Crossett in a manner by which stocking was built over EF that meets the minimal size criteria for response time. The stands had been harvested by the to release—a 20% live crown ratio, and diameter Crossett Lumber Company in 1915 to a 38-cm outside bark at the base of the live crown of 5 cm. A stump limit, roughly equivalent to a 30-cm di- similar rule of thumb for response to release would ameter limit. No management occurred on the be helpful to have in applying the selection method area until it was leased to the Forest Service to longleaf pine. (James M. Guldin) in 1934 (Guldin 2002). These stands were not fully stocked, homogeneous, and even-aged; rather, the stands showed considerable within- live crown is greater than 20% and if the stem stand heterogeneity at the start. The research diameter at the base of the live crown exceeds and demonstration work that began at that 5 cm (Fig. 5). point successfully restored understocked and The second was the use of herbicides to con- marginally stocked stands back to full stocking trol hardwoods that were competing with the through harvest of a portion of growth. Two pines. Effective hardwood control was critical elements of this work were especially impor- both as site preparation for the establishment tant. and development of new seedlings and also The first was the reaction of these pines to as a release treatment and liberation cutting removal of competition. In the upper West Gulf (Smith 1986) to free established pine saplings Coastal Plain, both loblolly and shortleaf pine and small merchantable stems. Thus, the abil- respond to release at advanced age. Data from ity to use herbicides effectively to control hard- studies at the Crossett EF (Baker et al. 1996) woods competing with pines, and to then have suggest that pine stems in the 10- to 15-cm the pines respond quickly to the growing space diameter class will respond to release if their made available, lies at the root of success in 7. Uneven-Aged Silviculture 229 using the selection method to manage loblolly– loblolly–shortleaf pine stands at the Crossett shortleaf pine stands in the West Gulf re- EF, the target residual basal area after a cutting gion. cycle harvest is roughly 14 m2 ha−1, and the stands grow approximately 0.55–0.7 m2 ha−1 in basal area annually. After 5 years, the stands Developmental Dynamics will have grown in basal area to about 17 m2 Uneven-aged stands exist in a delicate balance ha−1, and the subsequent cutting cycle harvest between understocking and growth. Baker can then be imposed. et al. (1996) describe this balance in the con- Because basal area can also be related to text of shade management. That balance is stand volume, operational feasibility of har- controlled by three factors: the distribution of vests can be tested. In south Arkansas, loblolly– basal area retained in the residual stand imme- shortleaf pine stands support after-cut volumes diately after a cutting cycle harvest, the length of roughly 26–27 m3 ha−1 of sawlog volume, of the cutting cycle, and the operability of the and the stands grow on the order of 2.2–2.4 future cutting cycle harvest. m3 ha−1 of sawlog volume annually. After After a cutting cycle harvest, the residual 5 years, the stands will support roughly 37– overstory trees grow and some degree of in- 40 m3 ha−1 of sawlog volume. Operable har- growth into the overstory also occurs. As a re- vests in the vicinity are about 9 m3 ha−1 of sult, stocking levels increase in the overstory sawlog volume. Thus, sawlog volume growth over time. But this increased stocking serves to of the stand can be cut on roughly a 5-year increasingly inhibit the development of regen- interval with operational harvests, which also eration. Subsequent cutting cycle harvests will maintains regeneration development. Metrics be needed to reduce overstory stocking suffi- such as these are needed for longleaf pine ciently to allow the continued development of stands. the initial regeneration cohort, and to obtain Another clue about the appropriate upper a new cohort. On the other hand, if a given limit of basal area is whether acceptable rates cutting-cycle harvest removes too much basal of height growth can be maintained in the re- area, the stand will not grow rapidly enough generation component. In loblolly and short- to allow an operable cutting cycle harvest in leaf pine stands, height growth of regenera- the subsequent cutting cycle. tion is a useful indication of maintaining the One simple metric for the upper limit of ability to recover full growth potential. Mini- acceptable basal area to carry in an uneven- mum acceptable annual height growth in these aged stand is to quantify the residual basal species is 0.15 m. If seedlings or saplings less area in a classic low thinning at which acciden- than 1.3 m in height are not growing at this tal regeneration just begins to be suppressed. rate, they will probably not survive. This point approximates the highest acceptable Finally, the Crossett EF experience sug- before-cut basal area in uneven-aged stands. gests a final visual clue for determination For example, pine regeneration can become of acceptable balance between overstory and established in even-aged Coastal Plain loblolly understory—the presence of foliage of the de- stands thinned to 16 m2 ha−1, but will cease to sired species at all levels of the canopy pro- make acceptable height growth if basal area file in the stand. If regeneration is success- exceeds 17–18 m2 ha−1. Because overstory fully established and making acceptable height tree distribution in uneven-aged stands is more growth, and if repeated cutting-cycle har- heterogeneous, these basal area levels repre- vests are successful in obtaining regeneration, sent an upper limit to the acceptable basal seedlings and saplings will be visible in the area range for successful uneven-aged pre- stand. The longer the period of successful silvi- scriptions. culture under the selection method, the more The lower limit is defined by maintaining ac- prominently will foliage of the desired species ceptable overstory growth over the expected be found at all levels of the canopy profile duration of the cutting cycle. In uneven-aged (Fig. 6). 230 III. Silviculture

EF, the selection method has a reputation as one that produces sawtimber of high quality (Guldin and Fitzpatrick 1991); the long-term application of a marking rule such as this contributes to that reputation. This rule raises distinctions between regulation by volume under the VCGDL and regulation by structure under the BDq method. It is easier to leave the best trees under the volume control regulation method versus the BDq, because the marking tally in the BDq is specific to a given diameter class whereas the marking tally of the volume control method cuts across diameter classes. Under VCGDL, a residual tree is judged to be part of the population of “best” trees regardless of diameter class. Conversely, in the BDq method, a tree that is retained is judged relative to other trees in that diameter class only, and the proportion to cut changes from one diameter class to another. For example, suppose that a BDq marking tally requires removal of 1 in 10 trees in the 45- cm class, but half of the trees in the 30-cm class. Field crews will invariably come across a tree FIGURE 6. A view of conditions in an uneven- in the 20th percentile of quality in the 45-cm aged stand of loblolly–shortleaf pine within the Poor class immediately adjacent to a smaller tree of Farm Forestry Forty Demonstration Area on the better absolute form and with better develop- Crossett EF immediately following the cutting cycle harvest in the spring of 2003. Note the presence mental potential in the 40th-50th percentile of foliage of the desired pines at all heights in the of quality in the 30-cm class, and will com- canopy profile, a simple visual clue that denotes sus- plain about marking the better tree and leaving tainable regeneration cutting over time in uneven- the poorer one. The answer for the field crews aged stands. (James M. Guldin) in that event is to use common sense, and to mark the poorer tree. Carrying that logic to its conclusion leads to a critical point relative to the BDq method. Of the B, D, and q variables, Marking Rules residual basal area is most important to retain, During his tenure at Crossett, CEF founding followed by maximum diameter; q is least im- scientist Russ Reynolds explicitly refused to portant. Some thought has been given to mod- identify his volume-control method as “single- ifications of the BDq method as a BD method tree” or “group” selection; he called it “se- (Baker et al. 1996); this would result in essen- lection.” Occasionally large openings would tially a basal area control method implemented occur, occasionally small ones would suffice. in a manner similar to regulation by volume, Reynolds’s key decision was whether the tree but in which the basis for compensation among being examined while marking was of accept- trees being retained is by equivalence of basal able form, size, and quality to retain. Reynolds area rather than volume. captured this concept in the simple phrase, “cut Reynolds’s marking rule also raises distinc- the worst and leave the best” (Reynolds 1959, tions between group selection and single-tree 1969). Attention to this simple marking rule selection. A rule that guides the forester to “cut ensured that stem quality was gradually im- the worst and leave the best” can be more proved over time. As practiced on the Crossett strictly followed in single-tree selection than 7. Uneven-Aged Silviculture 231 in group selection. In most instances, the lo- calized area of the stand within which a group opening is planned will contain trees that under an individualistic evaluation would qualify for retention. That might allow one to further refine the logic for placement of group openings—locate the opening in those parts of the stand where a disproportionate number of the trees within the planned group are of poorer condition than those in other parts of the stand. Such manipulation in the location of group openings is possible if the group selection method is being implemented using regulation by volume or structure. However, improvement in residual stand condition as a result of this marking rule is by definition un- likely to occur under area regulation of group selection, especially if imposed using strict ge- ometric patterns.

The Selection Method in Longleaf Pine

Interest in implementing the selection method in the longleaf pine forest type is driven by FIGURE 7. A view of the shelterwood method in ap- a number of considerations. Foremost among plication to longleaf pine in 1982 on the Escambia them is to develop habitat conditions in lon- Experimental Forest, Brewton, AL. The residual 2 −1 gleaf stands that favor the species that inhabit basal area in the overstory was 7 m ha , and these stands, such as bobwhite quail (Moser seedlings have emerged from the grass stage several et al. 2002) and the red-cockaded woodpecker years following the seed cut. (James M. Guldin) (McConnell 2002). To a certain extent, argu- ments about habitat condition that can be de- detailed studies of the reproductive biology veloped in uneven-aged stands of longleaf pine and silvics of longleaf pine were fundamen- are premature without a careful examination tal to the development of the even-aged shel- of what a sustainable application of the se- terwood method (Fig. 7). A subjective inter- lection method would look like in longleaf pretation of these sources suggests that a suc- pine, using the subjective metrics developed cessful prescription for the selection method from our understanding of the method in the in longleaf pine will require attention to re- loblolly–shortleaf pine forest type. generation establishment, the pattern of im- The state-of-the-art treatise on the selection plementation, the approach to regulation, and method in longleaf pine (Farrar 1996) is a pri- developmental dynamics. Among the largest mary source for managers to consider as the se- challenges will be the integration of prescribed lection method is operationally applied in lon- fire as a standard element of the method. gleaf pine. Equally important in application to the selection system in longleaf pine is research Regeneration on longleaf pine autecology that culminated three decades ago on the Escambia EF near The application of natural regeneration in a Brewton, AL (Croker and Boyer 1975), where selection method for longleaf pine will be 232 III. Silviculture difficult. Seed production is much less reliable in longleaf pine, where adequate seed crops only occur between 10 and 20% of the time (Wahlenberg 1946), than in loblolly pine. The degree of silvicultural attention required to make a successful prescription involving natural regeneration will be greater for longleaf pine than for loblolly pine, if for no other reason than the greater infrequency of adequate seed crops. This can be especially problematic in mixed-species southern pine stands that include longleaf pine as part of the mix, because the other pines will be more prolific seed producers. Careful attention to silvicultural detail is needed to ensure practical success with natural regeneration in longleaf pine. For example, the key to the development of the shelterwood method in longleaf pine was the detailed work by Croker (1973). He reported that over a 7-year period, cone production in a longleaf pine stand reached an optimum when the stand basal area of longleaf pine was 6.88 m2 ha−1 and declined as basal area decreased or increased from that level. Greater overstory basal area resulted in less seedfall and reduced numbers of seedlings. A uniform residual over- FIGURE 8. Longleaf pine seed after seedfall in fall story of 10.33 m2 ha−1 resulted in virtually 1982, a good seed year in a managed even-aged lon- no surviving saplings over time (Boyer 1993). gleaf pine stand in central Louisiana. A prescribed Fieldcraft such as that described in the devel- fire had been used that year to prepare the seedbed opment of the shelterwood method (Croker for the anticipated seed crop. (James M. Guldin) and Boyer 1975) would improve natural seedfall in any selection method applied in lon- window within which to complete a cutting gleaf pine stands. Because cone production cycle harvest. Hopefully, the period when the is a highly inherited trait genetically (Croker stand is actually cut would occur in conjunc- 1964), marking crews should include an eval- tion with an adequate seed year. However, uation of past cone production as a decision the administration of sales on National Forest element in whether to retain a tree during cut- lands precludes the ability to guarantee har- ting cycle harvests (Fig. 8). vest in conjunction with seed crops and recep- On national forest lands in the South, an- tive seedbeds. Logging contractors are typically other practical approach for management of given several years to harvest a timber sale, longleaf pine using the selection method is and cultural work to improve seedbed condi- to plan for natural regeneration, but to use tion must be programmed a year in advance. planting as a fallback position to prevent ex- The second chance is to catch the first good cessive delays in reforestation. There will be seed crop after the sale closes. Site preparation two opportunities for successfully obtaining should be conducted to prepare a receptive natural regeneration prior to planting. The seedbed when that seed crop occurs. This, too, first chance is that associated with the ini- is constrained by administrative procedures. tial harvest. Foresters with the USDA Forest Site preparation on National Forest lands is Service generally allow a logger a multiyear funded using proceeds from timber sales under 7. Uneven-Aged Silviculture 233 the Knutsen–Vandenberg Act of 1933, which More than any of the other southern pines, limits expenditures to a 5-year period follow- longleaf pine will benefit from some applica- ing sale closure. If longleaf pine produces one tions that involve planting as the source of good seed crop in 5 years, the chances are regeneration. However, research experience that site preparation can be timed to that ex- with that is limited. Probably a minimum num- pected seedfall. However, site preparation dol- ber of seedlings is rationally feasible to plant. lars must be requested in the fiscal year in ad- If natural regeneration is inadequate, and vance of that in which they would be spent— planting is needed, one should plant enough and the ability to predict a good seed year is seedlings to ensure a fully stocked stand of limited to 6 months in advance of seedfall (Cro- planted seedlings. The number to plant will ker and Boyer 1975). vary depending on seedling stock and the Practically, then, a forester with the USDA level of understory competition. For example, Forest Service has 4 years after the cutting cy- 1000–1200 seedlings per hectare may be suf- cle harvest is concluded to obtain natural re- ficient using containerized planting stock and generation; the fifth year must be devoted to a preplant herbicide treatment in a group se- spending the available funds to prepare the site lection opening. Without these refinements, and plant seedlings. Standards should be de- more seedlings will be needed. veloped that provide guidance to silviculturists Planting should be done in association with about when natural regeneration difficulties an effective site preparation prescription that are likely to be profound (such as poor seed promotes survival and height growth of the producers left on the site, an absence of ad- planted seedlings. Genetic improvement of vance growth, or the lack of adjacent stands planting stock has produced seedlings that that could contribute seed to a cutover area make rapid early height growth under open from adjacent mature trees). That informa- conditions with intensive site preparation. Al- tion could help foresters identify stands that though families selected for rapid growth in the might be good candidates to plant initially un- open would probably be successful if planted der residual overstories, rather than to tackle beneath a residual overstory or within a group the chain of efforts to synchronize site prepa- opening that is under the ecological influence ration to seedfall. of the overstory trees that surround the group, In the past decade, a focus of regeneration there are opportunities to explore the best fam- research in longleaf pine has been to exam- ilies to plant in conditions that are subject to ine longleaf seedling establishment, survival, partial shade or neighbor influence. However, and growth in openings of various size and there is little basis for this at present. condition. Results are generally of the opinion that a clumped residual overstory condition, suggestive of the group selection approach, Pattern of Implementation promotes early seedling development when Because the presence of overstory trees affects compared with homogeneous overstory con- longleaf pine seedling establishment and de- ditions. Seedlings initiate height growth pri- velopment, most recent research has concen- marily in the center of gaps, but height growth trated on the possibilities associated with use is reduced along the borders of gaps or adja- of the group selection method in longleaf pine. cent to residual trees. This border effect is on Farrar (1996, 1998) suggests that a “modified the order of 12–20 m. The major reasons for group selection” approach be used in which this seedling growth pattern are related to in- groups are not removed until regeneration is creased light intensity in gaps (Grace and Platt established beneath the group. He further sug- 1995; Palik et al. 1997; McGuire et al. 2001; gests that this be integrated with cyclic pre- Battaglia et al. 2002; Gagnon et al. 2003) and scribed burning (Fig. 9). Either the VCGDL or decreased levels of intraspecific competition the BDq regulation approaches would be fea- for soil resources in gaps (Brockway and Out- sible (Farrar and Boyer 1991). Under the BDq, calt 1998). Farrar’s suggested target structure for longleaf 234 III. Silviculture

FIGURE 9. Longleaf pine seedlings and saplings of natural origin in a group selection opening on the Escambia Experimental Forest in Brewton, AL, during the 1982 growing season. (James M. Guldin) pine includes a residual basal area target B the groups are treated as small shelterwood of 14 m2 ha−1, maximum retained diameter openings. Smith (1986) noted that although D of 50 cm, and a q ratio for 5-cm diameter group selection is usually imposed by remov- classes of 1.44, and a suggested cutting cycle ing all trees within the group, groups could cer- length of 10 years (Farrar 1996). His uneven- tainly be created that retained overstory trees aged marking guidelines contain a description within them as silviculturally appropriate. This of how to implement either method for lon- method would resemble a group selection with gleaf pine. The burning program is required reserves (cf. Helms 1998) except that the re- to keep competing hardwoods in check and to serves are explicitly retained for the silvicul- keep seedbeds prepared for any seedfall that tural purpose of obtaining natural regenera- might occur. When seedlings become estab- tion. lished at acceptable densities within an area A shelterwood-based group selection ap- (local distributions equivalent to 1800–2000 proach in longleaf pine would use groups trees ha−1), cutting cycle harvest to remove the within which longleaf pine seed trees are re- overstory trees will allow seedlings to initiate tained at shelterwood (6–7 m2 ha−1) residual height growth. Subsequent cutting cycle har- basal area levels. During one cutting cycle harvests can be used to expand existing groups or vest, groups would be marked to resemble the to establish new groups, and as a free thinning seed cut of a shelterwood residual basal area in the matrix of the stand between the group (Smith 1986), using the same decision vari- openings. ables for seed tree retention as described in Given the success in regenerating longleaf Croker and Boyer (1975). During the inter- pine naturally using the shelterwood method, vening cutting cycle, prescribed fires would be another possibility, or perhaps a variant of imposed as an element of the prescription, so Farrar’s modified group approach, is to adapt as to prepare the site for seedfall, and Farrar the shelterwood prescription for longleaf pine (1998) offers suggestions to accomplish that. (Croker and Boyer 1975) in the context of a The seed trees in the group openings would group selection regeneration method, where eventually maximize their ability to produce 7. Uneven-Aged Silviculture 235 cones, seedfall would be optimized 3 to 5 years even if only a few large trees are retained, their after the cutting cycle harvest, and at the next presence can adversely affect the development adequate seed year seedlings would become of the stand, because retaining them prevents established in the shelterwood group opening. growing space from being used by trees of The residual seed trees within the group would other sizes. The regulation method and mark- then be removed in the subsequent cutting ing guides must account for any trees that are cycle harvest to release established seedlings. retained for special purposes, simply because If the removal cut of the shelterwood resid- large trees usurp considerable growing space uals within the group opening is not timely, that if unaccounted for could disrupt stand de- seedling survival would be compromised. velopment. Alternatively, if groups are initially created For example, the BDq method calls for har- without residual trees, an area adjacent to the vest of the worst trees and retention of the best group opening could be retained at shelter- in diameter classes at or below the maximum wood basal areas such that group expansion retained diameter D, but calls for all trees above would occur in the subsequent cutting cy- the D to be cut. There are no active research cle harvest, resulting in an expanded group studies on the Crossett EF that test whether coalescence. The same suggestions regarding trees larger than D can be retained. As a starting marking, use of prescribed fire, seedling devel- point, the basal area of retained trees should be opment, overstory removal, and so on would included in the calculations of stand structure, apply in this modification of the practice. simply because they create a very large influ- It might be that stands larger than 15 ha ence in basal area calculations. An 80-cm tree could be managed this way, and also that the has a basal area of 0.5 m2, or roughly 4% of method would be amenable to group open- the residual basal area target of 13 m2 ha−1 af- ings larger than the 0.81-ha maximum for ter a typical cutting-cycle harvest. If three trees group selection suggested by the USDA For- above D per hectare were retained for special est Service. Research would be needed to de- purposes and the stand below D was man- termine an effective range of group opening aged for the after-cut residual basal area tar- size such that longleaf seedlings can initiate get of 13 m2 ha−1, the actual basal area in the height growth in a short period of time. It is stand would be 14.5 m2 ha−1, more than 10% likely that the size would be larger rather than higher than the target. Over time that would smaller. This approach could be regulated us- adversely affect stand development. Similarly, ing any of the usual regulation methods that under volume control, the volume and the vol- apply to uneven-aged stands. ume growth of those big trees must be averaged into the calculation used to determine the allowable cut and the guiding diameter limit. Approach to Stand Regulation Trees above the limit can be retained at the dis- More than in other southern pines, managers cretion of the marker, provided that an equiva- in longleaf pine are looking for ways to re- lent volume is then marked below the guiding tain larger trees, in some cases much larger, diameter limit. above the diameter limits that have been established in other uneven-aged experience. This Growth and Yield would be done to meet resource attributes, values, or needs for associated species within Empirical data on growth and yield of uneven- the longleaf forest ecosystem, such as legacy aged stands of longleaf pine are difficult to trees or for nest construction by species such find because uneven-aged stands managed for as the red-cockaded woodpecker. There is little a sufficient length of time are relatively rare theory available in the uneven-aged literature (Kush et al. this volume). One of the few pa- to account for the influence of large trees re- pers to cite data on growth and yield directly tained above the maximum diameter or in ad- under the selection system is that of Farrar dition to the desired target residual stand. But and Boyer (1991), who describe the growth of 236 III. Silviculture two uneven-aged stands on the Escambia EF Developmental Dynamics over a 10-year period in comparison to that of a demonstration stand being managed using If uneven-aged silviculture is applied in lon- the shelterwood method. In all three stands, gleaf pine stands similar to that in other for- sawtimber volume growth was comparable at est types where the method has been success- roughly 2 m3 ha−1, or roughly two-thirds the ful, a number of attributes will be apparent. rate expected in the study. At these rates of Longleaf pine trees in the sawtimber size class volume growth, cutting cycles of 10 years or will constitute roughly two-thirds of the resid- longer will be required to generate operable ual basal area, but only 25% of the number of harvests. stems 10 cm and larger in the stand (Farrar et Farrar and Boyer (1991) also speculated that al. 1984; Baker et al. 1996). Cutting cycle har- over time, volume yield from selection stands vests will require two visits by field crews to will likely be less than that produced from a the stand—one to obtain the before-cut stand forest of large even-aged stands, because the inventory upon which regulation calculations zone of competition between large and small are based and to determine operability of the trees is minimized with large blocks found in proposed cutting cycle harvest, and a second even-aged stands (Farrar and Boyer 1991). The to actually mark the stand using the mark- use of group selection, which would also have ing guidelines that the regulation calculations less area within the stand in this zone of com- produce. Marking will follow a pattern of cut- petition relative to single-tree selection, might ting the worst longleaf pines, and leaving the partly compensate for that hypothesized vol- best, during every cutting cycle. Some form ume shortfall. of competition control that meets the multi- In a study of longleaf pine regeneration ple silvicultural objectives of site preparation, development under varying residual over- release, and liberation cutting will be required story basal area levels, Boyer (1993) reported on the order of every other cutting cycle. In that even a few residual longleaf pine par- longleaf pine stands, that will probably take ent trees resulted in substantial growth re- the form of prescribed fire, with perhaps oc- ductions of the new age cohort. The growth casional herbicide use or mechanical felling of of two-aged stands in this study was less competing hardwood species. Regeneration of than half the growth reported in the natu- longleaf pine must be monitored after each rally regenerated even-aged stands released cutting cycle harvest. Some regeneration will from overstory competition when young. This be expected to become established and to ini- provides another estimate of the total mer- tiate height growth after every cutting cycle chantable volume growth that might be pro- harvest, and especially after the first cutting duced in uneven-aged stands—less than half cycle harvests in stands recovering from un- that found in comparable released even-aged derstocked conditions or being converted from stands. even-aged fully stocked conditions. After sev- Given the shortage of long-term data on eral decades of implementation using a 10- growth and yield from uneven-aged stands in year cutting cycle, visual examination of stands the literature, among the first priorities for re- will reveal foliage of longleaf pine present at all search is to better quantify the growth and levels of the canopy profile from the ground yield that one can expect from application of up through the main canopy. It will require the selection system in the longleaf pine for- three or more 10-year cutting cycle harvests est type. At a minimum, one should expect to approach an uneven-aged structure, and reduced rates of volume growth, especially longer to develop a well-balanced stand struc- in total merchantable cubic volume, in the ture. Fewer cutting cycles will be needed if the uneven-aged stands. This point has been ob- initial stand condition is understocked and if served elsewhere in application of the method multiple age cohorts are already present; more in southern pines (Guldin and Baker 1998). will be needed if the stand under management 7. Uneven-Aged Silviculture 237

FIGURE 10. A within-stand view of a new regeneration cohort in a longleaf pine stand managed as a quail plantation in southern Georgia. (James M. Guldin) was initially a well-stocked even-aged stand at maintain low residual basal area at 15 m2 ha−1 midrotation or later. or less; open midstory conditions are main- This description is at odds with several ex- tained, removals are made from below, and isting and notable approaches to management regeneration occurs in patches. The purpose of of longleaf pine—that proposed for manage- this variant of the single-tree selection method ment of the red-cockaded woodpecker, and is to provide habitat for northern bobwhite that reflected in the Stoddard–Neel approach quail. Graphs of the diameter distributions for for management of uneven-aged longleaf pine several quail plantations (Moser at al. 2002) stands for quail habitat. Open understory differ from the reverse J-shaped curve typically conditions are sought for both red-cockaded associated with uneven-aged silviculture at the woodpeckers and for bobwhite quail. An un- stand level, specifically in a lack of trees in derstory full of trees in the submerchantable the smallest diameter classes. For example, no and smallest merchantable size classes are 1-cm diameter class less than 10 cm in any of sought in an uneven-aged stand. The ques- the plantations managed using the Stoddard– tion is whether uneven-aged practices can Neel approach has more than 5% of the to- be, or have been, modified so as to simulta- tal number of pines per hectare. Conversely, neously maintain an open understory condi- the 1995 preharvest inventory on the Good tion, while concurrently maintaining the de- Farm Forestry Forty demonstration stand on velopment of regeneration cohorts in sufficient the Crossett EF showed 4450 trees ha−1 in number and distribution within the uneven- the submerchantable diameter classes (1.5–8.9 aged stand (Fig. 10). centimeters inclusive), corresponding to 92% Consider the Stoddard–Neel approach of of the total trees per hectare in the demon- uneven-aged silviculture, an excellent descrip- stration stand (Guldin unpublished data). This tion of which was recently published by Moser raises the question of whether regeneration et al. (2002) (see also boxes A and B). In is being recruited at a sufficient rate under this approach, single-tree selection is used to the Stoddard–Neel approach to ensure the 238 III. Silviculture long-term sustainability of not only uneven- The favorable elements will be useful in devel- aged stand structure, but also the habitat opment or refinement of the selection method values for which those stands are justifiably in longleaf pine. First, dominant or codomi- prized. nant longleaf pines respond to release, though Similarly, debate is currently active in the suppressed trees do not (Boyer 1990). Thus, literature about the habitat values provided cutting cycle harvests in which codominant or by uneven-aged stands for the red-cockaded better longleaf pines are released from com- woodpecker (Engstrom et al. 1996; Rudolph petition of others will stimulate a growth re- and Conner 1996; Hedrick et al. 1998). This de- sponse in the residual stand, which promotes bate has its roots within the application of the continued stand development. This attribute Stoddard–Neel selection system as well, since it has been a feature of the Stoddard–Neel vari- is largely upon that experience that habitat de- ant of the selection method in longleaf pine scriptions of uneven-aged longleaf pine stands as well (Moser et al. 2002). Second, it can be are based. But the debate embraces other sit- successfully managed using the shelterwood uations in which desired habitat for the red- method (Croker and Boyer 1975), which has cockaded woodpecker is inconsistent with in- been observed in other forest types where the formation available from the literature about selection method has been applied (Guldin uneven-aged stand dynamics. It is not really a 1996). question of the conditions that are appropriate Where longleaf differs most prominently for the red-cockaded woodpecker, which are from other southern pines is in the periodicity fairly well defined (Conner and Locke 1982; of seed crops and the difficulty in securing nat- Hooper 1988; Hooper et al. 1991; Conner et ural regeneration. This will require new inter- al. 1994; Ross et al. 1997). Rather, the ques- pretations of existing knowledge, and the de- tion is one in which a regeneration method velopment of new knowledge, to ensure that that provides desired residual stand conditions longleaf pine seedlings can become established is sustainable, according to the first rule of sus- and can develop properly following regen- tainability at the stand level—that an imposed eration cutting under the selection method. regeneration method must result in the suc- With respect to natural regeneration, refine- cessful establishment and development of re- ments of existing knowledge from the appli- generation. cation of the shelterwood method might be promising as a variation under modifications of the group selection method (Farrar 1996). Conversely, should natural regeneration tech- Summary niques fail or result in unacceptable delays in regeneration establishment or development, The successful practice of uneven–aged silvi- technology should be developed for applica- culture in mixed loblolly–shortleaf pine stands tion of planting as an alternative or a preferred of the upper West Gulf Coastal Plain has method of obtaining establishment of regen- been characterized by a number of attributes eration at acceptable levels. More than any (Guldin and Baker 1998). Key factors are at- other southern pine, or for that matter any tention to stand-level regulation, use of appro- other species in which the selection method priate residual basal area levels that approx- has been used, planting seedlings for reforesta- imate those found in slightly understocked tion of uneven-aged stands will have a promi- mature even-aged stands, establishment and nent place in the successful application of the development of regeneration, and attention selection method for longleaf pine. to a marking rule that cuts the poorest trees The biggest question that remains unre- and leaves the best across a range of diameter solved is the level at which regeneration de- classes. velopment can be considered acceptable in Longleaf pine shares some silvical attributes the selection method. The Stoddard–Neel se- with loblolly and shortleaf pines, but not all. lection method points in one direction about 7. Uneven-Aged Silviculture 239 this, and experience from the selection method Brockway, D.G., and Outcalt, K.W. 1998. Gap-phase as practiced in mixed loblolly-shortleaf pines, regeneration in longleaf pine wiregrass ecosys- and in other forest types, points in another. tems. For. Ecol Manage 106(2,3):125–139. The Stoddard–Neel approach differs from most Cain, M.D., and Shelton, M.G. 2001. Natural other instances of successful application of loblolly and shortleaf pine productivity through uneven-aged silviculture due to the smaller 53 years of management under four reproduction cutting methods. South J Appl For 25(1):7–16. number of stems in the submerchantable class. Conner, R.N., and Locke, B.A. 1982. Fungi and red- That difference leads to allied relationships in cockaded woodpecker cavity trees. Wilson Bull habitat condition that promotes open midstory 94:64–70. conditions in one case, and midstory condi- Conner, R.N., Rudolph, D.C., Saenz, D., and Schae- tions occluded by development of seedlings fer, R.R. 1994. Heartwood, sapwood, and fungal and sapling in the submerchantable diame- decay associated with red-cockaded woodpecker ter classes in other conditions. Ultimately, the cavity trees. J Wildl Manage 58:728–734. question relates to the degree to which de- Croker, T.C. 1964. Fruitfulness of longleaf trees viations from the reverse J-shaped structure more important than culture in cone yield. J For can be considered to be sustainable at the 65: 488. stand level. This is the most prominent re- Croker, T.C. 1973. Longleaf pine cone production in relation to site index, stand age, and stand density. search gap in our understanding of the regen- Research Note SO-156. New Orleans, LA: US DA eration dynamics of uneven-aged longleaf pine Forest Service, Southern Forest Experiment Sta- stands, and one that is critical in order to ulti- tion. mately evaluate what constitutes sustainabil- Croker, T.C., and Boyer, W.D. 1975. Regenerating ity of uneven-aged structure in longleaf pine longleaf pine naturally. Research Paper SO–105. stands in the long term. New Orleans, LA: USDA Forest Service, Southern Forest Experiment Station. deLiocourt, F. 1898. De l’amenagement des Sap- References inieres.` Bulletin of the Society of Foresters. Franche-Comte´ Belfort, Besançon, pp. 396–409. Baker, J.B., and Murphy, P.A. 1982. Growth and Engstrom, R.T., Brennan, L.A., Neel, W.L., Farrar, yield following four reproduction cutting meth- R.M., Lindeman, S.T., Moser, W.K., and Her- ods in loblolly–shortleaf pine stands—A case mann, S.M. 1996. Silvicultural practices and red- study. South J Appl For 6(2):66–74. cockaded woodpecker management: A reply to Baker, J.B., Cain, M.D., Guldin, J.M., Murphy, P.A., Rudolph and Conner. Wildl Soc B 24(3):334–338. and Shelton, M.G. 1996. Uneven-aged silvicul- Farrar, R.M., Jr. 1996. Fundamentals of uneven- ture for the loblolly and shortleaf pine forest cover aged management in southern pine. Miscella- types. USDA Forest Service, General Technical neous Publication 9. Tallahassee, FL: Tall Timbers Report SO–118. New Orleans, LA: Southern Re- Research Station. search Station. Farrar, R.M. 1998. Prescribed burning in selection Barnett, J.P. 1999. Longleaf pine ecosystem restora- stands of southern pine: Current practice and tion: The role of fire. J Sustain For 9(1/2):89–96. future promise. In Fire in Ecosystem Management: Battaglia, M.A., Mou, P., Palik, B., and Mitchell, Shifting the Paradigm from Suppression to Prescription, R.J. 2002. The effect of spatially variable over- eds. T.L. Pruden and L. A. Brennan, pp. 151–160. story on the understory light environment of an Tall Timbers Fire Ecology Conference Proceedings open-canopied longleaf pine forest. Can J For Res No. 20. Tallahassee, FL: Tall Timbers Research Sta- 32(11):1984–1991. tion. Boyer, W.B. 1990. Longleaf pine. In Silvics of North Farrar, R.M., Jr. and Boyer, W.D. 1991. Manag- America: 1. Conifers; 2. Hardwoods. Agriculture ing longleaf pine under the selection system— Handbook 654, tech. cords. R.M. Burns and B.H. Promises and problems. In Proceedings of the 6th Honkala, vol. 1, pp. 405–412. Washington, DC: Biennial Southern Silvicultural Research Conference; USDA Forest Service. 1990 October 30–November 1, Memphis, TN, pp. 357– Boyer, W.B. 1993. Long-term development of re- 368. General Technical Report SE–70. Asheville, generation under longleaf pine seedtree and shel- NC: USDA Forest Service, Southeastern Forest terwood stands. South J Appl For 17(1):10–15. Experiment Station. 240 III. Silviculture

Farrar, R.M., Murphy, P.A., and Willett, R. L. 1984. McConnell, W.N. 2002. Initiating uneven-aged Tables for estimating growth and yield of uneven- management in longleaf pine stands: Impacts on aged stands of loblolly–shortleaf pine on average red-cockaded woodpecker habitat. Wildl Soc B sites in the West Gulf area. Bulletin 874. Fayet- 30(4):1276–1280. teville: University of Arkansas, Arkansas Agricul- McGuire, J.P., Mitchell, R.J., Moser, E.B., Pecot, tural Experiment Station. S.D., Gjerstad, D.H., and Hedman, C.W. 2001. Gagnon, J.L., Jokela, E.J., Moser, W.K., and Huber, Gaps in a gappy forest: Plant resources, longleaf D.A. 2003. Dynamics of artiﬁcial regeneration in pine regeneration, and understory response to gaps within a longleaf pine ﬂatwoods ecosystem. tree removal in longleaf pine savannas. Can J For For Ecol Manage 172:133–144. Res 31(5):765–778. Grace, S.L., and Platt, W.J. 1995. Neighborhood Marquis, D.A. 1978. Application of uneven-aged sil- effects on juveniles in an old-growth stand of viculture on public and private lands. In Uneven- longleaf pine, Pinus palustris. Oikos 72(1):99– Aged Silviculture and Management in the United 105. States, pp. 25–61. USDA Forest Service. Timber Guldin, J.M. 1996. The role of uneven-aged silvi- Management Research, General Technical Report culture in the context of ecosystem management. WO-24. West J Appl For 11(1):4–12. Meyer, H.A., Recknagel, A.B., Stevenson, D.D., and Guldin, J.M. 2002. Continuous cover forestry in the Bartoo, R.A. 1961. Forest Management, 2nd edi- United States: Experience with southern pines. tion. New York: Ronald Press. In Continuous Cover Forestry: Assessment, Analysis, Moser, W.K., Jackson, S.M., Podrazsky, V., and Scenarios, eds. K. von Gadow, J. Nagel, and J. Larsen, D.R . 2002. Examination of stand struc- Saborowski, pp. 295–307. Boston: Kluwer Aca- ture on quail plantations in the Red Hills region demic Publishers. of Georgia and Florida managed by the Stoddard– Guldin, J.M., and Baker, J.B. 1988. Yield compar- Neel system: An example for forest managers. isons from even-aged and uneven-aged loblolly– Forestry 75(4):443–449. shortleaf pine stands. South J Appl For 12:107– O’Hara, K.L. 1996. Dynamics and stocking-level re- 114. lationships of multi-aged ponderosa pine stands. Guldin, J.M., and Baker, J.B. 1998. Uneven- For Sci 42(4), Monograph 33. aged silviculture, southern style. J For 96(7):22– Palik, B.J., Mitchell, R.J., Houseal, G., and Ped- 26. erson, N. 1997. Effect of canopy structure on Guldin, J.M., and Fitzpatrick, M.W. 1991. Compar- resource availability and seedling response in a ison of log quality from even-aged and uneven- longleaf pine ecosystem. Can J For Res 27(9):1458– aged loblolly pine stands in south Arkansas. South 1464. J Appl For 15(1):10–17. Pickett, S.T.A., and White, P.S., eds. 1985. The Ecol- Hedrick, L.D., Hooper, R. C., Krusac, D. L., and Dab- ogy of Natural Disturbance and Patch Dynamics. New ney, J. M. 1998. Silvicultural systems and red- York: Academic Press. cockaded woodpecker management: Another Reynolds, R.R. 1959. Eighteen years of selection perspective. Wildl Soc B 26(1):138–147. timber management on the Crossett Experimen- Helms, J.A., ed. 1998. The Dictionary of Forestry. tal Forest. USDA Technical Bulletin 1206. Bethesda, MD: The Society of American Foresters. Reynolds, R.R. 1969. Twenty-nine years of selection Hooper, R.G. 1988. Longleaf pines used for cavi- timber management on the Crossett Experimen- ties by red-cockaded woodpeckers. J Wildl Manage tal Forest. USDA Forest Service Research Paper 52:392–398. SO-40. Hooper, R.G., Krusac, D.L., and Carlson, D.L. 1991. Reynolds, R.R., Baker, J.B., and Ku, T.T. 1984. Four An increase in a population of red-cockaded decades of selection management on the Crossett woodpeckers. Wildl Soc B 19:277–286. Farm Forestry Forties. Arkansas Agricultural Ex- Landers, J.L., Van Lear, D.H., and Boyer, W.D. 1995. periment Station Bulletin 872. The longleaf pine forests of the Southeast: Re- Ross, W.G., Kuhlavy, D.L., and Conner, R.N. 1997. quiem or renaissance? J For 93(11):39–44. Stand conditions and tree characteristics affect Leak, W.B. 1964. An expression of diameter distri- quality of longleaf pine for red-cockaded wood- bution for unbalanced uneven-aged forests. For pecker cavity trees. For Ecol Manage 91:145– Sci 10:39–50. 154. Long, J.N. 1998. Multiaged systems in the central Rudolph, D.C., and Conner, R.N. 1996. Red- and southern Rockies. J For 96(7):34–36. cockaded woodpeckers and silvicultural practice: 7. Uneven-Aged Silviculture 241

Is uneven-aged silviculture preferable to even- Wahlenberg, W.G. 1946. Longleaf Pine: Its Use, Ecol- aged? Wildl Soc B 24(2):330–333. ogy, Regeneration, Protection, Growth, and Manage- Runkle, J.R. 1982. Patterns of disturbance in some ment. Washington, DC: Charles Lathrop Pack old-growth mesic forests of eastern North Amer- Forestry Foundation and USDA Forest Service. ica. Ecology 63:1533–1546. White, P.S. 1979. Pattern, process, and natural dis- Smith, D.M. 1986. The Practice of Silviculture, 8th edi- turbance in vegetation. Bot Rev 45:229–299. tion. New York: John Wiley & Sons. Zeide, B., and Sharer, D. 2000. Good forestry at a Troup, R.S. 1928. Silvicultural Systems. London: glance: A guide for managing even-aged loblolly Oxford at the Clarendon Press. pine stands. Arkansas Forest Resources Center Troup, R.S. 1952. Silvicultural Systems, 2nd edition. Series 003. University of Arkansas Division of Edited by E. W. Jones. London: Oxford University Agriculture, Arkansas Agricultural Experiment Press. Station. 242 III. Silviculture

BOX 7.1 The focus of the S-N approach is on longleaf pine, but it is not exclusive to this species The Stoddard–Neel and is applicable to more than just mature forests. Approach The S-N approach is unique not because A Conservation-Oriented of the uneven-aged focus or in how trees are selected for removal, but rather in its Approach overall guiding philosophy of resource man- Steven B. Jack, W. Leon Neel, agement and what is selected to remain on and Robert J. Mitchell the site after each harvest entry. The S-N Joseph W. Jones Ecological Research Center, approach is grounded in a strong land ethic, Newton, Georgia 39870 with the maintenance of a healthy and multifaceted forest as the major goal. Rather Uneven-aged systems of forest manage- than a primary focus on timber management have been recommended as good ment with other resources and amenities conservation-oriented resource manage- as secondary or ancillary, the S-N approach ment approaches. A multi-aged approach is seeks to preserve all characteristics of the thought to be especially appropriate for lon- ecosystem and then determine how much gleaf pine ecosystems managed for diversity timber is available for harvest. The result is and conservation purposes because it is be- that no one value of the ecosystem is max- lieved that historical longleaf pine forests imized at the expense of other amenities, contained multiple age classes with many but rather a balance of ecological, economic, cohorts older than is typically found in to- and aesthetic values is maintained. day’s landscape. Under the S-N approach the forest is Herbert Stoddard and Leon Neel devel- never terminated, and continual forest oped, over the course of several decades, cover is maintained with multiple regen- a unique approach (hereafter referred to eration events that lead to a multi-aged as the Stoddard–Neel or S-N approach) to structure. The S-N approach also incorpo- uneven-aged management of open pine– rates a long-term view to forest manage- grassland ecosystems on the Coastal Plain ment rather than a short-term, economi- region of north Florida and south Georgia. cally driven model for forestry, with most Their approach was developed on private results realized over many years from cu- hunting preserves in the region, where aes- mulative effects rather than from one or thetics and wildlife considerations were as two discrete management actions. The S- important, or even more important, than N approach is also inherently conservative, the production of timber volume. The over- where the capital in the form of stand- all objective of the S-N approach was and ing volume is protected and only a por- is to restore and perpetuate the forest and tion of the growth is harvested in any given all its components, while at the same time stand entry. Natural disturbance events in meeting individual land owner objectives, longleaf pine forests, which tend to be with the production of timber and subse- small in extent and of low intensity, are quent economic return as by-products of used as a guide to harvest operations. Fi- this focus on long-term stewardship. Prop- nally, the S-N approach embraces and cre- erties managed under their guidance have ates complexity and diversity in the forest produced substantial timber volumes over rather than trying to simplify forest struc- many years while perpetuating the longleaf ture or homogenizing structure over the pine ecosystem and its associated diversity. landscape. 7. Uneven-Aged Silviculture 243

These guiding principles of the S-N ap- for regeneration rather than targeting some proach lead to several attributes in imple- percentage of the forest to be in evenly dis- mentation. First, prescribed fire is crucial tributed gaps. While no elaborate calcula- and integral to the S-N approach within the tions are carried out to determine annual pine–grassland ecosystem. Fire addresses yield, periodic inventories are conducted to several objectives: (1) it controls hardwood guide the determination of stand growth encroachment and maintains the open mid- and how much volume is available to be story characteristic of the forest type; (2) harvested at any point in time. Maximiz- it helps propagate fire-adapted and fire- ing the growth of individual trees is not a dependent species of plants as well as ap- priority of the S-N approach, but individual propriate habitat for wildlife native to the tree quality is an important consideration ecosystem; and (3) over time it helps to and is one of the factors that help drive the define the upland areas for pine manage- economics of the approach. ment. In the S-N approach prescribed fire is Regeneration, especially of longleaf pine, objective driven and frequent such that all is of high concern in the S-N approach, but seasons for burning are potentially used to in comparison to other uneven-aged ap- meet objectives, and variability in fire effects proaches much less emphasis is placed on are achieved as objectives change over time. obtaining regeneration at each stand entry. Because prescribed fire is the primary man- This is due to two primary factors: (1) the agement tool utilized in the approach, silvi- long time scale of management planning cultural operations are carried out to ensure under the S-N approach and the longevity the presence of adequate fuels (e.g., pine of longleaf pine result in less need to obtain needles) to conduct frequent burns. regeneration establishment as frequently as Selection of trees for harvest is similar is necessary when managing for fixed maxi- to that found in other uneven-aged ap- mum age class, tree size, or a specified diam- proaches to forest management (i.e., “cut eter distribution; and (2) the conservative the worst and leave the best” with trees timber marking and less emphasis on pro- marked for removal based upon their vigor ducing timber means that there is a reduced and growth potential) but tends to favor requirement for continual recruitment to older and larger trees as residuals much replace large numbers of harvested trees. more than other approaches. With the S-N The result is that regeneration tends to occur approach a mix of age classes is maintained in small, discrete patches rather than being in the forest, including a substantial comple- fairly uniformly distributed throughout the ment of older trees beyond their “financial forest, and the understory and midstory ap- maturity.” Also, all trees with visible defect pear much more open than is true for many are not marked for removal because having uneven-aged forests. some trees with economic defect provides Because trees are maintained in the for- unique wildlife habitat and adds diversity est for a long time and residual trees to the forest. are carefully selected, harvesting opera- The S-N approach is not quantitative nor tions are carried out with great care and easily quantified. No uniform target stand close monitoring (as is the case with structure is specified to guide the timber all approaches to uneven-aged manage- marking process as is characteristic of most ment) in order to minimize damage to other uneven-aged approaches, with the re- regeneration, residual adult trees, ground- sult that the stand density and basal area cover, and soil. The S-N approach differs are highly variable in a forest managed from many other approaches, however, in under the S-N approach. Natural gaps are the high level of utilization of cut stems. utilized and expanded to provide “space” Intensive merchandizing of cut trees is 244 III. Silviculture

FIGURE 1. Data from Greenwood Plantation located near Thomasville, GA, showing results of management using Stoddard–Neel approach. (a) Total, standing, and harvested volume (for entire property) over 5 decades; (b) standing and harvested per hectare volumes by product class. Data are from periodic (every 10 years) timber inventories from 1945 to 1995 and from timber harvest records for same time period. Original volume data were in board feet (Scribner) per acre and were converted using the factors 1 m3 = 450 bf (green) and 1 bf/A = 0.00549 m3/ha.

important to implementing S-N approach varied structure of the forest also allow for because it helps to support the conser- greater ﬂexibility in responding to changing vative marking of trees. That is, maxi- timber markets. mizing the value of trees at the landing The S-N approach has been questioned by careful allocation to different product and criticized in regard to its long-term classes can greatly increase the return to the sustainability, and has been characterized landowner and offset the lower harvest vol- by some as only tending large trees with umes. The careful merchandizing and the no consideration for regeneration so that 7. Uneven-Aged Silviculture 245

the system will eventually collapse as all the evidence that the S-N approach is able to old trees either die or are harvested (i.e., achieve significant conservation objectives mining the limited resource). However, an and maintain and perpetuate the longleaf example from Greenwood Plantation near pine forest with a range of amenities while Thomasville, GA, managed using the S-N still producing substantial income from tim- approach for more than 60 years shows ber harvest. that the forest can be sustained and im- The S-N approach cannot be fully ex- proved utilizing this form of management plained in this overview because it relies (see box by Moser for other examples). Box heavily on the “art” aspects of silviculture A Figure 1 shows that over 50 years the (as opposed to “science” and quantifica- total standing volume on the plantation in- tion) and there is no numerical equation creased (in total and per acre) while still or “recipe” that can be distilled for direct supporting significant timber harvests. In application in the field. The S-N approach addition, other data (not shown) indicate is a product of the time and place where it that there is continual ingrowth to larger was developed, and is not appropriate for all tree sizes from regeneration and recruit- forests or landowner objectives. However, ment. Even though high timber volumes when conservation goals are paramount but were harvested from the property it did not some utilization and income are required, come at the expense of other resources: the S-N approach to forest management is excellent quail hunting was maintained, an excellent alternative to more traditional as were the diversity of the understory or prevalent uneven-aged silvicultural ap- flora and a large red-cockaded woodpecker proaches, particularly in pine-grasslands of population. These results provide strong the Southeast. 246 III. Silviculture

BOX 7.2 ecological attributes that make longleaf pine ecosystems important. Overall diversity is The Stoddard–Neel reduced by competition from a dense overstory and midstory, a necessary result of System a constant q-factor (Farrar 1996) in an Case Studies uneven-aged structure. The Stoddard–Neel system of uneven- W. Keith Moser aged management, practiced in the Red Hills USDA, Forest Service, North Central Research region of Florida and Georgia, uses single- Station, Forest Inventory and Analysis, St. Paul, tree selection to create a multi-aged for- Minnesota 55108. est that provides an open midstory and a great deal of growing space to the under- In the eighteenth century, the Coastal Plain story, thereby maintaining suitable habi- uplands of what is now the southeast- tat for species of interest, such as northern United States were covered with large, ern bobwhite quail, gopher tortoise, and variably stocked longleaf pine (Pinus palus- red-cockaded woodpecker (Engstrom et al., tris Mill.) stands. Hurricanes and tornadoes 1995) (see the box by Jack et al. for de- caused many areas to develop as even-aged tails of this approach). In this system, large stands. Smaller scale disturbances, such as old pines are considered valuable both for lightning strikes and fires of varying inten- their influence on the rest of the ecosys- sity, caused stands to develop diverse age- tem and as habitat for species of special con- and size-class structures (Frost 1993). With cern. Trees are moved through the midstory fires occurring every 1–10 years, longleaf in patches as opposed to trying to main- pine was well-suited to multi-aged forest tain a widespread midstory. Additionally, in conditions because it can respond to release the context of maintaining healthy quail at almost any age and because it tolerated habitat, basal areas were largely kept below − fire at all age classes better than its pine and 15 m2 ha 1 to ensure enough growing space hardwood competitors can (Boyer 1990), for the understory plants that provide food although fire will always impact growth for the bobwhite (Moser and Palmer 1997). (Boyer 2000). Longleaf pine also has some Starting in the 1920s, Herbert Stoddard reputation for insect and disease resistance determined the connection between quail (Moser et al., 2003), further limiting the op- population dynamics, prescribed fire, and portunity for widespread even-aged forests overstory structure (Stoddard 1931). Dur- caused by insect and disease attacks. ing World War II, Stoddard devised the As described in the accompanying chapter “Stoddard system” of selection system man- by Guldin, quantifying multi-aged forests agement to supply wood for the war ef- has assumed that a balanced forest has a fort while still maintaining a forest over- reverse J-shaped diameter distribution that story. Beginning in 1950, Leon Neel started can be described by a constant ratio of suc- working with Stoddard and continued ceeding diameter classes. While some may his management methods after Stoddard argue that the shape of the curve refers retired. to the diameter distribution of forests, not I present below the data from three stud- stands (de Liocourt 1898), many foresters ies. A 4-ha block of the Wade Tract, referred have sought to implement this structure at to as “Area WT” below, an 80-ha ecological the smallest scale possible, even in shade- preserve where no harvesting has occurred, intolerant species. provides a template of an unmanaged diam- Attempts to maximize fiber production eter distribution. (Originally collected by W. with such a structure can devalue the very J. Platt, reported in several articles including 7. Uneven-Aged Silviculture 247

FIGURE 1. Diameter distribu- 0.6 tion of plantations. Plantation BL 0.5 Plantation DK Plantation SG Plantation MR Plantation SW 0.4 Plantation MP Area WT 0.3

0.2

0.1 Percentage of Total Number Trees 0.0 0 20 40 60 80 100 DBH (cm)

Platt et al., 1988; the data were excerpted point. Next there is a “bump” or second from Cressie 1991.) In a second (pilot) mode in the larger diameter classes, after study, one section of a quail shooting plan- which the number of trees again decreases. tation was sampled for species, heights, and The open midstory is the result of fire and diameters using 0.04-ha plots at an inten- mechanical control of sapling ingrowth and sity of 10% (“Plantation MP,”Moser unpub- a preference of the Stoddard–Neel system lished data). In the third study, the struc- for retaining large trees. The shape of the ture of several Stoddard–Neel forests was diameter distribution of the plantations dif- examined over 3 years (“Plantations MR, fers from the shape of the distribution of SW, SG, BL, and DK”). This last investiga- the classic reverse-J curve. Given the quail- tion (Moser et al., 2002) looked at diame- hunting constraints of a low residual basal ter distribution and spatial arrangement at area and an open midstory, the Stoddard– four scales: ownerships, 40-ha blocks, tran- Neel system is a more appropriate structure sects (7–8 per block) and 0.04- or 0.08-ha in these longleaf pine stands. Plantations plots (about 7 to 8 of these per transect). MR and BL are the exceptions that validate The researchers took the forestwide diam- the rule: they have a higher proportion of eter distributions and converted them into loblolly pine, so maintaining the smaller size proportions for each diameter class. They classes is more difficult with an annual fire compared the diameter distributions at the regime. ownership level to the distributions of the Moser et al. (2002) then compared the re- constituent blocks, the diameter distribu- lationship between size classes: forestwide tions of the blocks to the constituent tran- versus plantation, plantation versus block, sects, and the distributions at the transect block versus transect, and transect versus level to those of the constituent plots. plot. They found little significant difference Box B Figure 1 shows a summary of between the forestwide, plantation, or block the plantation-level diameter distributions distributions. Of the 404 plots compared from all three studies. Under the Stoddard– with their parent transects, 161 were dif- Neel system, the number of trees decreases ferent, approximately 40%. There is greater rapidly as the diameter increases, up to a variation from the forestwide distribution at 248 III. Silviculture

the transect and plot levels than at the block Stoddard–Neel structure with a “low” thin- level, and a more distinctly two-storied (bi- ning (thinning the overstory from below). modal) structure (Moser et al., 2002). Longleaf pine crowns were not very dense At the plot level, the higher propor- (Gagnon et al., 2004), so the combination tion of plots different from their parent of large and frequent sunflecks through the transects reflects the preeminence of site- canopy and the openings created by tim- specific and tree-specific decisions. As one ber harvest or quail management would goes down in scale, the percentage of man- release growing space to the understory. agement units with distributions signifi- The value of the Stoddard–Neel system is cantly different from the forestwide distri- that it maintains a continuous forest canopy bution increases. These data suggest that while maintaining an open midstory. A for- the Stoddard–Neel system focuses on man- est maintained under the BDq system with aging plantations and blocks, not stands or a constant ratio between successive diam- smaller units, as uneven-aged entities. The eter classes would require a very flat dis- two-storied nature of the diameter distri- tribution (a very low “q”) to satisfy these butions at all scales suggests that a forester requirements, and would run the risk of in- could manage the smaller units as two-aged sufficient ingrowth at the upper diameter (really two-sized) stands. levels due to mortality losses by the smaller While the practitioners of the Stoddard– trees, particularly under an intensive pre- Neel system have stated that they use scribed burning regime. Longleaf pine is a single-tree selection to create and main- long-lived species, allowing the forest man- tain such forests, like most foresters, they ager to use the Stoddard–Neel system to incorporated the practice of negative (im- maintain a vigorous overstory while “filter- provement) thinning—focusing on removing” trees through the midstory at a den- ing poorer quality trees to increase the qual- sity and rate compatible with maintaining ity of the residual stand. Neel stated that the the hunting and ecological attributes of the forest grows at a rate of 4% per year (Leon Red Hills plantations; often cover blocks Neel personal communication) and that ap- or “ring-arounds,” areas left unburned to proximately 90% of annual growth can be provide nesting and cover habitat for quail harvested in periodic removals (Neel 1993). (Moser and Palmer 1997), serve the addi- What is unique about the system, however, tional function of providing regeneration is the emphasis on retaining large, old trees. opportunities for longleaf pine. Like any As there is only a fixed amount of grow- partial harvesting practice, the Stoddard– ing space, other parts of the age (size) class Neel system requires management disci- distributions must occupy a proportionately pline, in that one cannot remove overstory smaller area. volume at a rate greater than ingrowth. A Before Stoddard’s investigations in the suitable quantity of overstory is necessary 1920s and 1930s, landowners were un- to provide habitat and fuel for prescribed aware of the relationship between over- burning. story structure, fire, and quail; so plan- The key points of the system include tations accumulated volume and new age maintaining densities below 15 m2 ha−1, classes of trees. One can assume that the di- managing the overstory with removals from ameter (and perhaps age) distribution was below, maintaining a reproduction compo- normal at the inception of the Stoddard– nent in the stand, and allowing transition Neel system in the 1940s. A forest man- from reproduction to overstory on some ager could then create and maintain the small proportion of the area. 7. Uneven-Aged Silviculture 249

References ogy Conference, No. 18, the Longleaf pine ecosystem: Ecology, restoration and management, ed. S.M. Boyer, W.D. 1990. Pinus palustris Mill. Longleaf Hermann, pp. 17–43. Tallahassee, FL: Tall Tim- pine. In Silvics of North America. Vol. 1: Conifers, bers Research Station. comps. R.M. Burns and B.H. Honkala, pp. 405– Gagnon, J.L., Jokela, E.J., Moser, W.K., and Hu- 412. USDA Agriculture Handbook 654. ber, D.A. 2004. Characteristics of gaps and Boyer, W.D. 2000. Long-term effects of biennial natural regeneration in mature longleaf pine prescribed fires on the growth of longleaf pine. flatwoods ecosystems. For Ecol Manage 187(2– In Proceedings of the Tall Timbers Fire Ecology Con- 3):373–380. ference, No. 21, Fire and forest ecology: Innovative Moser, W.K., and Palmer, W.E. 1997. Quail habi- silviculture and vegetation management, eds. W.K. tat and forest management: What are the Moser and C.F. Moser, pp. 18–21. Tallahassee, opportunities? In Forest Landowner Magazine FL: Tall Timbers Research Station. Annual Landowners Manual, March/April, 56– Christensen, N.L. 2000. Vegetation of the south- 63. eastern Coastal Plain. In North American Ter- Moser, W.K., Jackson, S.M., Podrazsk´ y,´ V.V., and restrial Vegetation, 2nd edition, eds. M.G. Bar- Larsen, D. 2002. Examination of stand struc- bour and W.D. Billings, pp. 397–448. London: ture on quail plantations in the Red Hills re- Cambridge University Press. gion of Georgia and Florida managed by the Cressie, N.A. 1991. Statistics for Spatial Data. New Stoddard–Neel system: An example for forest York: Wiley. managers. Forestry 75(4):443–449. de Liocourt, F. 1898. De l’amenagement des Sap- Moser, W.K., Treiman, T., and Johnson, E.E. inieres. Bulletin of the Society of Foresters. 2003. Species choice and risk in forest restora- Franche-Comte Belfort, Besançon, pp. 396– tion: A comparison of methodologies for eval- 409. uating forest management decisions. Forestry Engstrom, R.T., Brennan, L.A., Neel, W.L., 76(2):137–147. Farrar, R.M., Lindeman, S.T., Moser, W.K., and Neel, W.L. 1993. An ecological approach to lon- Hermann, S.M. 1995. Silvicultural practices gleaf forestry. In Proceedings of the Tall Timbers and red-cockaded woodpecker management: Fire Ecology Conference, No. 18, the Longleaf pine A reply to Rudolph and Conner. Wildl Soc Bull ecosystem: Ecology, restoration and management, 24(2): 334–338. ed. S.M. Hermann, p. 335. Tallahassee, FL: Tall Farrar, R.M. 1996. Fundamentals of Uneven-aged Timbers Research Station. Management in Southern Pine. Tall Timbers Re- Platt, W.J., Evans, G.W., and Rathbun, S. L. search Station, Tallahassee, FL. Miscellaneous 1988. The population dynamics of a long- Publication No. 9. lived conifer (Pinus palustris). Am Nat 131: 491– Frost, C.C. 1993. Four centuries of changing 525. landscape patterns in the longleaf pine ecosys- Stoddard, H.L. 1931. The Bobwhite Quail: Its Habits, tem. In Proceedings of the Tall Timbers Fire Ecol- Preservation and Increase. New York: Scribner. Chapter 8 Longleaf Pine Growth and Yield

John S. Kush, J. C. G. Goelz, Richard A. Williams, Douglas R. Carter, and Peter E. Linehan

Introduction pine. Some limited data are available for projecting natural stands of longleaf that may Across the historical range of longleaf pine be extrapolated to yield estimates of potential (Pinus palustris Mill.), less than 10% of lands growth in planted stands, but there is a great previously occupied by longleaf ecosystems deal of uncertainty when gains in seedling are currently in public ownership (Johnson quality, competition control, fertilization, and and Gjerstad 1999; Alavalapati et al., this vol- other silvicultural techniques are factored in. ume). The remainder is owned by private en- Much of the reestablishment of longleaf pine tities ranging from the forest industry, to tim- taking place today is occurring on old ﬁelds berland investment organizations, to highly and pastures. At least half of that planted is varied nonindustrial private landowners. Any done so using containerized seedlings, usu- signiﬁcant recovery of longleaf is therefore de- ally employing both intensive site prepara- pendent on the participation of the private tion and follow-up herbaceous competition sector. Certainly, for the forest industry, and control to improve survival and accelerate many other investor-type groups, the need for growth. competitive returns from forest management Longleaf pine can grow competitively with, is extremely important. And although expe- or even exceed, the growth of other southern rience has indicated that economic return is pine species on many sites. If markets continue often not the primary motivator for nonindus- to award quality wood products, particularly trial landowners, it usually plays some role in utility poles, with premium prices, longleaf management decision-making. is highly competitive. Private industrial and One major area requiring more knowledge is nonindustrial landowners should therefore re- the need for models to reliably project growth spond positively to that possibility and make and, ultimately, economic value of longleaf longleaf a vital part of their portfolio.

r John S. Kush School of Forestry and Wildlife Sciences, Auburn University, Auburn, Alabama 36849. J. C. G. r r Goelz USDA Forest Service, Southern Research Station, Pineville, Louisiana 71360. Richard A. Williams School r of Forest Resources and Conservation, University of Florida, Milton, Florida 32572. Douglas R. Carter School r of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611. Peter E. Linehan The Pennsylvania State University, Mont Alto Campus, Mont Alto, Pennsylvania 17237.

251 252 III. Silviculture Historical Perspective and yield is root competition for soil moisture. Longleaf stands are subject to severe droughts. The first European explorers to visit the south- The slow juvenile growth and long taproot of eastern U.S. Coastal Plain found a vast park- the young tree indicate its adaptation to this land of low ground cover growing under a rel- condition. Very young stands of longleaf may atively open canopy of pine (Bartram 1791; be quite crowded and remain so for 50 to 80 Schwarz 1907). Depending on the geography years. But it was found that such stands, if and/or soils, the dominant tree was longleaf closely crowded, fell off in growth so badly pine. that there was a distinct loss of production. Schwarz (1907) noted that within pure Trees less than 100 years of age continued to stands of longleaf pine, certain minor varia- grow vigorously in diameter even in rather tions existed. The most important variation dense groups, provided such groups were iso- was in the density of trees. Ordinarily, the lated and did not form a complete stand. But stand of trees did not maintain uniformity over above this age it was found that groups of sev- more than a few hundred acres; often changing eral trees standing close together would have abruptly even within 50 acres. Schwarz (1907) differentiated themselves into dominant and gave this description: suppressed trees, one or two trees with large crowns showing continued growth, while the Thus we may enter a stand of mature timber, with rest were almost stationary. These occurred in trees from 90 to 120 feet in height and with am- groups surrounded by open space and could ple spaces here and there in the crown cover, giving not be accounted for by struggles for light. Root entrance to the light from overhead. After walking competition alone can account for the thin- perhaps only a few hundred paces, we may find the ning out of mature longleaf forests and the trees suddenly beginning to close up their crown wide spacing of veteran trees. It also accounts, spaces. They grow smaller and more numerous, un- to a greater extent than intolerance, for the til presently they form a tolerably dense grove; and absence of seedlings under the open crowns then they open up once more into the original stand of veteran trees, and their appearance only in of mature, tall trees. Occasionally, too, a tract of old trees of fairly uniform height is replaced by one in openings at some distance from such trees. The which the trees show diversity in size, ranging from indicated management for longleaf pine is to mere poles to veterans of the forest. avoid crowding and not to attempt rotations much longer than 100 years or the production Although an extremely intolerant tree, of large sizes. The ideal form for young stands which will thrive best in even-aged stands, of longleaf would be to have them stocked at the natural form of longleaf pine tends to- most with only about twice as many saplings ward small, even-aged groups of a few hun- as should be standing in the form of mature dred square feet. Being naturally resistant to trees at the end of the rotation. fire, large clearings are never caused by fire. In It is evident that under natural conditions, regions of severe winds, or tornadoes, larger even in the presence of repeated fires, the even-aged patches and strips are found, some- longleaf pine forest renews itself, young trees times one-quarter to one-half mile in width, coming in on areas left blank by the death of which have come in after blowdown. These are old timber. All trees in a stand do not grow pretty well interspersed with patches or single equally fast, nor continue to grow at the same survivors of the old forest, which have acted as rate. In longleaf pine this is especially notice- seed trees. Fire always has and always will be able. Only the largest trees, with the biggest an element in longleaf forests, and the problem crowns, continue to grow at a rapid rate after is not how fire can be eliminated but how it a stand has reached merchantable size. After can be controlled so as to secure reproduction; a longleaf pine stand reaches the age of about second, to prevent the accumulation of litter 120 years, the loss from red rot, fire, and sup- and reduce the danger of a disastrous blaze. pressed growth increases so fast that the net The factor that probably determines growth gain in growth on the stand would not pay the 8. Longleaf Pine Growth and Yield 253 taxes. This slow destruction of timber that off- more complex structurally and thus more dif- sets growth is due chiefly to the inability of the ficult to model. soil to support so many trees of large size. Other attempts at predicting growth have involved the use of empirical yield tables, which are direct estimates of growth based on stand Old Growth Information structure and volume tables. Davis (1966) indicated that good estimates of growth should from the Literature provide growth directly in hectares, include the fewest possible variables, require a mini- Based on timber tallies of 162 ha (400 acres) mum of field data, provide estimates in cubic of pure even-aged, old-growth longleaf pine foot volume, do not use age as a primary vari- stands in Tyler County, TX, Chapman (1909) able, and treat height growth differently than found that trees per hectare dropped from 148 diameter growth. to 27 (60 to 11 trees/acre) going from a stand Spurr (1952) proposed a method to account 100 years old to 320 years old with average di- for diameter growth and height growth in pro- ameter at breast height (DBH) increasing from jecting volume growth. One must first be able 37.0 to 77.9 cm (14.0 to 29.5 in.). Mean annual to predict basal area and tree height growth. growth [in thousand board feet (MBF) per acre Growth in basal area is largely a function of per year] reached at a maximum at 110 years. stand density while height growth is primarily This indicated that longleaf pine does not in- a function of site quality and stand age. Once crease much in yield after 120 years with the basal area and height measurements are esti- increase in mortality due to decay, fire, and mated and volume calculated for the present other disturbances. The increase in total yield stand, then estimates of volume for some fu- is very slow up to 250 years and then dimin- ture stand can be calculated. Assuming the ishes. Trees less than 100 years old will grow stand form factor remains unchanged, the re- vigorously in diameter even in rather dense lationship between the product of basal area stands provided groups are isolated and do not and height (for a given volume) will hold for form a complete stand. the future period, i.e., Wahlenberg (1946) noted that old-growth forests were aggregations of even-aged stands PBA × PHt FBA × FHt covering areas from a few hundred square = feet to several acres. Well-stocked stands had PVol FVol 74 to 247 merchantable trees/ha (30 to 100 trees/acre), and poorer sites as few as 5 to where 7 trees/ha (2 to 3 trees/ha). Using data from Forbes and Bruce (1930), Wahlenberg (1946) PBA is the present basal area observed that over wide regions of the South, PHt is the present height the forest contained many age and size classes PVol is the present volume of trees, with only small areas usually limited FBA is the future basal area to a single age class. FHt is the future height FVolis the future volume

The difference between future volume and Growth and Yield of present volume is the increment of growth for Natural Stands the period. Another approach is to use a stand, or stock There has been a great deal of research on table, projection. The essence of a stand table the growth and yield of longleaf pine. Farrar projection is to estimate the future stand based (1979a) published the first growth and yield on the present one. The problem is what diam- equations for thinned stands of even-aged nat- eter increment to use and how to apply the ex- ural longleaf pine. Uneven-aged stands are pected growth to the diameter class. Diameter 254 III. Silviculture increment data could be added to the mid- uneven-aged management with longleaf pine point of the diameter class. This method as- is that a lot of work is required to keep up with sumes that all trees in a given diameter class how a stand is growing. The fact that it takes a are at the midpoint and grow at the same rate. considerable amount of work to manage lon- This assumption reveals the limitation of this gleaf pine under an uneven-aged system has method of predicting stand growth as all trees discouraged past management of longleaf pine in a given diameter class will not be the same in this way. However, recently public agencies size and all trees will not be at the midpoint of such as state forest divisions or departments the diameter class. and the USDA Forest Service are attempting to A second method applies the average diame- manage longleaf pine using the uneven-aged ter growth to all trees within a class while rec- silvicultural approach. ognizing dispersion of individuals within the The best estimate of longleaf pine growth same class. This method uses a movement ratio and yield for natural stands can be found technique to predict what percentage of trees in Farrar (1979b). The USDA Forest Service within a diameter class move up into larger di- established a regional longleaf pine growth ameter class, and what percentage will remain study (RLGS) in the mid-1960s in south- in the same diameter class. The movement ra- west Georgia, northwest Florida, southern tio is defined as Alabama, and southern Mississippi. The data from this study and the subsequent formulas M = (g/i) × 100 represent the net volume growth and yield one might expect in the absence of adverse influ- where ences such as weather, insects, and disease. The equations for estimating growth and yield (in- M = movement ratio side bark) of thinned natural longleaf pine are g = diameter growth increment given below (the symbols and letters follow i = diameter class interval throughout all of the formulas). For example, assume the average diameter Basal area is calculated as growth was 5.3 cm (2.1 in.) over a period of time, and the diameter class interval is 5 cm BA = 0.2296B2 (2 in.), then the movement ratio is calculated B2 = e[(A1/A2) ln(B1) + 6.0594(1 − A1/A2)] to be 105. This means that 5% of the trees move up two diameter classes while 95% of where the trees move up only one diameter class. = Growth and yield models have been devel- BA projected basal area at the end of the oped for uneven-aged stands of loblolly and period in square meters per hectare = shortleaf pines (Farrar et al. 1984; Murphy B2 projected basal area at the end of the pe- and Farrar 1985). Some general inferences riod in square feet per acre = from these studies conducted in Arkansas were e exponential function = that loblolly and shortleaf pine average annual A1 initial stand age in years (beginning of growth could be expected to be around 0.3 m2 the period) = (3 ft2) of basal area and 2.4 to 3.3 m3 (84 to A2 stand age at the end of the period in years = 116 ft3) of merchantable volume growth. Vol- ln natural logarithm = ume growth on the Mississippi study locations B1 initial basal area in square feet per acre showed slightly lower production compared to Total volume is given by the Arkansas sites (Baker et al. 1996). Farrar (1996) provided guidelines for the TVIM2 = 0.06997 TVI2 uneven-aged management of longleaf pine. TVI = e[2.6776 + 0.015287(S) Nature managed longleaf pine as small 2 − . / + / patches of even-aged stands across an uneven- 21 909 A2 (A1 A2) ln(B1) aged landscape. The main drawback with + 6.0594(1 − A1/A2)] 8. Longleaf Pine Growth and Yield 255 where where

TVIM2 = projected stand total volume, inside V4IM = predicted stand merchantable volume bark, in cubic meters per hectare at the end in cubic meters per hectare, inside bark, of the period, DBH ≥ 1.5 cm (0.6 in.), 6.1- DBH ≥ 9.1 cm (3.6 in.), top DOB (diam- cm (0.2 ft) stump, S = site index in feet, eter outside bark) ≥ 7.6 cm (3 in.), 6.1-cm base age of 50 years (0.2 ft) stump TVI2 = projected stand total volume, inside V4I = predicted stand merchantable volume in bark, in cubic feet per acre at the end of cubic feet per acre the period These formulas have their limitations. First, When A1 = A2 and B1 = B2 or the growth pe- they were derived from the first 5 years riod is 0, then the above equation can be re- of the study. Actual growth beyond 5 years duced to has not been used to adjust formula coeffi- cients. Indeed, the study report indicates that TVIM = 0.06997 TVI the estimates of total cubic foot volume pro- TVI = e[2.6776 + 0.015287(S) vided the most reliable results while the for- − 21.909/A + ln(B)] mula for estimating merchantable volume was the least reliable. A second limitation is that where A, S, and B are current stand age, site these equations were derived for trees grow- index, and basal area, respectively, TVIM is the ing in stands that were thinned. These equa- predicted current stand volume in cubic feet tions used trees from thinned stands, a 5-year per hectare, and TVI is the predicted current growth period, and low mortality. Future yield stand volume in cubic feet per acre. predictions were also not reliable, and the es- The equation for calculating the predicted timates became unrealistically large for un- stand total volume (outside bark) is thinned stands. Even though the equations developed by the TVOM = 0.06997 TVO Forest Service and reported by Farrar (1979b) TVO = TVI [1 + e{−0.1785 + 43.629/S have their limitations, they do provide a useful + 1108.6/(SB) − 0.42802(ln(A)) estimate of a stand’s volume and an estimate of the stand’s future volume and basal area. − . / } 360 87(ln(A)) (SB) ] Stand predictions of growth and yield are useful to landowners attempting to manage their where forests. TVOM = cubic meters per hectare Quicke et al. (1994) used the RLGS database TVO = volume in cubic feet per acre and produced an individual tree basal area increment (BAI) model for longleaf pine. The The above equation for calculating the total model is an intrinsically nonlinear equation, stand volume (outside bark) can be used for ei- which is constrained so that it performs within ther the beginning or initial stand condition or the bounds of biologically reasonable outputs the final condition using the appropriate TVIM, for any combination of values for the indepen- TVI, age, site index, and basal area figures. dent variables. All parameters in the equation Farrar (1979b) also published formulas to were estimated simultaneously. This is a depar- determine merchantable volumes for both ture from the more traditional potential-times- present and future stands. The merchantable modifier approach in which parameters for a volume formula is potential growth function are estimated from V4IM = 0.6997 V4I a sample of trees exhibiting the fastest growth. Independent variables used to describe BAI are V4I = TVI/[1 + e{2.623 + 316.77/S stand basal area, the competitive position of an + − . (SB) 2 8248(ln(A)) individual tree within the stand calculated as − 3326.7(ln(A))/(SB)}] the sum of the basal areas of all trees larger 256 III. Silviculture than the subject tree, mean age of dominant top is given by and co-dominant trees, and individual tree di- YM = 0.4536 Y ameter outside-bark at breast height. Notice- =− . + . 2 ably absent from the model is an independent Y 36 83043 0 15608(D Th) variable that explicitly characterizes site differ- The study gives the cubic foot volume of the ences. stem to a 10.2-cm (4 in.) top as Further work by Quicke et al. (1997) cre- = . ated an individual tree annual survival rate VM 0 2832 V model. Variables used in the model were pre- V =−0.84281 + 0.02216(D 2Th) dicted diameter increment and diameter at where breast height (DBH). Predicted annual survival rates ranged from 0.92 for a tree with a 2.54 VM = volume in cubic meters cm (1 in.) DBH and an annual diameter in- V = volume in cubic feet crement of 0.13 cm, to over 0.99 for any tree For longleaf pine, the paper also gives green larger than 15 cm in DBH. Stand level verifica- weight to 17.8-cm (7 in.) and 22.9-cm (9 in.) tion was based on 102 comparisons of observed tops; green weight of wood, bark, and foliage; and predicted trees per acre. Mean residuals, wood volume to 17.8- and 22.9-cm tops; board expressed as a percentage of observed final foot (Scribner) volumes; green weight of saw- trees per acre, were 3% and 6% for projec- timber per MBF; and pulpwood weights and tion periods of 5 and 10 years, respectively. volumes. The model predicts noncatastrophic mortality. In conjunction with a basal area increment model, it can be used to predict changes in the structure of longleaf pine stands. Meldahl et al. Growth and Yield (1997) used the RLGS dataset to calculate nee- of Planted Stands dle fall, standing biomass, net primary productivity, and projected leaf area. In addition, cli- The most broadly based system of stem pro- matic variables were included in tree and stand file equations (and hence volume) is provided models. by Clark et al. (1991). Clark et al. (1991) in- Another study by Saucier et al. (1981) de- clude equations to predict stem diameter at any veloped weight, volume, board-foot, and cord height, given diameter at breast height and at tables for major southern pine species, includ- Girard’s Form Height (5.3 m or 17.3 ft), to- ing longleaf pine. Data for this study were de- tal height, and the height at which diameter rived by felling sampled trees and measuring is to be predicted. The equations can also be diameter, total height, and height to various used to estimate height at a given minimum merchantability limits. diameter (merchantable height), and volume The equation for predicting the total tree in cubic feet to any minimum diameter, or vol- green weight using DBH and total height is ume between a maximum and minimum diameter (such as pulpwood volume above saw- YM = 0.4536 Y timber volume). There also are equations to use if height to a top diameter, rather than to- =− . + . 2 Y 44 418879 0 20297(D Th) tal height, is known. If diameter at form height is not measured, Clark et al. give an equation where to predict it from diameter at breast height and total height. Bark thickness at breast height YM = total tree weight in kilograms can also be estimated from an equation that Y = total tree weight in pounds predicts diameter inside bark. Clark et al. fur- D = DBH in inches nish parameter estimates for longleaf pine that Th = total height in feet represent South-wide estimates as well as sub- The total tree green weight to a 10.2-cm (4 in.) regions, namely, Coastal Plain (Atlantic coast 8. Longleaf Pine Growth and Yield 257 and eastern Gulf—Alabama and points east measured on every tree, it should be possible and north), Piedmont (all states), and Deep to assign crown ratio class reasonably accu- South (Texas, Louisiana, and Mississippi). rately to trees as height is measured. These ta- Thomas et al. (1995) provide biomass and taper and volume equations should be useful for per equations for longleaf pine in thinned and estimating volume in longleaf pine stands in unthinned plantations in Louisiana and Texas. the western Gulf states (Texas and Louisiana). Their taper equation predicts upper stem di- Brooks et al. (2002) provide taper and cubic ameters as a function of relative height, diam- foot volume equations for young plantations in eter at breast height, and plantation age. Vol- southwest Georgia; they compared their equa- ume is obtained by the integral of the taper tions to the equations of Baldwin and Polmer equation, between limits of merchantability. (1981) and Baldwin and Saucier (1983), but Weight of the bole is obtained by doubly inte- they did not measure crown ratio, and so could grating (across diameter and height) a function not fully utilize Baldwin and Polmer’s (1981) for specific gravity. However, their equation for equations. The taper equation of Brooks et al. specific gravity was unique for the ages of the (2002) was slightly superior to that of Baldwin sampled trees (35, 45, or 50 years), and thus and Polmer (1981), and their volume equa- has somewhat restricted applicability. Thomas tion was slightly superior to those of Baldwin et al. compared their taper equations for four and Polmer (1981) and Baldwin and Saucier classes of stands: (1) an unthinned Louisiana (1983); however, this is expected because they plantation; (2) a thinned Louisiana plantation; used the same data to fit and test their equa- (3) a different thinned Louisiana plantation tion, while the Baldwin equations were fit to and two thinned Texas plantations; (4) natural other data. It is reasonable to expect bias when and plantation-grown longleaf from various an existing equation is applied to a new dataset stands in Alabama. These four classes produced that is geographically distinct from the data taper equations that differed; the Alabama ta- on which the equation was estimated. How- per curve was particularly different from the ever, the bias was not large, and when the bias others. This suggests that site and management was made equal to the bias of the equation differences can have large effects on stem ta- of Brooks et al. (2002), the taper equation of per. If stem taper varies so much from stand Baldwin and Polmer (1981) would have pro- to stand, this suggests that regional volume duced lower absolute error than the equation equations will be poor estimators for any given of Brooks et al. (2002) (calculations by the sec- stand, unless the volume/taper equation in- ond author). This suggests that Baldwin and cludes some measurement of form beyond Polmer (1981) might be more broadly applica- simply measuring diameter at breast height ble than Brooks et al. (2002). and total height. This suggests that Clark et al. Baldwin and Saucier (1983) provide above- (1991), with diameter at form height deter- ground weight and volume estimators for mined from local data, would be preferable to unthinned planted longleaf pine, using 111 regional volume equations. Clark et al. (1991) of the 113 trees sampled in Baldwin and can be applied to natural and planted longleaf, Polmer (1981). Rather than using a taper equa- and they have estimated their equation for all tion, they used a combined variable equation common species or species groups that are as- (Clutter et al. 1983) given by sociates of longleaf pine. log(Y) = b + b ∗ log(D 2 H) Baldwin and Polmer (1981) used data from 0 1 some of the same plantations as Thomas et al. where Y is either volume or some biomass (1995). They fit three different taper equations component, D represents diameter at breast for different classes of crown ratio (less than height, and H represents total height. To es- 36%, between 36% and 50%, and greater than timate volume to some minimum top diam- 50%). Crown ratio potentially can reflect the eter, they estimate a volume ratio equation differences in taper among trees within and that is multiplied by the value for total vol- among stands. Even when crown ratio is not ume. As they did not use any variable, such 258 III. Silviculture as crown ratio, that could explain stand-to- height divided by diameter outside bark at stand and tree-to-tree variability in taper, we breast height). Use of the equations may be suggest the taper equations of Baldwin and more efficient for some individuals than look- Polmer (1981) would be preferable for es- ing up the values in a table, although the table timation of volume. However, Baldwin and look-up can be programmed. Wiant (1986) as- Saucier’s (1983) biomass equations are prob- sumed a 3% change in volume for each point ably the only regional equations to estimate of form class change from 78 (higher volumes green and dry weight of wood, bark, branches, would be obtained with higher values of form and foliage, given the limitations in applica- class). A landowner may have a good idea of tion of the bole biomass equations of Thomas the form class for his holdings, perhaps as a et al. (1995). There are biomass equations for function of diameter and height of the tree. Al- longleaf in very young natural stands (presum- ternatively, Clark et al. (1991) have an equa- ably natural stands, the publication only in- tion for diameter inside bark at 5.3 m (17.3 dicates they are even-aged; Edwards and Mc- ft). Because of the structure of the equation, Nab 1977) and an old natural stand (Taras and form class is a constant for a given total height; Clark 1977). The results of Taras and Clark that is, diameter at breast height does not af- (1977) are included in USDA (1984) tables. fect form class. This is counterintuitive, as it To estimate board foot volume, there are a would seem likely that form class should de- few different approaches. The first approach pend on diameter and height, but the rela- is to use volume tables derived from natural tionship seems to hold for the data of Clark stands, such as those found in USDA (1929). et al. (1991). The corresponding form class pro- These volume tables use diameter at breast duced by their equation is provided in Table 1. height and total height to predict board foot On the other hand, Parker (1998) suggests that volume by several different log rules. The sec- taper could be constant within diameter classes ond approach employs form class volume ta- rather than height classes. The taper equations bles, such as Mesavage and Girard (1946). of Thomas et al. (1995) suggest that form class Wiant (1986; Wiant and Castaneda 1977) cre- varies in response to diameter, height, and age, ated equations that approximate the Mesavage and does so differently for thinned and un- and Girard (1946) tables for form class 78 thinned stands. Age has a relatively small dif- (form class is diameter inside bark at 17.3 feet ference on taper, but DBH and height have

TABLE 1. Form class (in percent) by height relationships for three subregions and South-wide, as calculated from an equation of Clark et al. (1991) Total height (ft)

Region 40 50 60 70 80 90 100 110 120

South-wide 66 73 77 79 81 82 83 83 84 Coastal Plain 67 73 77 79 80 81 82 82 83 (AL to VA) Piedmont 65 73 77 80 81 82 83 84 84 Deep South 62 72 77 80 83 84 85 86 86 (TX, LA, MS) Total height (m)

Region 12.2 15.2 18.3 21.3 24.4 27.4 30.5 33.5 36.6

South-wide 66 73 77 79 81 82 83 83 84 Coastal Plain 67 73 77 79 80 81 82 82 83 (AL to VA) Piedmont 65 73 77 80 81 82 83 84 84 Deep South 62 72 77 80 83 84 85 86 86 (TX, LA, MS) 8. Longleaf Pine Growth and Yield 259

TABLE 2. Form class (in percent) at age 40 related to way, although it is tedious if the calculations diameter (DBH) and total height for thinned and are done by hand rather than programmed. unthinned stands in Louisiana and Texas, calculated A program to calculate board foot volume in from taper equation of Thomas et al. (1995) and this way is available from the second author bark thickness equation of Clark et al. (1991) ([email protected]) or from the programmer, DBH (in.) Height (ft) Thinning Form class Daniel Leduc ([email protected]). 9 50 Thinned 74 Poles are a high-value product, greatly ex- 9 80 Thinned 81 ceeding the value of sawtimber per board foot, 12 50 Thinned 75 and thus it is critical to determine yield of poles. 12 80 Thinned 82 9 50 Unthinned 73 Any of the taper equations can also be used to 9 80 Unthinned 79 determine whether a tree of a given diame- 12 50 Unthinned 74 ter at breast height and total height possesses 12 80 Unthinned 80 the minimum dimensions for top diameter and DBH (cm) Height (m) Thinning Form class length of poles. ANSI (1987) provides specifi- 22.9 15.2 Thinned 74 cations for dimensions of poles for 10 quality 22.9 24.4 Thinned 81 classes for pole lengths of 6.1 to 38.1 m (20 to 30.5 15.2 Thinned 75 125 ft), in increments of 1.5 m (5 ft). The spec- 30.5 24.4 Thinned 82 ifications are in terms of circumference (diam- 22.9 15.2 Unthinned 73 eter times π) at 1.8 m (6 ft) from the butt of 22.9 24.4 Unthinned 79 30.5 15.2 Unthinned 74 the pole. The taper equation can be solved for 30.5 24.4 Unthinned 80 diameter at a height of 2.2 m (7.3 ft; counting stump and sawkerf). Hawes (1947) assumed a constant ratio of diameter inside bark at a a larger effect (Table 2). There are relatively height of 1.8 m (6 ft) to be 0.88 times diam- small differences in absolute numbers between eter outside bark at breast height. Using taper Tables 1 and 2. However, there are fundamen- and bark thickness equations would probably tal differences in the choice of the factors that be more accurate for a specific tree. Quicke affect form class. This could be very important and Meldahl (1992) used a taper equation for when contrasting different silvicultural prac- natural longleaf to create such tables. Any ta- tices. If silvicultural practices affect form class, per equation and bark thickness equation for the chosen equation might not reflect those planted longleaf could be used in a similar differences and thus the real difference among fashion. The differences between the tables of treatments might not be apparent. Quicke and Meldahl (1992) and Hawes (1947) The values in Table 1 can be used in con- were greatest for long poles. Busby et al. (1993) cert with Mesavage and Girard’s (1946) ta- found that 90% of trees in their plantations bles or Wiant’s (1986) equation. Borders and of longleaf pine in Louisiana were sufficiently Shiver (1995) used the taper equation of Clark free of defect to produce a pole if they met the et al. (1991) to produce board foot tables for diameter and length requirements. loblolly, slash, and shortleaf pines. Their tables suggested greater volume than Mesavage and Girard (1946). This might suggest that Evaluating Site Quality Mesavage and Girard (1946) underestimate board foot volume for longleaf pine as well. Site quality is a critical component of most Borders and Shiver (1995) present the final growth and yield models. Lohrey and Bailey procedure for calculating board foot volume. (1977) created a growth and yield model for They used a taper equation to calculate inside unthinned plantations of longleaf pine. In it, bark diameters at the scaling diameter of each they used the site index equations produced log of fixed length, and directly applied their by Farrar (1973). Farrar (1973) developed chosen log rule to calculate volume of each his equations to reproduce the graphical site log. Any taper equation could be used in this curves provided for natural second growth 260 III. Silviculture

TABLE 3. Number of trees (A) per acre and (B) per hectare to be planted to achieve thinning thresholds of GLSDI of 50% of maximum or 4/3 of crown closure for different levels of quadratic mean diameter and survival (to age of ﬁrst thinning) 90% survival 70% survival 50% survival

1/2 4/3 crown 1/2 4/3 crown 1/2 4/3 crown DBH (in) maximum closure maximum closure maximum closure 4 1355 1014 1742 1303 2439 1824 6 604 463 777 595 1088 834 8 341 264 438 340 613 475 10 218 170 280 219 393 307 12 151 119 195 153 273 214 DBH (cm) 10.2 3348 2506 4305 3220 6027 4507 15.3 1493 1144 1920 1470 2689 2061 20.4 843 652 1082 840 1515 1174 25.4 539 420 692 541 971 759 30.5 373 294 482 378 675 529

longleaf pine in USDA (1929). It should be in the anamorphic curves, but is obtained in noted that while the USDA (1929) curves in- the polymorphic curves of Goelz and Leduc dicate zero height before age 5, Farrar’s (1973) (2003). Brooks (2004) describes an equation to curves are not conditioned in this way, and predict dominant stand height for young lon- do not adequately represent the USDA (1929) gleaf plantations in southwest Georgia. curves at ages below 15 years. Apparently, Boyer (1980, 1983) suggested that plant- Lohrey and Bailey (1977) found the curves for ing site (old field, unprepared cutover, me- natural longleaf pine stands (USDA 1929; Far- chanically prepared cutover) and stand density rar 1973) adequately represented the planta- (survival at 10 years) affected early growth of tion grown longleaf data that they had avail- longleaf pine. Boyer used a simple Schu- able. Goelz and Leduc (2003) provided prelim- macher equation: inary site index curves using the data of Lohrey = + 1 and Bailey (1977), as well as additional mea- log10(H) bo b1 surements on the same plots, and supplemen- A tal plots that were not available to Lohrey and to fit a common guide curve for all site condi- Bailey (1977). The curves of Goelz and Leduc tions, where H and A represent height and age, (2003) are very similar to the USDA (1929) respectively, and the bi are parameters. Then, curves for site index of 70 at base age of 50 he expanded the equation by making b1 a lin- (Fig. 1). However, the curves for site index ear function of surviving trees per unit of land 50 and 90 are considerably different from the and height at age 15 (height at age 15 was only USDA (1929) curves. The anamorphic curves included for the two cutover sites). This struc- represented by USDA (1929) do not change ture produced site curves that are anamorphic shape as site index changes. There is a very for old-field situations and polymorphic for common phenomenon that arises in polymor- the two cutover situations and having a com- phic site index curves where the curves for mon asymptote for all combinations of site in- the lower sites tend to be more linear while dex and site condition. Although he labeled the curves for the higher sites tend to be more his curves by height at age 25, by including curved, achieving a higher proportion of the height at age 15 as a predictor variable, he was asymptotic height at younger ages (Goelz and essentially creating base age specific site curves Burk 1996). This expected pattern is missed with 15 as the base age. 8. Longleaf Pine Growth and Yield 261

FIGURE 1. Comparison of site index curves from USDA (1929), dashed lines, and Goelz and 90 Leduc (2003), solid lines, for site index (base age 50) of 50, 70, and 90 feet.

50 Height (feet)

10 10 20 30 40 50 60 70 Age (years)

Goelz and Leduc (2004) suggested that a ity suggested loams or sandy loam soils (6– base age of 50 years provided more reliable 9%). Rainfall is clearly related to geography, estimates of site index than a base age of and thus the effect of rainfall is confounded 25 years. The trade-off seems to be to either with soil differences and perhaps difference model this early height growth directly, or lose in constituents of the competing vegetation. some of the capacity to define site quality, or Moisture-holding capacity affects competing to ignore the early, largely random variation vegetation, and thus the effect may be due and concentrate on the intrinsic productivity to competition rather than moisture avail- of the site. We suspect that the different bias ability per se. An earlier analysis at age 15 of the perspectives of Boyer (1980, 1983) and (Shoulders and Walker 1979) suggested that Goelz and Leduc (2004) arises from the differ- the effect of rainfall was different on sites with ent natures of their datasets, predominantly droughty soils compared to wet or interme- 15 years or younger versus predominantly 16 diate sites. Rainfall was positively related to years and older. height for droughty soils while negatively re- Shoulders and Tiarks (1980) produced related to height for wet and intermediate sites. gression models for planted longleaf pine in This suggests that aeration may be limiting to Louisiana and Mississippi relating height at age growth where soils may become saturated, but 20 to rainfall, slope, and available soil mois- not on inherently dry sites. To estimate site in- ture (1/3 atmosphere percent minus 15 atmo- dex, a user could estimate height at age 20, sphere percent) in the B2 horizon. Longleaf using the equation of Shoulders and Tiarks pine height was greatest where rainfall was (1980), then apply this height to a site index least 122 cm (48 in.) per year (evenly split equation or curve. between warm season, April through Septem- Harrington (1990) produced an expert sys- ber, and cool season), on very modest slopes tem to predict site index of both natural (1.6%), and where moisture-holding capac- and planted longleaf pine from 25 soil and 262 III. Silviculture physiographic variables. Most of the variables maximum SDI of shortleaf and slash pines, but need only be given qualitative values, or as- less than the maximum for loblolly pine (450). signed to classes of continuous variables. A The southern variant of the Forest Vegetation user can make reasonable guesses for the more Simulator (Donnelly et al. 2001) uses a max- difficult to measure variables, such as percent imum Reineke stand density index of 390 for phosphorus, from soil survey information or longleaf pine, based on their data. regional data. The system can predict base age The actual maximum density line has lim- 50-site index within 1.7 m (5.5 ft). It is pre- ited applicability to management since most sented as a stand-alone computer application, forests would be maintained at densities much but the source code could be extracted and in- less than the maximum. However, a line can corporated into larger growth and yield sys- be defined that is parallel to the limiting den- tems. As Shoulders and Tiarks (1980) devel- sity line that represents the threshold of signif- oped their equation from data arising only icant density-dependent mortality (Drew and from Louisiana and Mississippi, Harrington’s Flewelling 1979; Dean and Baldwin 1993). (1990) system might be more suitable outside This line typically represents 50% to 55% of those two states. the maximum density. Under typical management, a stand would be maintained at or below this level of density. Evaluating Growing Stock In Fig. 2, we plot a limiting relationship for and Stand Density longleaf pine plantations from a large database of Longleaf Pine Stands 9 Stand density affects growth and shape of individual trees as well as understory plant communities and affects habitat quality for wildlife (Grelen and Lohrey 1978; Clutter et al. 1983; Haywood et al. 1998). Stand density can be 7 described as simple measurements of number of trees, basal area, or volume per acre. Or the variables may be combined into a stand LNTPA density index. Most stand density indices are 5 functions of two or more of (1) basal area per acre, (2) trees per acre, or (3) average tree size (in terms of quadratic mean diameter, volume, or weight). Most stand density indices are independent of stand age and site qual- 3 ity, except as those variables influence tree size 1.0 1.5 2.0 2.5 3.0 3.5 and mortality. Reineke’s (1933) stand density LNQMD −1.605 index (SDI) is N(10/Dq) , where N rep- IGURE resents number of trees per acre, Dq repre- F 2. A limiting density relationship for lon- sents quadratic mean diameter, and −1.605 is gleaf pine plantations. Only plots with a Reineke an empirically derived constant for all species. stand density index of 100 or greater were used. The line was fit by minimizing the function: The Reineke relationship arose out of the ob- −1.992 12 = − 2 Dq servation that there seemed to be a limiting loss (observed predicted) N 10 . This straight-line relationship when the logarithm weighting ensures that the line will approach the of quadratic mean diameter was plotted versus limit of the data. Note that there are no plots with the logarithm of trees per acre. Reineke (1933) large quadratic mean diameter near the limiting suggested a maximum stand density index of line; plots with the largest quadratic mean diameter 400 for longleaf pine, which is the same as the were thinned fairly heavily. 8. Longleaf Pine Growth and Yield 263

(Goelz and Leduc 2001). However, we found threshold for significant density-dependent the exponent to be −1.992, rather than mortality, and a line that represents crown Reineke’s −1.605. If the exponent was −2, this closure, as determined by the equations of would indicate that the limiting density re- Smith et al. (1992) for crown diameter of open- lationship would represent a maximum basal grown longleaf pine. As the GLSDI is based on area for all levels of quadratic mean diame- an exponent very close to two, and as Smith ter. As −1.992 is very close to −2, the max- et al. (1992) predict open-grown crown di- imum basal area is 49.1 m2/ha ± 0.19 (214 ameter as a linear function of diameter, basal ft2/acre±2) across a wide range of quadratic area is nearly constant across a broad range mean diameter from 6.4 to 66 cm (2.5 to of diameter. Thus, maximum basal area is ap- 26 in.). The southern variant of the Forest proximately 44.7 m2/ha (215 ft2/acre), the Vegetation Simulator (FVS) employs a maxi- threshold for significant density-dependent mum basal area of 48.9 m2/ha (213 ft2/acre) mortality is 22.2 m2/ha (107 ft2/acre), and (Donnelly et al. 2001). This is very similar to crown closure occurs at approximately 14.5 the maximum indicated by the data of Goelz m2/ha (63 ft2/acre). Appropriate levels of basal and Leduc (2001). However, this should not area would vary depending on management be surprising, as Donnelly et al. (2001) incor- objectives. However, a basal area of 14.5 m2/ha porated the data of Goelz and Leduc (2001) (63 ft2/acre) would provide high rates of indi- with data they had available from natural vidual tree growth while not sacrificing much stands. whole stand growth. Higher levels of basal We used the limiting density relationship area would produce slower individual tree to build a density management diagram for growth, but somewhat greater whole stand longleaf pine plantations. As the exponent growth, somewhat higher log quality, and is not −1.605, we call the index the Goelz– losses to mortality would be slight if stands Leduc stand density index (GLSDI) for longleaf were thinned before they exceeded 24.6 m2/ha pine, rather than the Reineke stand density in- (107 ft2/acre). Although the data of Goelz and −1.992 dex. It may be calculated as N(Dq/10) , Leduc (2001) included some plots with greater with the maximum SDI calculated to be ap- than 49.4 m2/ha (215 ft2/acre) of basal area, proximately 393. Three lines are present on the plots were small (roughly 0.04 ha, 0.1 acre) Fig. 3. There is a maximum density line, a and were not likely indicative of larger plots line representing 50% of maximum, or the (0.4 ha, 1.0 acre) or stands.

3119

Maximum GLSDI 1563 50% of Maximum Crown Closure 783

393

197 Trees per Acre 99

FIGURE 3. Density management diagram for longleaf pine planta- 49 tions. Lines are drawn for maximum density, 50% of maximum, 25 and for crown closure. Note the 4.0 5.0 6.3 7.9 10.0 12.6 15.8 20.0 axes are scaled logarithmically. Quadratic Mean Diameter (inches) 264 III. Silviculture The GLSDI also can be used to help guide Acknowledgments initial planting density. We will consider two thresholds for thinning—50% of maximum Drs. Tom Croker and William Boyer are ac- GLSDI and 4/3 times the density at crown clo- knowledged for the work they did over the sure. The former threshold implies removal decades at the Escambia Experimental Forest of one-third to one-half of the basal area and for their teachings on how to naturally re- while the latter implies removal of about generate longleaf pine and manage it. We ex- one-fourth to one-third of the basal area tend our sincere thanks to Dr. Robert M. Farrar, in the thinning. Merchantability of a thin- Jr. It was his tenacity that has provided much ning will be influenced by size of the trees of the data available today for modeling lon- that are harvested. We consider first thin- gleaf pine growth and yield. nings at quadratic mean diameters between 10.2 and 30.5 cm (4 and 12 in.). We explore three levels of survival to first thinning— 50%, 70%, and 90%. Note this is survival to first thinning, not initial survival. Finally, References our calculations assume there are no de- ANSI. 1987. American national standard for wood sirable volunteer trees in the plantation. If poles—specifications and dimensions. ANSI 05.1– quadratic mean diameter is 10.2 cm (4 in.), 1987. and survival to first thinning is 90%, 3033 Baker, J.B., Cain, M.D., Guldin, J.M., Murphy, P.A., trees/ha (1355 trees/acre) must be planted and Shelton, M.G. 1996. Uneven-aged silvicul- to produce 24.6 m2/ha (107 ft2/acre). At a ture for the loblolly and shortleaf pine forest cover types. General Technical Report SO-118. USDA quadratic mean diameter of 15.2 cm (6 in.), Forest Service, Southern Research Station. 1492 trees/ha (604 trees/acre) are required, Baldwin, V.C., Jr., and Saucier, J.R. 1983. Above- and at 20.3 cm (8 in.), 138 surviving trees/ha ground weight and volume of unthinned, planted (341 trees/acre) are needed. So, if thinning longleaf pine on West Gulf forest sites. Research a stand with a quadratic mean diameter of Paper SO-191. USDA Forest Service, Southern 15.2 cm (6 in.) is practicable, and the manager Forest Experiment Station, New Orleans, LA. sought to maximize total volume growth, ini- Baldwin, V.C., Jr., and Polmer, B.H. 1981. Taper tial planting density should be about 1483/ha functions for unthinned longleaf pine plantations (600/acre), assuming survival of 90%. If sur- on cutover West Gulf sites. In Proceedings of the vival were 70%, more than 1903 trees/ha First Biennial Southern Silvicultural Research Confer- (770 trees/acre) would be planted. The conse- ence, ed. J.P.Barnett, pp. 156–163. November 6–7, 1980, Atlanta, GA. General Technical Report SO- quence of lower than expected survival means 34. USDA Forest Service, Southern Forest Exper- that thinning could be delayed a few years iment Station. until quadratic mean diameter is greater. For Bartram, W. 1791. Travels through North & South Car- example, if 90% survival were anticipated, olina, Georgia, East & West Florida, the Cherokee Coun- but 50% was achieved, thinning would be try, the Extensive Territories of the Muscogulges, or Creek delayed until quadratic mean diameter was Confederacy, and the Country of the Chactaws. Pub- at least 5.1 cm (2 in.) larger than antici- lished as the Francis Harper’s naturalist edition. pated, which is predicted to be about 5 years Athens: University of Georgia Press. 1998. later using the model of Lohrey and Bailey Borders, B.E., and Shiver, B.D. 1995. Board foot vol- (1977). Besides the delay to first thinning, ume estimation for southern pines. South J Appl there would likely be a reduction in log qual- For 19: 23–28. Boyer, W.D. 1980. Interim site-index curves for lon- ity. A stand that is not sufficiently dense to gleaf pine plantations. Research Note SO-261. thin until it has a quadratic mean diameter USDA Forest Service, Southern Forestry Exper- from 25.4 to 20.5 cm (10 to 12 in.) will likely iment Station, New Orleans, LA. have more defects due to knots and persis- Boyer, W.D. 1983. Variations in height-over-age tent dead branches, unless pruning was per- curves for young longleaf pine plantations. For formed. Sci 29: 15–27. 8. Longleaf Pine Growth and Yield 265

Brooks, J.R. 2004. Predicting and projecting stand Edwards, M.B., and McNab, W.H. 1977. Biomass dominant height from inventory data for young prediction for young southern pines. J For 77: longleaf pine plantations in southwest Georgia. In 291–292. Proceedings of the Twelfth Biennial Southern Silvicul- Farrar, R.M., Jr. 1973. Southern pine site index tural Research Conference, ed. K. Connor, pp. 187– equations. J For 71: 696–697. 188. Biloxi, MS. February 24–28, 2003. General Farrar, R.M., Jr. 1979a. Growth and yield predic- Technical Report SRS-71. Asheville, NC: USDA, tions for thinned stands of even-aged natural lon- Forest Service, Southern Research Station. gleaf pine. Research Paper SO-156. USDA Forest Brooks, J.R., Martin, S., Jordan, J., and Sewell, C. Service, Southern Forest Experiment Station. 2002. Interim taper and cubic-foot volume equa- Farrar, R.M., Jr., 1979b. Volume production of tions for young longleaf pine plantations in south- thinned natural longleaf. In Proceedings Longleaf west Georgia. In General Technical Report SRS- Pine Workshop, ed. W.E. Balmer, pp. 30–48. 48, pp. 467–470. Asheville, NC: USDA Forest Ser- Mobile, AL, October 17–19, 1978. USDA Tech- vice, Southern Research Station. nical Publication SA-TP3. USDA Forest Service, Busby, R.L., Thomas, C.E., and Lohrey, R.E. 1993. SE Area, State & Private Forests, Atlanta, GA. Potential product values from thinned longleaf Farrar, R.M., Jr. 1996. Fundamentals of uneven- pine plantations in Louisiana. In Proceedings of aged management in southern pine. Tall Timbers the Seventh Biennial Southern Silvicultural Research Research Station Miscellaneous Publication 9. Conference, ed. J. C. Brissette, pp. 645–650. Mo- Farrar, R.M., Jr., Murphy, P.A., and Willett, R.L. bile, AL, November 17–19, 1992. General Tech- 1984. Tables for estimating growth and yield of nical Report SO-93. USDA Forest Service, South- uneven-aged stands of loblolly-shortleaf pine on ern Forest Experiment Station, New Orleans, average sites in the West Gulf area. Bulletin 874. LA. Fayetteville: University of Arkansas, Arkansas Carmean, W.H. 1975. Forest site quality evaluation Agricultural Experiment Station. in the United States. Adv Agron 27:209–267. Forbes, R.D., and Bruce, D. 1930. Rate of growth of Chapman, H.H. 1909. A method of studying growth second-growth southern pines in full stands. U.S. and yield of longleaf pine applied in Tyler County, Department of Agriculture Circular 124. Texas. Proceedings, Society of American Foresters 4: Goelz, J.C.G., and Burk, T.E. 1996. Measurement 207–220. error causes bias in site index equations. Can J For Clark, A., III, Souter, R.A., and Schlaegel, B.E. Res 26: 1585–1593. 1991. Stem profile equations for southern tree Goelz, J.C.G., and Leduc, D.J. 2001. Long-term species. Research Paper SE-282. USDA Forest studies on development of longleaf pine planta- Service, Southeastern Forest Experiment Station, tions. In Proceedings of the Third Longleaf Alliance Asheville, NC. Regional Conference, Forests for Our Future, ed. J.S. Clutter, J.L., Fortson, J.C., Pienaar, L.V., Brister, Kush, pp. 116–118. Alexandria, LA, October 16– G.H., and Bailey, R. L. 1983. Timber Management: 18, 2000. Auburn, AL: Longleaf Alliance, Auburn A Quantitative Approach. New York: John Wiley & University. Sons. Goelz, J.C.G., and Leduc, D.J. 2003. A model for Davis, K.P. 1966. Forest Management: Regulation and growth and development of longleaf pine plan- Valuation, 2nd edition. New York: McGraw–Hill. tations. In Proceedings of the Fourth Longleaf Al- Dean, T.J., and Baldwin, V.C., Jr. 1993. Using a liance Regional Conference, ed. J.S. Kush. Southern density-management diagram to develop thin- Pines, NC, November 17–20, 2002. Auburn, AL: ning schedules for loblolly pine plantations. Longleaf Alliance, Auburn University. Research Paper SO-275. USDA Forest Service, Goelz, J.C.G., and Leduc, D.J. 2004. Reproducibil- Southern Forest Experiment Station, New Or- ity and reliability: How to define the population leans, LA. of trees that represent site quality for longleaf Donnelly, D., Lilly, B., and Smith, E. 2001. The pine plantations. In Proceedings of the Twelfth Bi- southern variant of the forest vegetation simu- ennial Southern Silvicultural Research Conference, ed. lator. USDA Forest Service, Forest Management K. Connor, pp. 189–195. Biloxi, MS, February Service Center, Ft. Collins, CO. 24–28, 2003. General Technical Report SRS-71. Drew, J.T., and Flewelling, J.W. 1979. Stand den- USDA Forest Service, Southern Research Station, sity management: An alternative approach and NC: Asheville. its application to Douglas-fir plantations. For Sci Grelen, H.E., and Lohrey, R.E. 1978. Herbage yield 25: 518–532. related to basal area and rainfall in a thinned 266 III. Silviculture

longleaf plantation. Research Note SO-232. (Pinus palustris Mill.). Am Midl Nat 131(4):491– USDA Forest Service, Southern Forest Experi- 525. ment Station, New Orleans, LA. Quicke, H.E., and Meldahl, R.S. 1992. Predicting Harrington, C.A. 1990. PPSITE—A new method of pole classes for longleaf pine based on diameter site evaluation for longleaf pine: Model develop- breast height. South J Appl For 16:79–82. ment and user guide. General Technical Report Quicke, H.E., Meldahl, R.S., and Kush, J.S. 1994. SO-80. USDA Forest Service, Southern Forest Ex- Basal area growth of individual trees: A model de- periment Station, New Orleans, LA. rived from a regional longleaf pine growth study. Hawes, E.T. 1947. A method of determinating For Sci 40(3):528–542. southern pine pole classes from d.b.h. J For 45: Quicke, H.E., Meldahl, R.S., and Kush, J.S. 1997. 204–205. A survival rate model for naturally regener- Haywood, J.D., Boyer, W.D., and Harris, F.L. 1998. ated longleaf pine. South J Appl For 21(2):97– Plant communities in selected longleaf pine land- 101. scapes on the Catahoula Ranger District, Kisatchie Reineke, L.H. 1933. Perfecting a stand-density in- National Forest, Louisiana. In Proceedings of the dex for even-aged forests. J Agric Res 46:627– Ninth Biennial Southern Silvicultural Research Con- 638. ference, ed. T.A. Waltrop, pp. 86–91. Clemson, SC, Saucier, J.R., Phillips, D.R., and Williams, J.G., Jr. February 25–27, 1997. General Technical Report 1981. Green weight, volume, board-foot, and SRS-20. USDA Forest Service, Southern Research cord tables for the major southern pine species. Station, Asheville, NC. Georgia Forest Research Paper No. 19. Georgia Johnson, R., and Gjerstad, D. 1999. Restoring the Forestry Commission, Research Division. longleaf pine forest ecosystem. Alabama’s Trea- Schwarz, G.F. 1907. The Longleaf Pine in Virgin For- sured Forests Fall: 18–19. est: A Silvical Study. New York: John Wiley & Lohrey, R.E., and Bailey, R.L. 1977. Yield tables Sons. and stand structure for unthinned longleaf pine Shoulders, E., and Tiarks, A.E. 1980. Predicting plantations in Louisiana and Texas. Research height and relative performance of major south- Paper SO-133. USDA Forest Service, South- ern pines from rainfall, slope, and available mois- ern Forest Experiment Station, New Orleans, ture. For Sci 26:437–447. LA. Shoulders, E., and Walker, F.V.1979. Soil, slope, and Mesavage, C., and Girard, J.W. 1946. Tables for Es- rainfall affect height yield in 15-year-old southern timating Board-Foot Volume of Timber. Washington, pine plantations. Research Paper SO-153. USDA DC: USDA Forest Service. Forest Service, Southern Forest Experiment Sta- Means, D.B. 1996. Longleaf pine forest, going, go- tion, New Orleans, LA. ing, . . . . In Eastern Old-Growth Forests: Prospects for Smith, W.R., Farrar, R.M., Jr., Murphy, P.A., Yeiser, Rediscovery and Recovery, ed. M. B. Davis, pp. 210– J.L., Meldahl, R.S., and Kush, J.S. 1992. Crown 229. Washington, DC: Island Press. and basal area relationships of open-grown Meldahl, R.S., Kush, J.S., Rayamajhi, J.N., and southern pines for modeling competition and Farrar, R.M., Jr. 1997. Productivity of natural growth. Can J For Res 22:341– 347. stands of longleaf pine in relation to competition Spurr, S.H. 1952. Forest Inventory. New York: Ronald and climatic factors. In The Productivity and Sus- Press. tainability of Southern Forest Ecosystems in a Chang- Taras, M.A., and Clark, A., III. 1977. Aboveground ing Environment, ed. R. A. Mickler and S. Fox, pp. biomass of longleaf pine in a natural sawtimber 231–254. New York: Springer-Verlag. stand in southern Alabama. Research Paper SE- Murphy, P.A., and Farrar, R.M., Jr. 1985. Growth 162. USDA Forest Service, Southeastern Forest and yield of uneven-aged shortleaf pine stands in Experiment Station, Asheville, NC. the interior highlands. Research Paper SO-218. Thomas, C.E., Parresol, B.R., Le, K.H.N., and USDA Forest Service, Southern Forest Experi- Lohrey, R.E. 1995. Biomass and taper for trees in ment Station. thinned and unthinned longleaf pine plantations. Parker, R.C. 1998. Field and computer application South J Appl For 19:29–35. of Mesavage and Girard form class volume tables. USDA. 1929 [revised 1976]. Volume, yield, and South J Appl For 22: 81–87. stand tables for second-growth southern pines. Platt, W.J., Evans, G.W., and Rathbun, S.L. 1988. USDA Miscellaneous Publication 50. Washington, The population dynamics of a long-lived conifer DC. 8. Longleaf Pine Growth and Yield 267

USDA. 1984. Tables of whole-tree weight for se- Wiant, H.V., Jr. 1986. Formulas for Mesavage and lected U.S. tree species. General Technical Report Girard’s volume tables. North J Appl For 13:147– WO-42. USDA Forest Service, Washington, DC. 148. Wahlenberg, W.G. 1946. Longleaf Pine: Its Use, Ecol- Wiant, H.V., Jr., and Castaneda, F. 1977. Mesavage ogy, Regeneration, Protection, Growth, and Manage- and Girard’s volume tables formulated. Resource ment. Washington, DC: Charles Lathrop Pack Inventory Notes BLM 4. USDI, Bureau of Land Forestry Foundation. Management. Denver Service Center. Section IV Restoration

269 Chapter 9 Restoring the Overstory of Longleaf Pine Ecosystems

Rhett Johnson and Dean Gjerstad

Introduction Less well quantified but no less appre- ciated are the ecological benefits of lon- Restoring longleaf pine trees to the southeast- gleaf forests, particularly when they are re- ern landscape is a daunting task, because more stored to ecological function (Frost 1990). The than 97% of the original area has been lost grassy and herbaceous understory is rich in to other uses (Landers et al. 1995; Frost this species, some present only in fire-maintained volume). However, many of the disincentives longleaf ecosystems (Brockway and Lewis and difficulties in managing for longleaf pine 1997; Glitzenstein et al. 1998; Kush et al. have been addressed and solved or exposed 1999a). Wildflowers abound, particularly in as misconceptions, and landowners across the the fall, and are attended by scores of butter- region are expressing renewed interest in re- flies. Wildlife responds well to these forests. turning this once-dominant southern pine to Bobwhite quail (Colinus virginiana) do partic- their lands. Several recent publications provid- ularly well in longleaf forests, as do fox squir- ing information to landowners and natural re- rels (Sciuris niger). Wild turkeys (Meleagris galla- source managers on longleaf pine restoration pavo), deer (Odocoileus virginiana), and gopher and management have appeared (Earley 1997, tortoises (Gopherus polyphemus) find good habi- 2002; Franklin 1997; Kush 1997, 1999, 2001, tat there and many songbird species are reg- 2003; Johnson 1999; Mitchell et al. 2000). ular residents. Several threatened or endan- The economic and silvicultural benefits of gered species of birds, reptiles, and plants are longleaf pine are generally well-known: high- dependent on burned longleaf forests for their quality products; fire tolerance; insect and dis- existence. ease resistance; wind firmness; the ability to Interested landowners face two distinct but grow and thrive on harsh sites; and the abil- related tasks in their efforts to restore longleaf ity to respond to thinning at virtually any time pine to their lands. One, the establishment of in its long life (Johnson 1999). The burgeoning the trees themselves, is challenging but there pine straw market promises to yield even more are tested and reliable strategies for most sit- financial benefit to landowners who manage uations (Franklin 1997; Boyer 1999; Hainds their forests for its production. 1999, 2001, 2003a,b). The second, restoration

r Rhett Johnson Solon Dixon Forestry Education Center, School of Forestry and Wildlife Sciences, Auburn University, r Andalusia, Alabama 36420. Dean Gjerstad School of Forestry and Wildlife Sciences, Auburn University, Auburn, Alabama 36849.

271 272 IV. Restoration of the entire longleaf pine forest community, ful regeneration. Seed loss to predation and or at least enough of it to provide all of the seedling mortality following seed germination functions necessary to provide the benefits that make it necessary to sow from 1.7 to 3.4 kg make it unique, is considerably more difficult of seed per hectare (Mann 1970; Barnett and and may be very expensive. It will certainly be Baker 1991). In addition, to prevent exces- a long-term effort. sive predation by birds and rodents, the seed This chapter describes 10 forest conditions must be treated with repellents. Successful re- common to the longleaf pine region today generation using this technique is dependent and suggests ways to begin the overstory on adequate rainfall in the days and weeks restoration process. These situations, in gen- following sowing, otherwise germination will eral order of the level of difficulty in restoring be poor and mortality of newly germinated ecosystem structure and function from most seedlings will be high (Williston and Balmer to least difficult, include: (1) agricultural fields 1971). and pastures; (2) “off-site” upland hardwoods; (3) abandoned cutover forest land; (4) mixed Seedlings hardwood–pine or pine–hardwood forest; pine stands without a significant longleaf compo- In situations where there are no residual lon- nent and (5) no fire history or (6) with a recent gleaf pines to provide seed, the most viable op- history of fire; pine stands with greater than tion is to plant longleaf seedlings. There are 4.5 m2 basal area (ba) per hectare of longleaf two types of seedlings available on the mar- pine in the overstory and (7) no recent fire his- ket today, bareroot and containerized (Barnett tory or (8) with a recent history of fire; stands et al. 1990). Both types of seedlings can be that are predominately longleaf pine and (9) planted successfully; however, a regional sur- with no recent fire history or (10) with a re- vey of practitioners found that average survival cent history of fire. was 20% higher for container stock (85%) than for bareroot stock (Boyette 1996). Bareroot seedlings are grown for approxi- Regeneration Options mately 1 year from seed in nursery beds and then lifted with no soil on the roots immedi- Artificial Regeneration ately before planting (Mexal and South 1991). Direct Seeding The lifting process causes most fine roots to be lost and these fine roots must be regener- The success of planting from seed, called “di- ated after outplanting to enable the seedling rect seeding” by foresters, depends on creat- to survive. Left unchecked in the nursery bed, ing a well-prepared seedbed, sowing sufficient longleaf lateral and taproots can grow to 1 m quantity of seed and having favorable weather or longer in length, making them unwieldy conditions for seed germination and seedling to transplant. Thus, roots are usually pruned early growth (Barnett and Baker 1991). Cre- several weeks prior to lifting to produce a ating a well-prepared seedbed can range from more compact root system and to stimulate a low-cost prescribed fire to intensive mechan- fine root growth (Mexal and South 1991). The ical and chemical site preparation for sites with pruned roots are easier to plant and the resid- thick sod or heavy hardwood competition. The ual fine roots will eventually regenerate new purpose of site preparation is to provide ex- roots when outplanted. Root pruning should posed mineral soil for seed germination and only be done by trained personnel in the nurs- to reduce competing vegetation that can im- ery seedbed and never by tree planters prior pact the survival and early growth of seedlings to outplanting. Typically, root pruning by tree (Lowery and Gjerstad 1991). A major deter- planters results in a significant increase in rent to direct seeding is that the cost of seed seedling mortality (Wakeley 1953; Mexal and necessary for successful regeneration is higher South 1991). Characteristics of quality bare- than the cost of seedlings required for success- root seedlings include healthy green foliage, 9. Restoring the Overstory 273 root collar diameters of at least 12 mm, and firm plugs, and light brown or white root tips several healthy lateral roots. (indicating root growth) visible along the sides Containerized seedlings are grown in a vari- of the plug (Barnett et al. 2002; Hainds and ety of hard-walled vessels or in peat pots from Barnett 2003). seed. The entire plant, with root mass intact in the potting medium, is removed from the container, placed in shipping boxes, and shipped to Site Preparation Alternatives the planting site (Brissette et al. 1991). In this Preparation of a site for planting can take many case, the fine feeder roots are intact at the forms and vary considerably in intensity of time of outplanting and frequently begin root site disturbance (Table 1). Primary reasons to growth immediately. Characteristics of qual- perform site preparation include better plant- ity containerized seedlings include healthy fo- ing access and control of unwanted vegeta- liage, root collar diameters of at least 6 mm, tion, thus enhancing the survival and early

TABLE 1. Advantages and disadvantages of mechanical and chemical site preparation techniques. Mechanical Technique Advantages Disadvantages

Bulldozing 1. Complete vegetation removal 1. Short-term vegetation control (windrow, rootrake) 2. Easy to administer 2. Negative impact on herbaceous plants 3. Machine plant 3. High cost 4. Potential loss in soil productivity and erosion 5. Loss of area in windrows 6. Potential water quality and sedimentation problems Shear and pile 1. Less soil and site disturbance, therefore 1. Short-term vegetation control less potential for erosion and site 2. High cost degradation 2. Easy to administer 3. Machine plant Disking 1. Easy to administer 1. High cost 2. Effective for natural regeneration 2. Potential erosion on steep slopes 3. Machine plant 3. Short-term vegetation control 4. Negative impact on herbaceous plants Bedding 1. Improved survival on wet sites 1. Expensive 2. Machine plant 2. Short-term vegetation control 3. Easy to administer 3. Negative impact on herbaceous plants 3-in-1 (combines plowing, 1. Easy to plant 1. Expensive subsoiling, and bedding 2. Concentration of organic matter in bed 2. Competing vegetation between in one pass) tincreases soil fertility near seedling beds not affected 3. Reduces woody plant competition in 3. Likely severe impact on bedded area herbaceous plants in bedded zone 4. Beds need to settle prior to planting Drum chopping 1. Fewer water quality and erosion problems 1. Considerable hardwood resprouting 2. Minimal impact on herbaceous plants commonly occurs 3. Machine plant (if obtain good burn) 2. Must be used in combination with 4. Usually provides for a good ﬁre ﬁre to clear debris 5. Easy to administer 3. Short-term vegetation control

(Cont.) 274 IV. Restoration

TABLE 1. (Continued ) Mechanical Technique Advantages Disadvantages Sub soiling (ripping) 1. Breaks up plow pan allowing improved 1. Do not plant in rip as seedlings drainage and root growth tend to sink resulting in mortality; plant to the side of the rip Scalping 1. Removes sod, pest, and pathogens in 1. Must follow terrain to avoid erosion planting zone on old agricultural ﬁelds 2. Low cost 3. Easy to administer Burning 1. Low cost 1. Temporary air pollution 2. Positive impact on herbaceous plants 2. Short-term vegetation control 3. Reduction of fuels 3. Possibility of escape 4. Controls many small woody plants 4. Normally used in combination with other methods Total utilization 1. If done correctly, little needed 1. Rarely accomplished for site prep. 2. Easy to plant 2. Short-term vegetation control 3. Often considerable hardwood resprouting

Chemical 1. Lower capital expenditures 1. Narrow window of application 2. Higher productivity (more acres per day) (day/season) 3. Less resprouting of hardwoods 2. Must train personnel 4. Can be used on more rugged terrain 3. Need follow-up burn to facilitate 5. Less soil disturbance planting 6. Growth response of crop trees 4. Controversial 5. May reduce herbaceous cover and species richness 6. Difficult to administer growth of newly established seedlings (Low- In situations following a timber harvest, ery and Gjerstad 1991). In pastures and agri- coarse woody debris can be a problem for tree cultural fields, scalping and subsoiling provide planters. Treatments to reduce or redistribute some measure of both. In improved pastures, tree tops and other debris may be used to a herbicide treatment is recommended prior to make the site more accessible for replanting scalping to control introduced grasses like fes- (Lowery and Gjerstad 1991). Controlling un- cue (Festuca arundinacea), Bermuda grass (Cyn- wanted vegetation on the site prior to planting odon dactylon), and Bahia grass (Paspalum nota- will improve access and reduce competition for tum). These grasses must be controlled prior to the new stand. To that end, techniques such as planting because their well-developed peren- chopping, subsoiling, bedding plows, and her- nial root systems will outcompete longleaf bicides, all used alone or in combination and in seedlings for soil water and result in poor lon- conjunction with fire, can damage rootstocks gleaf pine survival. Additionally, there are no of residual woody vegetation and allow the effective post plant treatments to control pas- new stand greater access to the site’s resources ture grasses. Subsoiling or ripping is intended (Lowery and Gjerstad 1991). Subsoiling and to break up the hardpan or plow pan common bedding treatments can prepare a good root- to many agricultural soils and allow drainage ing environment for the new trees as well as and root growth. Scalping can provide a weed- control competing vegetation. Herbicides pro- free zone immediately after planting, lessen vide cost-effective control of unwanted veg- disease and insect damage potential, and im- etation prior to planting in practically every prove soil moisture conditions. situation. Modern herbicides are commonly 9. Restoring the Overstory 275

TABLE 2. Herbicides commonly used to manage longleaf pine forests. Common name of Herbicide active ingredient family Usesa Species controlled Resistant species Glyphosate Unclassified SP, CS Most plants when absorbed Plants not receiving foliar through foliage spray Imazapyr Imidazolinone SP, CS, release, Grasses, broadleaf, Legumes, pines, elms, HWC hardwoods blackberry, wax myrtle Triclopyr Pyridine SP, CS Hardwoods, broadleaf Grasses weeds Sulfometuron Sulfonyl urea SP, HWC Many grasses and broadleaf Andropogon sp., Bermudagrass, methyl weeds trumpet creeper Hexazinone Triazine SP, CS, BS, Hardwoods and herbaceous Some grasses, Vaccinium sp. release, species HWC Picloram Pyridine SP, CS Hardwoods, pine, broadleaf grasses weeds a SP = site preparation; CS = cut surface; BS = basal soil; release = controlling hardwoods in pine stands; HWC = herbaceous weed control. selective in their effect on various plant species a good inventory of what species are present (Table 2). For example, some herbicides selec- prior to choosing a chemical. tively kill grasses (sethoxydim) while others Chopping, a mechanical treatment which (imazapyr) do not kill legumes and still oth- employs a heavy ribbed metal drum pulled ers (triclopyr, hexazinone) kill only broadleaf across the site, causes minimal soil disturbance plants (Minogue et al. 1991). This trait allows and seedbed disruption, and has little impact knowledgeable managers the opportunity to on herbaceous plants, but tends to stimulate almost surgically extract undesired species and sprouting of woody growth if no other treat- create room for desirables. The choice of herbi- ment is applied (Lowery and Gjerstad 1991). cides or prescription should be made by a trained An effective follow-up treatment to control professional to best suit the vegetation to be woody sprouts is a second chop in late summer controlled. or early fall when root carbohydrate reserves By the same token, intensive site prepa- are low. ration has the potential to destroy desirable As in most forest management situations, components of the understory or to hamper cost is a factor in any decision and will vary their recovery (Bengston et al. 1993). If un- with the intensity of site preparation. Cost- derstory restoration is a primary consideration effectiveness can only be determined if desired for a landowner, a balance must be struck be- outcomes and priorities are clearly understood tween maximizing site preparation intensity to and stated. increase tree growth and survival and minimizing the impact on other ecosystem compo- Planting nents. Optimally, fire alone does little to retard understory development and, in fact, is likely Longleaf pine can be either hand planted or to enhance it. In most situations where sites are planted with a planting machine. No matter being converted back to longleaf pine, how- which method is used, planting depth is a ever, fire is not sufficient to allow optimal lon- critical factor in determining the survival and gleaf pine growth and survival. Selective herbi- early growth of longleaf seedlings. Research cides, used in conjunction with fire, offer some conducted by The Longleaf Alliance (Hainds of the benefits of site control while sparing 2003a,b) has indicated that planting container- some of the desirable plant community. The ized seedlings with the terminal bud and root landowner/manager should have a clear idea collar slightly above the soil surface results in of what understory species are desirable and excellent survival. Seedlings planted with the 276 IV. Restoration terminal bud covered by soil suffer high mor- managers use late planting to replace mortality tality and slow growth. Maintaining an optimal from earlier plantings. planting depth is particularly important with Planting with bareroot seedlings generally containerized seedlings and may be difficult calls for more care than is required for con- to obtain with a planting machine. It is eas- tainerized seedlings as success is more depen- ier to hand plant containerized seedlings than dent on uncontrollable factors such as weather bareroot seedlings. Bareroot seedlings should and the window for planting is much nar- be planted so that the root collar and termi- rower. Bareroot seedlings must regrow the fine nal bud are right at or within 6 mm of the feeder roots lost in the lifting process before soil surface after the soil settles (Barnett et al. seedling growth can take place. Experienced 1990). Machine-planting usually results in a individuals and organizations can be success- slight berm around the seedlings created by the ful planting bareroot seedlings, but the mar- packing wheels. That berm typically subsides gin of error is smaller than with container- over time, so that seedlings planted using a ma- ized seedlings and any misstep can lead to low chine should be planted with the terminal bud survival and/or poor early growth. Cost can slightly below the soil berm surface. Most lon- also be a factor in the choice of seedling type. gleaf planters find it easier and more efficient Currently, bareroot seedlings cost about 40– to plant bareroot seedlings by machine because 50% less than containerized seedlings. How- of the long coarse lateral roots found on many ever, the cost of seedlings is a small part of the longleaf seedlings (Barnett et al. 1990). Hand- total cost of reforestation and replanting after a pruning these roots leads to high mortality and failure can be very expensive. Simply put, con- stuffing them into the small slit typical of hand tainerized seedlings cost more, but offer insur- planting can cause root binding or “J-rooting.” ance against failure as well as improved early Most tree planters recommend that as much growth. of the root mass that will fit be inserted into the planting hole, and the remainder left out to Release Treatments air prune. Because containerized seedling roots are in a compact plug, that problem is avoided. Because longleaf responds poorly to compe- Planting with bareroot seedlings should take tition, it is often desirable to “release” it by place any time in the late fall or early winter controlling undesirable vegetation. The most when the soil is moist. It is generally prefer- important competitors for longleaf seedlings able to plant in November or December, but typically are grasses. Aggressive grass compe- all planting should be complete before March tition can cause high seedling mortality and (Long 1991). Bareroot longleaf pine seedlings retard growth of surviving seedlings (Hainds do not store as well, even in cold storage, as do 2003b). The longer seedlings stay in the grass other pines and should be planted as quickly stage, the more susceptible they are to brown as possible after lifting (Barnett et al. 1990). spot needle blight (Mycosphaerella dearnessii), Containerized seedlings can be planted at any an important disease of longleaf pine seedlings. time of the year as long as there is moisture in Woody competition has a more limited effect the soil and can begin growing immediately on on early survival of longleaf pine seedlings, outplanting (Hainds 2003b). Very good results but can also retard growth. Woody compe- have been obtained by planting in November tition can be controlled after the longleaf is and December, giving the seedlings a chance planted with fire, herbicides, or a combina- to develop a root system while competitors are tion of both. Grasses and forbs, on the other dormant. This head start can be very impor- hand, are generally only top-killed by fire tant in the spring, when grasses, forbs, and and can rebound aggressively. The most im- woody competitors begin early growth. Plant- portant competition takes place below ground ing containerized seedlings in the late spring for growing space, soil moisture, and nutri- and summer has also proven to be successful ents. Burning grasses does little to reduce as long as there is adequate soil moisture. Some that competition. Treating herbaceous plants 9. Restoring the Overstory 277

FIGURE 1. Shelterwood systems employ a series of a cuts over several years and can result in even- or two-aged stands. Trees marked with an “X” are to Pines be harvested in the indicated cutting cycle. a. The first cut removes nonlongleaf pines unless Hardwoods they are needed to provide fuel for prescribed fires and all hardwoods. Approximately 11.5–16 m2/ha of basal area of well-distributed longleaf pine should be left across the site. b. The second cut comes after several years and several fires to control underbrush. This cut will leave a residual basal area of 5.5–7 m2/ha of well- b distributed, well-formed longleaf pines, all proven cone and seed producers. X X c. A winter or spring burn before seed fall a year X X when there is a good cone crop prepares a seed X bed for the new germinants. If stocking is adequate (greater than 11,000/ha), the seed trees can be re- X X moved in the second or third year following seed fall X to yield an even-aged stand. Seed trees should be re- X moved while the seedlings are in the grass stage. If X X a two-aged stand is desired, some or all of the seed c trees may be retained. Fire should be excluded for the first year after seed fall.

with herbicides provides seedlings with a diameter at breast height (DBH) and preferably competition-free rooting zone, allowing them 40 cm or larger before being considered as re- to establish healthy and competitive root sys- liable seed producers. If enough longleaf trees tems. Selective herbicides can be used to favor of sufficient size are present and evenly dis- some plants while controlling others. Banded tributed across the site, natural regeneration is applications only treat the immediate area an option. The longleaf shelterwood system, as around the seedling rows, preserving desirable developed at the Escambia Experimental For- native vegetation between bands. The effec- est in Alabama, is a modified seed tree sys- tive period for most of the chemicals used for tem adapted to allow for the heavier seed of release is one growing season or less. longleaf pine (Croker and Boyer 1975; Den- nington 1990). In this shelterwood system, a longleaf stand is first thinned to a basal area Natural Regeneration (BA) of about 12 m2/ha or about 125–150 Even-Aged Systems trees at least 30 cm in diameter per hectare (Fig. 1; Boyer and White 1990). The resulting Where adequate seed sources exist, longleaf stand is left to grow while the tree canopies pine may be regenerated from natural seed expand into the gaps between them. Approx- (Dennington and Farrar 1983; Barnett and imately 5 years later, a second cut is made to Baker 1991). Trees should be at least 30 cm in reduce the number of trees to 60–75 trees/ha 278 IV. Restoration

700 FIGURE 2. Stand structures and appearances of traditional All-Aged Stand Even-Aged Stand uneven-aged and even-aged ma- 600 ture longleaf stands are shown. Regulated uneven-aged stands 500 have a stem diameter distribution curve that resembles a reverse “J,” while even-aged stands 400 typically have a bell-shaped /ha distribution curve. 300 Stems/

200

100

0 5 1015202530354045505560 DBH (cm)

(BA of 5.5–7 m2/ha). The residual trees should overstory removal. Generally, 2500 to 3500 be selected for desirable qualities, because they seedlings/ha in height growth and free of over- will provide the seed for the new forest. Dur- head competition are considered a successfully ing this period, prescribed fire should be used established stand. Over time such stands will frequently to keep the woody brush in the un- develop into an even-aged stand with bell- derstory under control. Longleaf cones take shaped diameter distribution (Fig. 2). 2 years to mature, so it is possible to anticipate good seed crops by monitoring flower Uneven-Aged Systems production more than a year before actual seed fall (Croker 1971). Failing that, green Uneven-aged longleaf pine stands with a re- cones are easily spotted in the late spring verse J diameter distribution can be created and early summer of the actual seed year. through several approaches (Figs. 2 and 3). If It is recommended that an average of 2500 an uneven- or all-aged stand is desired, the cones/ha be present to achieve successful re- quickest and easiest approach is a modified generation. With each cone producing approx- shelterwood system where all or a portion of imately 50 seeds, 2500 cones will produce the overstory trees are retained. This approach nearly 125,000 seeds/ha. Typically, 90% of the will create a two-aged stand with the first seed seed and newly germinated seed are lost to pre- fall and, over time and successive seed crops, dation by wildlife or die due to adverse envi- can come to contain several age classes. How- ronmental conditions. The desired goal is that ever, such stands are not commonly found on 7500 to 15,000 seedlings/ha be present before the landscape. The trade-off is reduced growth the overstory is removed. A prescribed burn of the new seedlings due to competition from in the winter or spring before seed fall will the overstory trees (Boyer 1993; Palik et al. greatly enhance seed “catch” and germination. 1997, 2003). Another option is to create gaps When the new stand is established and at least of 0.1–0.2 ha in the canopy across the stand 1 year old, the overstory should be removed by removing several trees in clumps (Fig. 4; to achieve maximum growth. Some seedlings McGuire et al. 2001). This approach mimics will be damaged by the logging activity, but in natural gaps resulting from overstory mortal- virtually every incidence more than adequate ity due to lightning strikes, insect infestations, numbers of seedlings will be present after the wind throw, and other causes of mortality. 9. Restoring the Overstory 279

pete vigorously with the new seedlings and can prevent successful establishment (McGuire et a al. 2001). If the gaps are too large, the heavy longleaf seed might not reach the middle and it can be very difficult to get fire to burn across the gap due to the lack of continuous pine needle litter to carry the fire. This can result in establishment of undesirable woody species in the gaps. In open stands, single-tree removal can create adequate openings for regeneration, but the removal of several trees in clumps at each entry may be more cost-effective and provide better sites for seedling growth (Fig. 5). An even-aged stand can be converted to an b all-aged or all-sized stand over a series of harvests using the BDq method (Fig. 6). The first harvest typically spares the smallest size classes and most of the largest except where stems exceed the predetermined diameter limit. Most of the harvest takes place in the middle diameters, where stem counts exceed those indicated by the desired distribution curve. Typ- ically, to retain the desired basal area, some of the larger trees are retained and removed in later cuts. Over several harvests and seed c crops, the stand distribution will begin to resemble the desired reverse J shape and the all- aged condition reached. A version of this technique is possible when there is no natural longleaf seed source if fire can be used in the existing stand. Cutting gaps into existing pine stands and planting with containerized seedlings can be successful. Ex- perience has suggested that at least minimal FIGURE 3. There are various approaches to uneven- site preparation and fairly large (greater than aged management. (a) Single tree selection, 0.1 ha) openings are necessary for satisfac- (b) group selection, and (c) modified shelterwood tory results. Again, repeating the process over cuts are visualized above. time with several cuts should result in an all- Again, fire must be used to retard invasion aged longleaf stand. This approach may result by undesirable hardwoods and woody brush. in growth reductions (Farrar and Boyer 1991; As the gaps seed in with longleaf, they are Boyer 1993), but does allow the retention of gradually enlarged by removing trees around mature trees on the site at all times. the margins. The gaps soon begin to resemble small “domes,” with the oldest, tallest, and fastest growing saplings in the middle and the Situations Requiring youngest, slowest growing, and smallest at the edges or margins of the openings. Over time Artificial Regeneration and successive seed crops and cuts, these gaps Agricultural Fields and Pastures merge into each other and trees representing many age classes occupy the site. If the Although old agricultural fields and pastures gaps are too small, the residual trees com- can appear to be “blank slates,” they typically 280 IV. Restoration

a FIGURE 4. Small group selection harvests can yield uneven-aged longleaf pine stands. Trees marked Pines with an “X” are to be removed. Symbol size indi- Hardwoods cates the relative size of the tree represented. X X X a. The first cut removes hardwoods and pines in X XX clusters of 0.1 to 0.2 ha in area and at least 30 m X X X wide. Seed-producing longleaf surrounding these X X X clusters are retained. Prescribed fire must be utilized to keep these openings free of hardwoods and pines other than longleaf until regeneration with b longleaf is achieved. b. After skipping a year of fire to protect the new seedlings, burning is resumed. Mature trees adjoin- ing the openings may suppress growth of the new seedlings and may be removed. c. Another series of clusters can be selected and removed on the desired cutting cycle, following the same regime as above. The seedlings from the first clusters are now small saplings. d. Continued removal of mature trees in small c groups or clusters will eventually release the X X X X X younger trees, the first of which are now entering X X X XX X larger size classes. An all-aged stand or forest is developing, with trees of several age and size classes.

are the most difﬁcult of all sites to restore Preparing the site for planting is critical. If to longleaf pine ecosystem function (Hainds pasture grasses are present, they should be 2003b). The pasture grass and agricultural killed before planting is attempted (Hainds weed complex is aggressive and most of these 2003b). Killing grasses like Bermudagrass, fes- sites have little or no residual seed bed from the cue, and Bahiagrass is difﬁcult and usually re- early native forest. Reforestation, of course, quires herbicides. An effective technique in- must be through planting. Direct seeding tech- cludes mowing, allowing the grass to grow niques are seldom successful. If seed is broad- back to about 15 cm in height, and then ap- cast, germination is uncertain and predation by plying glyphosate, imazapyr, or a combina- birds and small mammals can be severe. Plant- tion of the two herbicides. Glyphosate can be ing individual seeds by hand or drilling the seed applied anytime the grass is growing. Imaza- into the soil is labor intensive and success is pyr is most effective in late spring or early usually spotty. summer. Follow-up spot treatments will likely 9. Restoring the Overstory 281

FIGURE 5. Single tree selection harvests can also yield a an all-aged condition, although growth of younger age X X X classes is likely suppressed. Trees marked with an “X” X X Pines are to be harvested. Symbol size represents tree size. X X X Hardwoods a. Individual trees are selected for removal through- X X X out the stand, choosing suppressed trees, trees with X bad form, and trees of species other than longleaf for X X X harvest. Some hardwoods may be retained for wildlife X X purposes and some nonlongleaf pines may spared to X provide continuous fuel for prescribed ﬁre. b. Seedlings should accumulate in small patches b throughout the stand, particularly in the small gaps created by the earlier harvest. A second round of tree removals based on tree quality, position in the canopy, and stand density may be performed. c. The thinning process continues on regular intervals, retaining the most desirable individuals and removing some to release nearby seedlings and saplings. d. Prescribed ﬁre is continuously applied. Thinning continues to be used to improve stand quality and to free established regeneration from competition. An c all-aged stand comprised of trees of several age and size classes is developed. X X X X

X X X

be necessary to control these tough competi- follow contours when scalping. Soil movement tors, especially Bermudagrass. along the scalped furrow can not only doom In both pastures and old fields, the next step the planting effort but damage water quality is to scalp the site (Hainds 2001). Scalping peels and ruin the productivity of the field for future back the upper 5–7.5 cm in old fields and 7.5– generations as well. Unless the site has been 10 cm in sod pastures in a furrow 60–90 cm subsoiled within the past 2 years, it is usually wide. Scalping implements are available com- a good idea to pull a ripper bar along the fur- mercially or may be constructed from modi- row to break up plow pans caused by previous fied fire plows. There are commercial vendors agricultural tillage. Depth of the rip should be in many areas who will provide scalping ser- from 45 to 60 cm. If planting is to be done using vices. It is extremely important to rigorously a mechanical planter, it is recommended that 700 700

600 600 Stand table after 5th cut 500 500

400 400

300 300

200 200

100 100

0 0 5 1015202530354045505560 5 1015202530354045505560

700 700 All-Aged Stand 600 600 Even-Aged Stand Stand table after 6th cut 500 500

400 400

300 300

200 200

100 100

0 0 5 1015202530354045505560 5 1015202530354045505560 700 700

Stand table after 7th cut 600 Stand table after 1st cut 600

500 500

400 400

300 300

200 200

100 100

0 0 5 1015202530354045505560 5 1015202530354045505560 700 700

600 Stand table after 2nd cut 600 Stand table after 8th cut /ha

500 500

400 400 Stems/ 300 300

200 200

100 100

0 0 51015 20 25 30 35 40 45 50 55 60 5 1015202530354045505560 700 700

Stand table after 3rd cut Stand table after 9th cut 600 600

500 500

400 400

300 300

200 200

100 100 0 0 5 1015202530354045505560 5 1015202530354045505560

700 700

600 Stand table after 4th cut 600 Stand table after 10th cut

500 500

400 400

300 300

200 200

100 100

0 0 51015202530354045505560 5 1015202530354045505560 DBH (cm)

FIGURE 6. An even-aged stand can be converted to an all-aged or all-sized stand over a series of harvests using the BDq management method. Existing and desired stand distribution curves are developed and superimposed. The first harvest typically spares the smallest size classes and most of the largest except where stems exceed the predetermined diameter limit (D). Most of the harvest takes place in the middle diameters, where stem counts exceed those indicated by the desired distribution curve. Typically, to retain the desired basal area (BA), some of the trees above the “line” are retained and removed in later cuts. Over several harvests and seed crops, the stand distribution should begin to resemble the desired reverse “J” shape and the all-aged condition reached. In this example, a cycle of 10 harvests is depicted. 9. Restoring the Overstory 283 the rip be oriented approximately 30 cm off should be at least 95 liters/ha. Before spraying, center in the scalped area. This is done so that one or two seedlings should be excavated and the seedlings may be planted in the center checked for new root growth. If the seedlings of the furrow and not in the ripped area. If have not begun to initiate new root growth trees are to be planted by hand, they should outside the plug, delay spraying until that oc- not be planted in the rip, but 15 cm or more curs. Also, areas where soil pH is high (greater to one side. This is because seedlings planted than 6.0) should not be treated with OustTM. in the ripped zone tend to sink several cen- OustTM is a root growth inhibitor, and high- timeters into the soil, resulting in mortal- pH soils tend to exacerbate its effectiveness. If ity. Scalping greatly improves survival in agri- root growth on new seedlings is retarded by the cultural areas (Barnard et al. 1995; Hainds chemical, April and May droughts may cause 2001). The furrow typically remains virtu- high mortality. ally weed free for much of the first grow- Another herbicide combination that has ing season, reducing competition for moisture proven successful in these situations is and nutrients. In addition, there is evidence ArsenalTM and OustTM (Hainds 2003b). that scalping reduces loss to a number of ArsenalTM should not be applied over longleaf pests and diseases, including white fringed seedlings earlier than 4–6 weeks after foliage beetle (Graphognathus spp.) grubs and char- growth initiation in the spring. One treatment coal root rot (Rhizoctonia spp.) (Barnard et al. combines an early treatment with 0.15 or 1995). 0.2 kg of OustTM per treated hectare in March Planting should be accomplished as soon in or April followed by a subsequent treatment, the fall as there is adequate moisture in the if needed, with 0.35 or 0.4 kg of ArsenalTM per soil (Hainds 2003b). All planting should be treated hectare in May or later. The ArsenalTM completed before Christmas if at all possible. and OustTM may be applied together in one Planting earlier in the fall gives the seedlings treatment 4–6 weeks after initiation of foliage a chance to initiate root growth before en- growth in the spring. Late germinants such as tering dormancy and gives them a head start crabgrass can prove disastrous to old-field and the following spring. It is recommended that pasture plantings if only the early treatment is planting be done with containerized seedlings used. in these situations. Recent research indicates Since these sites generally have no remnants that seedlings should be planted so that 1.2 of the native forest understory vegetation and to 2.5 cm of the plug is exposed, especially little or no seedbed to draw on, the use of in scalped areas (Hainds 2003b). Soil tends chemicals to control vegetation does little dam- to move into these areas and seedlings with age to the total restoration effort. Once the terminal buds covered by soil are not likely trees are established on agricultural sites, they to survive. In the spring following planting, can be burned almost immediately. In fact, it herbaceous weed control treatment might be may be difficult to use fire effectively on many desirable. One frequently recommended treat- of these sites because there are often few fine ment is 0.15–0.2 kg of OustTM per hectare in fuels on the ground to carry the fire. Old- March or April. The addition of 0.75 kg of Vel- field weeds like dog fennel (Eupatorium capillo- par DFTM per hectare gives broader spectrum folium), ragweed (Ambrosia spp.), and golden- control and may extend the effectiveness of the rod (Euthamia spp.) do not burn well and do treatment (Hainds 2003b). A premixed version not carry fire. Broomsedge (Andropogon spp.) of these chemicals, OustarTM, can also be ap- is a better fuel and will carry a fire if there plied at a rate of 0.75 kg per treated hectare. is a wind of 5 km/h or more. The trees are These treatments are applied directly over the fire tolerant throughout their life, but the stage top of the seedlings and generally in a band when they are 15 cm to 1 m tall is the most 1–1.5 m wide or in spots of 1.5-m diame- vulnerable. Trees in this size class should not ter over each seedling. In banded treatments, be burned during the growing season when total volume of spray (chemical plus water) they are “candling” (when the terminal bud 284 IV. Restoration is elongating and is tender and white). Young cutover site, site preparation can proceed. Be- longleaf may be burned successfully with strip cause hardwoods are vigorous resprouters, it headfires, although other types of fires may be is usually desirable to control regrowth with successful as well. Mowing between rows may a herbicide. These chemicals can be applied be helpful for access and to help carry fire, aerially, with ground-mounted sprayers, or by but care should be taken to avoid damaging hand with backpack sprayers. The chemical seedlings. may be broadcast, banded, or directed at sprout clumps. It is usually recommended that the site lay out for one growing season after harvest to Off-site Upland Hardwoods allow sufficient resprouting to aid herbicide ef- Many longleaf sites become occupied by hard- ficacy. wood species if clearcut or other heavy har- Choice of chemical should be based on vest is followed by fire suppression and no at- species to be controlled as well as species to fa- tempt is made at longleaf pine regeneration. vor (Cantrell 1985; Minogue et al. 1991). Her- Typically, light seeded species like sweetgum bicides differ in their mode of action and which (Liquidambar styraciflua) and loblolly pine (Pi- species they control. For instance, chemicals nus taeda) are the first tree species to colonize whose active ingredients include glyphosate the site. Without disturbance, particularly fire, (e.g., RoundupTM or AccordTM) are only fo- more shade-tolerant hardwood species come liar active and affect only the plants that are to dominate the site, in the process provid- sprayed. They control grasses and forbs as ing enough shade to preclude regeneration well as woody plants, but are much less ef- by the residual pines. These hardwood species fective on waxy-leafed plants. Products con- in the longleaf pine region typically include taining hexazinone as an active ingredient, on oaks (Quercus spp.) of various species, hicko- the other hand, are soil active. That is, they are ries (Carya spp.), and maple (Acer spp.). With- taken up through the plant’s root system and out further disturbance, the site will soon be work internally in the plant. Hexazinone prod- crowded with woody brush and mixed hard- ucts, like VelparTM, PrononeTM, and ULW TM, wood regeneration, resulting in a shaded for- are ineffective on many grasses like the com- est floor with little or no herbaceous vegeta- mon Andropogon species (e.g., broomsedge, tion. On mesic sites, this can be a fairly rich bluestems) and wiregrasses (e.g., Aristida spp., mix of species. On xeric or very dry sites, the Muhlenbergia spp.) at low rates (Brockway dominant species are typically scrub oaks like and Outcalt 2000). This may be an impor- turkey oak (Quercus laevis Walt.), bluejack oak tant factor where understory restoration is an (Quercus incana Bartr.), sand post oak (Quercus objective. Likewise, desirable wildlife species margaretta Ashe), and blackjack oak (Quercus such as American beautyberry (Callicarpa amer- marilandica Muench.). The effect, if left undis- icana) and many Vaccinium species (e.g., blue- turbed for long periods, is similar; no pines in berry, sparkleberry, huckleberry) are also tol- the overstory or understory, a mixed woody erant of hexazinone. Products like ArsenalTM midstory and understory, and very few grasses or ChopperTM, which contain imazapyr as or forbs. Restoring longleaf pine and longleaf an active ingredient, afford broad scale con- function to this forest type is slow, but can trol of many species, but have little effect on be accomplished. Longleaf pine restoration can nitrogen-fixing legumes which can also pro- only be accomplished by artificial means, i.e., vide important food for many wildlife species planting. (Minogue et al. 1991). Tank mixes of her- If the hardwoods on the site are of commer- bicides afford broader spectra of control, but cial size, the first step is to remove them with care should be taken to ensure that they are a timber harvest. Examination of the site af- compatible. In short, prescriptions for herbi- ter harvest will determine the presence of de- cides should be tailored carefully to fit both sirable understory components. If there is no species that are to be controlled and species significant desirable native vegetation on the that are to be spared. 9. Restoring the Overstory 285

In the case of mesic off-site hardwood con- able professionals. Planting with container- version, there is likely to be little desirable ized or bareroot longleaf pine seedlings can vegetation under the stand to protect and then usually proceed. Many native species typ- the choice of site preparation method is more ically have a diminished presence on these likely to depend on what species are to be con- sites and may recover without further assis- trolled prior to planting (Haywood et al. 2004). tance beyond prescribed fire, but depleted seed Herbicide labels are a good source of infor- banks may require planting of some desired mation on efficacy of the chemical on target species. species. It is advisable to obtain professional Mesic sites occupied by vigorous hardwood assistance, however, before choosing and ap- species can be restored to longleaf pine in a plying an herbicide. two-step process. After clearing the hardwoods Mechanical site preparation in combination through harvest and controlling the sprouts with fire is another choice for effective off- and other woody competitors with chemicals, site hardwood conversion (Table 1). Shear- a “nurse crop” of loblolly or slash pine might be ing, raking, plowing, and bedding in prepa- planted and fire introduced as early as possible ration for planting are commonly applied in the rotation, usually between ages 8 and 12, practices (Lowery and Gjerstad 1991). All of and kept in the system until final harvest of the these techniques, however, are likely to de- pines, either as pulpwood in 12–15 years or as stroy existing native ground cover and make solid wood products, with thinnings, at 25–30 its reestablishment more difficult (Bengston years. This will help control unwanted woody et al. 1993). This delay in herbaceous re- vegetation through the early shading provided covery can delay the reintroduction of pre- by the faster growing loblolly or slash pine and scribed fire. Chopping with a drum chopper provide pine fuel for subsequent fires. Keeping is effective where debris is light and can be fire in the system on a 2- to 4-year rotation, done with minimal impact on native under- with occasional growing season fires as part of story. These mechanical treatments typically the mix, should prepare the site for planting result in many hardwood sprouts from roots with longleaf and encourage any native species and stumps and may threaten the survival to assert themselves prior to establishment of and growth of longleaf pine seedlings unless the longleaf pines. controlled. In the case of scrub oaks (Quercus spp.) on xeric sites, mechanical treatments or herbi- Abandoned Cutover Forest cides used in combination with fire or fire alone will often stimulate vegetative growth Land by desirable grasses and forbs (Brockway and Cutover woodland sites typically do not have Outcalt 2000). Before longleaf pine can be a longleaf pine seed source present, so they reestablished, however, the site must be made must be regenerated by planting if longleaf accessible for the planters and the oaks must pine is to be reestablished. Depending on the be reduced in number to minimize compe- length of time since longleaf occupied the site, tition with longleaf pine seedlings and allow site preparation for planting can take various the development of herbaceous fuels to carry forms. If the site has been cut over and aban- later fires. These oaks are fairly fire-tolerant doned for several years, it is very likely to be and are best killed with a herbicide. Choices occupied by woody brush, hardwood seedlings include a hexazinone product applied as a and saplings, grapevine (Vitis rotundifolia) and broadcast or banded treatment, a directed or other vines, and even loblolly pine, an ag- broadcast treatment with an imazapyr prod- gressive pioneer species. There is typically lit- uct, or stem injection with a variety of chemi- tle herbaceous understory if the site has been cals. Again, choices should be based on existing cut over for very long. In this case, site prepa- competition and desirable vegetation and pre- ration is usually most successful if it is done scriptions should be prepared by knowledge- with herbicides. The chemical(s) chosen would 286 IV. Restoration depend on the species mix and the mode of ap- tems and are well suited to that forest type. plication on the density and height of the vege- For a variety of reasons, including optimizing tation. If herbicides are chosen for site prepara- ecological and economic value, many man- tion, they should be followed in a timely fash- agers desire to restore longleaf pine to these ion by fire to augment their effectiveness and sites. Doing so requires the planting of longleaf to make access for planting easier. Fire alone pine seedlings. One technique that has been is seldom effective in these situations due to a applied successfully begins with the commer- lack of fine fuels and the lack of total kill on cial harvest of the hardwood component, fol- hardwood rootstocks. If the cutover is recent, lowed by a growing season fire, preferably in then fire might be effective assuming there is late March through mid-May. Clearing logging sufficient fine fuel, usually grasses and forbs, slash away from the boles of the residual pines to carry the fire across the site. It is usually manually or mechanically is a wise precaution. desirable to follow up the fire with directed The intense heat these piles or tops can gen- herbicide treatments either before or after erate can damage or kill the roots and cam- planting. bium layer and scorch the crowns of nearby When the site is prepared for planting, trees (Swezy and Agee 1991; Hanula et al. containerized or bareroot longleaf seedlings 2002). Continuing to burn on a 2- or 3-year should be planted as early in the fall as fea- rotation, using growing season and dormant sible with the limiting factor being adequate season burns, should control the hardwood soil moisture (Hainds 2003b). A spring herba- sprouting and most of the pine regeneration. ceous release treatment with selective herbi- When the woody understory is diminished, cides carefully applied will enhance survival the next step is to remove the nonlongleaf and growth of the newly planted seedlings. pine component in a final harvest. Following Fire may be introduced as early as the first the harvest and prior to replanting, site prepa- year postplanting and a mixture of dormant ration is typically needed to remove unmer- and growing season fires should encourage ex- chantable hardwoods and pines. The method isting native vegetation (Walker 1998). Cau- of site preparation should be chosen with the tion is advised when the young seedlings are level of residual competition and the presence in the candling stage and before they are 1.5–2 or absence of desired ground-cover species in m tall. If the desired native plant community mind. If the fire has been effective, the woody does not appear within the first 2 or 3 years of competition should be minimal and any resid- burning, it may be necessary to reintroduce it ual ground cover should have begun to ex- or supplement remnant populations with seed press itself. Postplanting control of hardwood or seedlings (Glitzenstein et al. 1998). sprouts might be desirable and can be achieved by directed spray with foliar active herbicides or spot treatments with hexazinone products Mixed Hardwood–Pine or (Minogue et al. 1991). Restoration of understory vegetation is likely achievable through Pine–Hardwood Forest repeated use of fire, particularly growing sea- Many historical longleaf pine sites are occupied son fire, throughout the rotation unless the site today by forests composed of loblolly and/or was in cultivation for an extended period prior shortleaf pine (Pinus echinata Mill.) and mixed to becoming reforested (King et al. 1997). If hardwoods. In much of the natural longleaf desired understory recovery does not materi- range, this is the typical forest condition fol- alize within 5 years or less, it may be neces- lowing a total harvest if no attempt was made sary to artificially supplement it with seed or to reforest. These forests can produce valuable seedlings. timber, excellent wildlife habitat, and are of- Another approach to restore mixed pine– ten aesthetically pleasing. Still, they occupy hardwood stands is to leave a portion (no sites that once supported longleaf pine ecosys- more than 2.3 m2/ha of basal area (BA)) of 9. Restoring the Overstory 287 the loblolly, slash (Pinus elliottii) or shortleaf include mowing and/or herbicides prior to the pine stand intact after the pine harvest and burn to reduce “laddering” of fuels and get underplant with longleaf seedlings. This would fuels down onto the ground where they can create a two-aged stand and would help pro- deteriorate prior to the fire. In any event, if vide fuel for subsequent fires. There are sev- fuel buildup is a concern, the first burn should eral cautions, however. There is evidence that be a cool one. For example, the burn should even as little as 2.3 m2/ha BA can signifi- be in the dormant season on a cool day with cantly retard growth of the new stand (Boyer moderate humidity (30–50%), fuel moisture, 1993). More importantly, there may not be ad- and wind (e.g., 10–20 km/h) to carry the heat equate pine fuel to carry sufficiently hot fires out of the canopy quickly. The type of fire to control the loblolly, slash, or shortleaf pine can vary, but strip headfires with 20 m or seedlings resulting from this technique. These less between strips is a good compromise be- pine seedlings can be important competitors tween a slow backfire and hotter fires (Wade for the longleaf pine and choke out desirable and Lunsford 1988). On-site observations are understory vegetation if not controlled. The necessary to prescribe fire correctly. It may be likely outcome of this approach is the creation necessary to follow this initial fire with a sec- of a mixed species pine stand. ond dormant season fire within 2 years under similar conditions. The first fire will reduce fine fuels on the forest floor, but only top kill Pine Stands without a woody vegetation, creating a potential prob- Significant Longleaf Component lem for the next fire. A second dormant sea- There are many essentially pure stands of pines son fire will “knock down” and begin to con- of species other than longleaf pine on former sume some of that dead fuel. An assessment of longleaf pine sites. Some of these are the re- fuel conditions should be made after the sec- sult of earlier removal of longleaf pine and ond fire to determine if fuel conditions will al- recolonization by loblolly, shortleaf, or slash low a growing season fire. A thinning might be pine. Many more are the result of deliberate performed at this point to encourage ground replacement of longleaf pine by planting ei- cover plants. If a thinning is performed, plan- ther loblolly or slash pine. These sites must be ning for the next fire should take into con- reforested to longleaf by planting as there is no sideration the resulting slash and protect the available seed source. The task of restoring the residual trees (Haywood et al. 2004). After the ground cover is directly linked to the previous thinning and the first growing season burn, history of fire in the stand and prior land use fire should be continued on at least 3-year (e.g., agricultural uses) (Frost 1993). and preferably 2-year intervals for at least two cycles. Ideally, at least two successive growing season burns should be performed prior Pine Stands without Longleaf in to harvest of the overstory in preparation for the Overstory and No Recent reforestation because hardwood brush rootstocks are persistent and will occupy growing History of Fire space with overstory removal unless otherwise To restore the form and function of a lon- killed. gleaf pine ecosystem to this forest type re- The regeneration harvest can take two quires the reintroduction of fire into the sys- forms. The first approach is to clearcut the ex- tem first. If the period without fire has been isting stand and, after site preparation, replant long (greater than 10 years) and fuel accu- with longleaf seedlings. If burning has been ef- mulations are high (greater than 8000 kg/ha), fective, site preparation can be minimal, con- some type of fuel treatment may be neces- sisting of reducing or removing the woody de- sary to avoid damage to the existing stand bris to allow the planters access to the site. (Haywood et al. 2004). These treatments might Spot herbicide treatments might be desirable 288 IV. Restoration to control persistent woody clumps, but broad- Pine Stands without Longleaf in cast chemicals should not be necessary. Opti- the Overstory and with a Recent mal growth of the seedlings will require control of the herbaceous vegetation with chemicals, Fire History but care should be taken to protect desirable The task of restoration is made much easier native understory species. in this situation. At least two growing sea- A second technique involves retaining 5– son fires with a 2-year interval should be per- 2 8m/ha BA of the existing pine stand informed prior to harvesting the overstory pines. tact as a fuel source for future fires and to The next steps can be similar or identical to maintain the appearance of a forest while the the scenarios described above, beginning with new stand is developing. Harvesting should be the harvest of the overstory. As above, the har- done so as to create scattered gaps through- vest can take either the form of a clearcut or a out the stand. Underplanting with container- thinning to create gaps in which to plant lon- ized seedlings can be accomplished in the gleaf seedlings. newly created gaps. Growth and survival of the new stand will likely be affected negatively (McGuire et al. 2001; Palik et al. 2003). Com- Situations in Which petition by regeneration from seed from the residual trees poses a greater problem for the Natural Regeneration longleaf pine seedlings. Fire will be necessary Is an Option to control that regeneration while the longleaf pine becomes established. Eventually, the rem- Pine Stands with BA > 4.5 nant pine overstory should be removed in a m2/ha Longleaf Pine series of steps, creating or enlarging existing gaps for underplanting. Gap size is important. in the Overstory If gaps sizes are greater than 0.1 ha, it is difficult Some natural pine stands in longleaf pine’s to get fire to carry into the centers to control range contain a significant longleaf pine com- invading pines or hardwoods. Smaller gaps al- ponent in the overstory. Some of these stands low few resources for the new seedlings and offer the option of natural regeneration if there make survival and growth problematic (Palik is an adequate longleaf pine seed source and et al. 2003). that seed source is well distributed. Longleaf In all cases, fire must be used regularly to pine seeds are the heaviest of the southern pine maintain the stand and to encourage desirable seeds and do not usually disperse farther from native understory plants. The new seedlings the parent tree than the tree’s height, typically can be safely burned within a year of planting 30 m or less (Boyer and Peterson 1983). Conse- (Wahlenberg 1946). When the seedlings are quently, it is recommended that no fewer than just emerging from the grass stage (0.2–1.5 m 50 seed trees/ha greater than 40 cm in diame- tall), they are vulnerable to fire and should ter at breast height (DBH) be present for ade- be burned very carefully or not at all to avoid quate coverage of the area with seed (Barnett damage. Once the trees are 1.5 m tall or taller, and Baker 1991). The only way to determine they can be burned fairly safely unless they if this is true is a good ground inventory or are candling. Once they are 2 m tall, they can “cruise.” Plot tallies should be kept separately be burned safely in any season. One recom- because an average “trees per acre” figure is of mended technique employs narrow strip head- limited use if the trees are not well distributed. fires for the early burns (Mobley et al. 1978). If there is an evenly distributed and ade- Later fires can take many forms, but growing quate longleaf pine seed source available, then season fires should be part of the mix. If the it might be possible to regenerate the stand understory does not recover, it may be because with longleaf pine naturally via some form of the site was farmed prior to the previous stand, the shelterwood method. In the classic shelter- not uncommon in the Southeast. wood method used in longleaf pine systems, a 9. Restoring the Overstory 289 preparatory cut is made to obtain a BA of ap- moved when an adequately stocked new stand proximately 14 m2/ha to allow the crowns of is achieved. The seed trees may be retained, at potential seed trees to expand, therefore pro- the owner’s discretion, to create a two-aged ducing more cones (Boyer and White 1990). stand. There is evidence that the retention of When the crowns have expanded into the as little as 2.3 m2 BA/ha can significantly retard canopy gaps, a second cut is made to reduce growth of seedlings, however (Boyer 1993). If the BA to approximately 7 m2/ha, favoring the stand is a mixed species pine stand with good cone producers well distributed across longleaf as a minor component, it may be nec- the site. Fire is used throughout this process essary to use fire or other means to control to control competing vegetation, particularly seedlings of the other species while the shel- woody shrubs, and to encourage a desirable terwood process is underway. It will proba- herbaceous understory. bly be necessary to leave some trees of other Once the desired overstory stocking level is pine species in the first shelterwood cut to obtained, the cone crop must be monitored provide fuel and shade to control invading in anticipation of a good seed year. Longleaf woody brush. In the second cut, however, it pine produces good seed years sporadically, is unwise to leave any loblolly or other pines on the average of every 6 years across most to provide seed to compete with the longleaf of its range, although it may vary from one pine seedlings. Because these species are an- every 3 years to one every 30 years (Cro- nual and prolific producers of widely dispersed ker and Boyer 1975; Boyer and White 1990; seed, they are excellent colonizers of unoccu- Boyer 1999). Prescribed fire is necessary while pied sites. It is preferable to leave gaps in the waiting on a good seed crop to maintain con- stand than to retain loblolly seed trees. Once trol of the understory. Because longleaf has loblolly seedlings attain2mormore in height, a “2-year” cone, inventory of the potential typically in 3 or 4 years, they are tolerant of seed crop can begin in the year previous to fire and are difficult to control that way. seed fall. Binocular counts of female flowers There is always the option to clearcut and the year prior are an index of the coming plant with longleaf pine seedlings after estab- seed crop. In the spring and summer of the lishing a burning regime and controlling the next year, binocular surveys of tree crowns woody understory. This is a more reliable and can be conducted to estimate the number of quicker way to reestablish the longleaf pine maturing cones (Croker 1971). When the seed component in these stands, but sacrifices the trees have an average of 30 cones/tree or 2500 appearance of the forest for the short term and cones/ha, a burn should be performed to pre- creates an even-aged stand. Mature forest ob- pare a seedbed. This burn should take place ligate species such as red-cockaded woodpeck- in late winter or early spring (March–April) to ers will be eliminated by this practice. In some control vegetation and reduce the litter layer, cases the management goal is to maintain the allowing the seed to fall on mineral soil. Suf- native genetic population. In such situations, ficient time should be allowed for litter and choices include natural regeneration or collect- vegetation regrowth to cover the soil and con- ing seed from the site for use in artificial regen- ceal the seed from seed-eating birds and small eration. The new stand should still be managed mammals until it can germinate. Longleaf pine with fire as described below, with caution ad- seed typically falls in October and November vised when the seedlings are most vulnerable. and germinates as soon as it encounters adequate moisture (Boyer and White 1990). If seeds fall on heavy litter, they may germinate Stands with No Recent History but suffer from May droughts or be killed in the first fire because their roots are not in mineral of Fire soil (Boyer and White 1990). Fire must be a part of the restoration pro- Classic shelterwoods are even-aged regen- cess, but must be reintroduced very carefully eration systems, and the seed trees are re- if it has been long excluded. Excessive fuel 290 IV. Restoration buildups can result in damaging or lethal fires, is not the case, the process might be “jump- even in mature stands (Haywood et al. 2004; started” by planting or seeding desired species. Outcalt and Wade 2004). In stands with a pine overstory other than longleaf, forest floor fuel depths may not be as great, because loblolly Stands with a Recent History and shortleaf pine needles are shorter and de- of Fire cay more rapidly than do longleaf pine nee- Pine stands with a recent fire history may be dles. Still, if fuel accumulations are very deep, restored in the same manner, but it is much feeder roots may have grown up into the lit- more likely that the woody understory and ter and can be damaged or killed by a hot fire. accumulated forest floor fuel will not require In addition, litter buildups composed of bark, special treatment. Burning should begin with a cones, and needles surrounding the base of the growing season fire in late March or April and boles of standing timber can smolder for hours follow with both growing season and dormant and even days, damaging the cambium where season burns until the stand is regenerated. At the bark is thin. In addition, long-term fire ex- that point, fire should be continued as often clusion usually results in a woody mid- and as needed to control fuels and to encourage understory. These fuels can feed hot fires, es- restoration of the desired understory. pecially if needles drape on vines and shrubs and “ladder” flames up into the canopy. The first fire in a long unburned pine stand Stands that Are Predominately should be as cool as possible, burning on a cold, breezy day with moderate humidity (30–60%) Longleaf Pine and relatively high fuel moisture. In extreme There are several options for regeneration and cases, fuel treatments such as mowing or rak- maintenance of stands dominated by longleaf ing around the boles of overstory trees might pine. Regeneration of these stands can be be employed. If longleaf pine is to be favored in by artificial or natural means and can take these stands, special attention should be paid place all at once, in several stages, or contin- to the bases of existing longleaf pine. Raking or uously. Even-aged stands may be maintained wetting the fuel around the trees might lessen in an even-aged condition by regenerating by the danger of damage from fire if scale of op- clearcutting and replanting or by use of the eration allows. Strip headfires move the heat shelterwood method. The existing stand can through the stand relatively quickly without usually be retained for a relatively long time allowing intensity to build to dangerous lev- (200 years or more) given the long life span els (Wade and Lunsford 1988). The second fire of longleaf pine (Matoon 1922; Wahlenberg might also be a dormant season strip headfire 1946). to further reduce fuel loads, encourage root If the objective is an uneven- or all-aged growth into deeper soils, and consume or top- stand or forest, there are several techniques ple woody shrubs top-killed by the first fire. available. Creating gaps in the canopy of an The interval between the first and second fire even-aged mature longleaf pine forest allows should be no more than 3 years. Subsequent longleaf pine regeneration to occur when seed fires should take place at 2- or 3-year inter- crops are available. If the gaps are created vals, depending on fuel conditions. The third prior to a seed crop, fire must necessarily be or fourth fire should be a growing season fire, used to keep those gaps from becoming oc- ideally in late March through early May. From cupied with woody brush or hardwoods like this point on, at least every third burn should turkey oak (Quercus laevis) or water oak (Quer- be a growing season burn, extending the win- cus nigra). Another approach includes creat- dow into early summer. As fuel loads are re- ing gaps after seed fall opening holes in the duced, hotter fires can be safely used. canopy after new seedlings have established The understory should begin to respond fol- themselves and giving them space to grow lowing the first growing season fire and con- into. Optimum gap size is a subject of some tinue to develop with successive burns. If this discussion among managers and researchers, 9. Restoring the Overstory 291 but there are some generally accepted trade- duced into these situations very cautiously. offs. Smaller gaps decrease seedling survival Fuel treatments such as raking around existing and growth rate (McGuire et al. 2001). The trees, wetting areas immediately around exist- removal of single trees throughout a longleaf ing trees, and mowing or otherwise remov- pine stand can work with longleaf pine man- ing standing fuels are suggested techniques agement systems if the stand is fairly open in to avoid or reduce mortality of the overstory. nature (less than 7 m2/ha) and growth rates Longleaf pine is fire-tolerant, not fire-proof. are not particularly important. Small group se- Feeder roots frequently grow into the litter lection harvests of 0.1 to 0.2 ha are adequate layer in long unburned stands and can be for longleaf pine regeneration (McGuire et al. severely damaged by even moderately hot sur- 2001; Palik et al. 2003). Typically, subsequent face fires. Because it is common for fuel to cuts are used to expand these openings in an- accumulate around the base of large longleaf ticipation of new seedlings and perpetuating pine in fire-excluded situations, fires can smol- the range of age classes of an all-aged for- der in these piles of bark, foliage, and dead est. “Domes” of young longleaf pine saplings wood for long periods, damaging the cambium and seedlings are created, with the oldest and layer of the bole at a location where the in- tallest saplings in the center of the gaps and the sulating bark is often thin to begin with and height and age decreasing with distance from killing or damaging feeder roots, making them the center. Eventually, as the openings are en- more vulnerable to invading diseases (Kush larged, they merge into each other and a forest et al. 1999b; Hanula et al. 2002). High mortal- with groups of trees representing many seed ity levels in mature longleaf pine stands are not years results. This stand structure is consid- unusual when fire is reintroduced in long un- ered typical of much of the presettlement lon- burned stands. If scale permits, extraordinary gleaf pine forest (Matoon 1922; Wahlenberg efforts are warranted to prevent this mortal- 1946). Patch size undoubtedly varied in preset- ity. If it does not, the first fires should be con- tlement forests, because natural forces of vary- ducted on cool or cold days, with moderate ing intensity created them. Lightning strikes, humidity (40% or greater) and moderate to fires, tornados, and insects likely caused small high fuel moisture levels. Wind speeds should gaps in the forest, while hurricanes and catas- be moderate (e.g., 8–12 km/h) and steady. trophic fires had the potential to flatten large Strip headfires are recommended, with strips areas. It is postulated that constant frequent 30 m or less apart (Wade and Lunsford 1988). fire relegated slash and loblolly pines and hard- This allows the fires to move fairly rapidly woods to wetter, less fire-prone areas, leaving through the fuel without building excessive the field, so to speak, to longleaf pine (Schwarz intensity and without igniting the smolder- 1907; Chapman 1932). Recurring fires prob- ing fires so dangerous to longleaf pine sur- ably kept invading species at bay, while the vival. As the fuel is gradually reduced with suc- longleaf pine gradually reoccupied large areas cessive fires, more latitude with fire intensity from seed and seedlings in place and residual and season is allowed. Eventually, when most seed trees that survived the catastrophic event feeder roots are in mineral soil and the fuel (Heyward 1939). load is reduced, growing season fires may be introduced into the fire regime to encourage the herbaceous understory typical of longleaf Longleaf Stands with No Recent pine forests. When regenerating longleaf pine stands naturally, it is important to remember Fire History that longleaf pine seed should fall on mineral Longleaf pine forests are more likely than other soil or very light litter to achieve best germi- pine forests to accumulate high levels of for- nation and survival and fires should be timed est floor litter. Longleaf needles are larger and accordingly. It is equally important to remem- more decay resistant than those of other pines ber that longleaf pine seedlings are extremely and can build up litter depths of 25 cm or vulnerable to fire in the first year after germi- more (Hermann 2001). Fire must be reintro- nation and mortality is high on seedlings with 292 IV. Restoration root collar diameters less than 0.6 cm (Boyer Barnett, J.P., Hainds, M.J., and Hernandez, G.A. 1979). Once seedlings start height growth, fires 2002. Interim guidelines for growing longleaf should be applied cautiously or withheld when seedlings in containers. General Technical Report a majority of the seedlings are between 15 cm SRS-60. Southern Research Station, USDA Forest and 1.5 m in height, as this size class is vul- Service. nerable to fire, particularly when the terminal Bengston, G., DuPre, J., Twomey, W.,and Hooper, R. 1993. Longleaf ecosystem restoration in the wake bud is in the “candle” stage (Grelen 1982). of Hurricane Hugo. In Proceedings of the Tall Tim- bers Fire Ecology Conference, No. 18, ed. S. M. Her- mann, pp. 339–347. Tall Timbers Research Sta- Longleaf Stands with a Recent tion, Tallahassee, FL. History of Fire Boyer, W.D.1979. Regenerating the natural longleaf pine forest. J For 77:572–575. The key to managing longleaf pine stands Boyer, W.D. 1987. Volume growth loss: A hidden with a recent fire history is to maintain the cost of periodic prescribed burning in longleaf fire regime and decide on the method of re- pine? South J Appl For 11(3):154–157. generation. If the understory is still occupied Boyer, W.D. 1993. Long-term development of re- by woody shrubs, a switch to growing sea- generation under longleaf pine seedtree and shel- son burns will encourage the proliferation terwood stands. J For 17:10-15. of grasses and forbs typically indicative of Boyer, W.D. 1999. Longleaf pine: Natural regenera- restored longleaf pine forests (Lewis and tion and management. Alabama’s Treasured Forests Harshbarger 1976; Brockway and Lewis 1997; 18:7–9. Glitzenstein et al. 1998; Haywood et al. Boyer, W.D., and Peterson, D.W. 1983. Longleaf 2001; Kush 2001). The choice of regeneration pine. In Silvicultural Systems for the Major Forest Types of the United States. Agricultural Handbook method will shape the stand’s structure for the 445. pp. 153–156. Washington, DC: US Depart- future. Both even- and uneven-aged condi- ment of Agriculture. tions are natural in longleaf forests (Croker and Boyer, W.D., and White, J.B. 1990. Natural regener- Boyer 1975; Boyer 1979, 1999; Dennington ation of longleaf pine. In Proceedings of the Sympo- and Farrar 1983; Boyer and White 1990; Bar- sium on the Management of Longleaf Pine, April 4–6, nett and Baker 1991; Farrar 1996). By main- 1989, Long Beach, MS, pp. 94–113. General Techni- taining fire in this system, the common char- cal Report SO-75. USDA Forest Service, Southern acteristic is a fire-maintained plant and animal Forest Experiment Station. community that can for the foreseeable future Boyette, W.G. 1996. 1995 survey of longleaf pine be maintained. restoration efforts in the South. North Carolina Division of Forest Resources, Raleigh. Brissette, J.C., Barnett, J.P., and Landis, T.D. 1991. References Container seedlings. In Forest Regeneration Man- ual, eds. M.L. Duryea and P.M. Dougherty, Barnard, E.L., Dixon, M.N., and Ash, E.C. 1995. pp. 117–141. Norwell, MA: Kluwer Academic Benomyl root dip and scalping improves perfor- Publishers. mance of longleaf pine on pest-infested agricul- Brockway, D.G., and Lewis, C.E. 1997. Long-term tural croplands. Tree Planters Notes 46:93–96. effects of dormant-season prescribed fire on plant Barnett, J.P., and Baker, J.B. 1991. Regeneration community diversity, structure and productivity methods. In Forest Regeneration Manual, eds. M.L. in a longleaf pine wiregrass ecosystem. For Ecol Duryea and P. M. Dougherty, pp. 35–50. Norwell, Manage 96:167–183. MA: Kluwer Academic Publishers. Brockway, D.G., and Outcalt, K. W. 2000. Restoring Barnett, J.P., Lauer, D.K., and Brissette, J.C. 1990. longleaf pine wiregrass ecosystems: Hexazinone Regenerating longleaf pine with artificial meth- application enhances effects of prescribed fire. For ods. In Proceedings of the Symposium on the Manage- Ecol Manage 137:121–138. ment of Longleaf Pine, April 4–6, 1989, Long Beach, Cantrell, R.L. 1985. A guide to silvicultural herbi- MS, pp. 72–93. General Technical Report SO-75. cide use in the southern United States. School of USDA Forest Service, Southern Forest Experi- Forestry and Wildlife Sciences, Auburn Univer- ment Station. sity, AL. 9. Restoring the Overstory 293

Chapman, H.H. 1932. Is the longleaf type climax? species diversity in longleaf pine groundcover: Ecology 8(4):328–334. Effects of fire regimes and seed/seedling intro- Croker, T.C., Jr., 1971. Binocular counts of longleaf ductions. In Proceedings of the Longleaf Pine Ecosys- pine strobili. SO-127. New Orleans, LA. USDA tem Restoration Symposium, Society for Ecologi- Forest Service, Southern Forest Experiment Sta- cal Restoration 9th Annual International Confer- tion. ence, ed. J.S. Kush, pp. 72–75. Longleaf Alliance Croker, T.C., Jr., and Boyer, W.D. 1975. Regenerat- Report No. 3. The Longleaf Alliance, Andalusia, ing longleaf pine naturally. SO-105. New Orleans, AL. LA. USDA Forest Service, Southern Forest Exper- Grelen, H.E. 1982. May burns benefit survival and iment Station. growth of longleaf pine seedlings. In Proceedings Dennington, R. W. 1990. Regenerating longleaf pine of the Second Biennial Southern Silvicultural Research with the shelterwood method. R8-MB47. At- Conference, ed. E. P. Jones, pp. 70–73. General lanta, GA. USDA Forest Service, Southern Re- Technical Report SE-24. Asheville, NC: USDA, gion. Forest Service, Southeastern Forest Experiment Dennington, R. W., and Farrar, R. M., Jr. 1983. Station. Longleaf pine management. R8-FR3. Atlanta, GA. Hainds, M.J. 1999. Successfully planting longleaf USDA Forest Service, Southern Region. pine. Alabama’s Treasured Forests 18:10–11. Earley, L.S., ed. 1997. A Working Forest: A Hainds, M.J. 2001. Scalping aids survival of longleaf. Landowner’s Guide for Growing Longleaf Pine in Alabama’s Treasured Forests 20(3):24–27. the Carolina Sandhills. Sandhills Area Land Trust, Hainds, M.J. 2003a. Determining the correct plant- Southern Pines, NC. ing depth for container-grown longleaf pine Earley, L.S., ed. 2002. Managing the Forest and the seedlings. In Proceedings 4th Biennial Regional Lon- Trees: A Private Landowner’s Guide to Conservation gleaf Alliance Conference, pp. 66–68. November 17– Management of Longleaf Pine. Zebulon, NC: Theo 20, 2002, Southern Pines, NC. Davis Sons Inc. Hainds, M.J. 2003b. Establishing longleaf seedlings Farrar, R.M., Jr. 1996. Fundamentals of uneven- on agricultural fields and pastures. In Proceedings management in southern pines. Miscellaneous 4h Biennial Regional Longleaf Alliance Conference, pp. Publication No. 9. Tall Timbers Research Station, 69–73. November 17–20, 2002, Southern Pines, Tallahassee, FL. NC. Farrar, R.M., Jr., and Boyer, W.D. 1991. Manag- Hainds, M.J., and Barnett, J.P. 2003. Container- ing longleaf pine under the selection system— grown longleaf pine quality. In Proceedings 4th Bi- Promises and problems. In Proceedings of the ennial Regional Longleaf Alliance Conference, pp. 63– Sixth Biennial Southern Silvicultural Research Con- 65. November 17–20, 2002, Southern Pines, NC. ference, October 30–November 1, 1990, Memphis, Hanula, J.L., Meeker, J.R., Miller, D.R., and TN, pp. 357–368. General Technical Report 70, Barnard, E.L. 2002. Association of wildfire with Asheville, NC: USDA Forest Service Southeastern tree health and numbers of pine bark beetles, re- Forest Experiment Station. production weevils and the associates in Florida. Franklin, R.M. 1997. Stewardship of longleaf pine For Ecol Manage 170:233–247. forests: A guide for landowners. Longleaf Alliance Haywood, J.D., Harris, F. L., and Grelen, H. E. 2001. Report No. 2. The Longleaf Alliance, Solon Dixon Vegetation response to 37 years of seasonal burn- Forestry Education Center, Andalusia, AL. ing on a Louisiana longleaf pine site. South J Appl Frost, C.C. 1990. Natural diversity and status of lon- For 25(3):122–130. gleaf pine communities. In Forestry in the 1990s— Haywood, J.D., Bauman, T.A., Goyer, R.A., and Har- A Changing Environment, eds. G. Youngblood and ris, F.L. 2004. Restoring upland forests to longleaf D.L. Frederick, pp. 26–35. Regional Technical pine: Initial effects on fuel load, fire danger, for- Conference. Pinehurst, NC: Society of American est vegetation, and beetle populations. In Proceed- Foresters. ings of the 12th Biennial Southern Silvicultral Research Frost, C.C. 1993. Four centuries of changing land- Conference, ed. K. F. Connor, pp. 299–303. General scape patterns in the longleaf pine ecosystem. In Technical Report SRS-71. Asheville, NC: USDA Proceedings of the Tall Timbers Fire Ecology Conference, Forest Service, Southern Research Station. No. 18, ed. S. M. Hermann, pp. 17–43. Tall Timbers Hermann, S.M. 2001. A brief overview of fire and Research Station, Tallahassee, FL. season of burn in native longleaf pine ecosystems. Glitzenstein, J.S., Streng, D.R., Wade, D.D., and In Forest for Our Future. Restoration and Management Platt, W.J. 1998. Maintaining and restoring of Longleaf Pine Ecosystems: Silvicultural, Ecological, 294 IV. Restoration

Social, Political, and Economic Challenges. Proceed- Lewis, C.E., and Harshbarger, T.J. 1976. Shrubs ings of the 3rd Longleaf Alliance Regional Con- and herbaceous vegetation after 20 years of preference. October 16–18, 2000, comp. J.S. Kush, scribed burning in the South Carolina coastal pp. 53–57. Alexandria, LA. Longleaf Alliance Re- plain. J Range Manage 29(1):13. port No. 5. Long, A.J. 1991. Proper planting improves per- Heyward, F. 1939. The relation of fire to stand formance. In Forest Regeneration Manual, eds. composition of longleaf pine forests. Ecology M. L. Duryea and P. M. Dougherty, pp. 303–321. 20(2):287–304. Norwell, MA: Kluwer Academic Publishers. Johnson, R. 1999. Restoring the longleaf pine forest Lowery, R. F., and Gjerstad, D.H. 1991. Chemi- ecosystem. Alabama’s Treasured Forests 18(4):18– cal and mechanical site preparation. In Forest Re- 19. generation Manual, eds. M.L. Duryea and P. M. King, S. E., Grace, S. L., and Hedman, C.W. 1997. Dougherty Norwell, MA: Kluwer Academic Pub- The importance of site history on ground cover lishers. species of longleaf pine systems: Implications for Mann, W.F., Jr. 1970. Direct-seeding longleaf pine. restoration. In Longleaf Pine: A Regional Perspec- USDA Forest Service Research Paper SO-57. tive of Challenges and Opportunities, Proceedings of Southern Experiment Station. the First Longleaf Alliance Conference, Septem- Matoon, W.R. 1922. Longleaf pine. US Department ber 17–20, 1996, comp. J.S. Kush, pp. 102–103. of Agriculture Bulletin No. 1061. Washington, Mobile, AL. Longleaf Alliance Report No. 1. DC. Kush, J.S., comp. 1997. Longleaf Pine: A Regional McGuire, J.P., Mitchell, R. J., Moser, E.B., Pecot, S. Perspective of Challenges and Opportunities, Proceed- D., Gjerstad, D. H., and Hedman, C.W. 2001. Gaps ings of the First Longleaf Alliance Conference, in a gappy forest: Plant resources, longleaf pine September 17–20, 1996, Mobile, AL. Longleaf Al- regeneration, and understory response to tree reliance Report No. 1. moval in longleaf pine savannas. Can J For Res Kush, J.S., comp. 1999. Longleaf Pine: A Forward 31:765–778. Look, Proceedings of the 2nd Longleaf Alliance Mexal, J.G., and South, D.B. 1991. Bareroot Conference, November 17–19, 1998, Charleston, seedling culture. In Forest Regeneration Manual, SC. Longleaf Alliance Report No. 4. eds. M.L. Duryea and P. M. Dougherty, pp. 89– Kush, J.S., comp. 2001. Forest for Our Future. Restora- 116. Norwell, MA: Kluwer Academic Publishers. tion and Management of Longleaf Pine Ecosystems: Minogue, P. J., Cantrell, R. L., and Griswold, H. C. Silvicultural, Ecological, Social, Political, and Eco- 1991. Vegetation management after plantation nomic Challenges, Proceedings of the 3rd Longleaf establishment. In Forest Regeneration Manual, eds. Alliance Regional Conference, October 16–18, M. L. Duryea and P. M. Dougherty. Norwell, MA: 2000, Alexandria, LA. Longleaf Alliance Report Kluwer Academic Publishers. No. 5. Mitchell, R.J., Neel, W.L., Hiers, J.K., Cole, F. T., and Kush, J.S., comp. 2003. Longleaf Pine: A Southern Atkinson, J.B., Jr. 2000. A model management Legacy Rising from the Ashes, Proceedings of the 4th plan for conservation easements in longleaf pine- Longleaf Alliance Regional Conference, Novem- dominated landscapes. Joseph W. Jones Ecologi- ber 17–20, 2002, Southern Pines, NC. Longleaf cal Research Center, Newton, GA. Alliance Report No. 6. Mobley, H.E., Jackson, R.S., Balmer, W.E., Ruziska, Kush, J.S., Meldahl, R. S., and Boyer, W.D. W.E., and Hough, W.A. 1978. A guide for pre- 1999a. Understory plant community response scribed fire in southern forests. USDA Forest Ser- after 23 years of hardwood control treatments in vice, State and Private Forestry, Atlanta. natural longleaf pine (Pinus palustris) forests. Can Outcalt, K.W., and Wade, D.D. 2004. Fuels man- J For Res 29:1047–1054. agement reduces tree mortality from wildfires Kush, J.S., Varner, J.M., and Meldahl, R. S. 1999b. in southeastern United States. South J Appl For Slow down, don’t burn too fast . . . Got to make 28(1):28–34. that old-growth last. In Longleaf Pine: A Forward Palik, B., Mitchell, R.J., Houseal, G., and Pederson, Look, Proceedings of the 2nd Longleaf Alliance N. 1997. Effects of canopy structure on resource Conference, comp. J. S. Kush, pp. 109–111. Lon- availability and seedling responses in a longleaf gleaf Alliance Report No. 4. pine ecosystem. Can J For Res 27:1458–1464. Landers, J.L., Van Lear, D.H., and Boyer,W.D. 1995. Palik, B., Mitchell, R.J., Pecot, S., Battaglia, M., The longleaf forests of the Southeast: Requiem or and Pu, M. 2003. Spatial distribution of over- renaissance? J For 93(11):39–44. story retention influences resources and growth 9. Restoring the Overstory 295

of longleaf pine seedlings. Ecol Appl 13:674– ment. Washington, DC: Charles Lathrop Pack 686. Foundation. Robbins, L.E., and Myers, R.M. 1992. Seasonal ef- Wakeley, P.C. 1953. Planting the southern pine. fects of prescribed burning in Florida: A review. Government Printing Office, Washington, DC. Tall Timbers Research, Inc. Miscellaneous Publi- USDA Agricultural Monograph No. 18. cation No. 8. Tallahassee, FL. Walker, J. 1998. Ground layer vegetation in Schwarz, G.F. 1907. The Longleaf Pine in Virgin longleaf pine landscapes: An overview for Forest—A Silvical Study. New York: John Wiley & restoration and management. In Proceedings Sons. of the Longleaf Pine Ecosystem Restoration Sym- Swezy, D.M., and Agee, J.K. 1991. Prescribed-fire posium, Society for Ecological Restoration effects on fine-root and tree mortality in old- 9th Annual International Conference, Lon- growth ponderosa pine. Can J For Res 21(5):626– gleaf Alliance Report No. 3, comp. J.S. Kush, 634. pp. 2–13. The Longleaf Alliance, Andalusia, Wade, D.D., and Lunsford, J.D. 1988. A guide for AL. prescribed fire in southern forests. Technical Pub- Williston, H.L., and Balmer, W.E. 1971. Direct lication TP11. Atlanta, GA. USDA Forest Service, seeding of southern pines—A regeneration al- Southern Region. ternative. Forest Management Bulletin. US For- Wahlenberg, W.G. 1946. Longleaf Pine: Its Use, Ecol- est Service, State and Private Forestry, Atlanta, ogy, Regeneration, Protection, Growth, and Manage- GA. Chapter 10 Restoring the Ground Layer of Longleaf Pine Ecosystems

Joan L. Walker and Andrea M. Silletti

The longleaf pine ecosystem includes some of however, we believe there are general patterns the most species-rich plant communities out- within this ecological system that can guide side of the tropics, and most of that diversity restoration protocol development. In addition, resides in the ground layer vegetation. In ad- restorationists have practical experience that dition to harboring many locally endemic and is not documented in the peer-reviewed liter- otherwise rare plant species (Peet this volume) ature at this time but nevertheless converges and enhancing habitat for the resident fauna on some necessary steps for successful restora- (Costa and DeLotelle this volume), the ground tion. In this chapter we summarize the gen- layer vegetation produces fine fuel needed to eral lessons learned from ongoing restoration carry low-intensity surface fires that perpetu- efforts, and discuss ecological aspects of the ate the ecosystem. Ecosystem restoration re- ground layer vegetation that guide us in ex- quires the restoration of both the ground layer trapolating this information to other sites. Our plant community and the pine canopy. purpose is to share information so that we Ground layer restoration in longleaf pine might advance the restoration of the ground communities is an area of active investigation, cover in longleaf pine communities by mini- through both adaptive management projects mizing avoidable mistakes and by identifying and formal research. However, there is no critical information needs. restoration manual for the longleaf pine com- In the first section we provide an overview munity. Instead, restoration practitioners de- of the ground cover vegetation in the long- velop their action plans based on an ecological leaf pine ecosystem. We then describe extant reference model and project goals, and achieve conditions in longleaf pine sites often targeted their objectives using conventional natural refor restoration, including the ways they dif- sources management and horticultural meth- fer from reference conditions and their deriva- ods. Given the natural heterogeneity of the tion from widespread historical land uses. The longleaf pine ecosystem at multiple scales and next two sections summarize lessons learned the differences imposed by a varied land use from research and restoration projects that em- history, one could argue that there never will phasize (1) altering canopy structure to favor be a manual to adequately describe or pre- ground-layer restoration or (2) starting new scribe restoration protocols for all situations; populations of ground cover species. We close

r Joan L. Walker and Andrea M. Silletti USDA Forest Service, Southern Research Station, Clemson, South Carolina 29634.

297 298 IV. Restoration with an assessment of information needed to the beak rushes (Rhynchospora spp.), and lilies. advance ground cover restoration in the lon- More unusual plants include orchids and car- gleaf pine ecosystem. nivorous species, often associated with wet, Our review of existing restoration projects nutrient-poor sites. shows that they are being conducted on a Locally, ground layer composition and struc- relatively narrow subset of possible longleaf ture vary with fire frequency and soil con- pine habitats. Significant projects we know ditions, typically characterized by soil texture of are concentrated in the Atlantic and Gulf and interpreted as variation in soil moisture Coastal Plains; mesic savannas and flatwoods, status (Peet and Allard 1993). Overall, fre- loamy upland sites, and xeric to subxeric sites quently burned sites have more species at small in the Fall-line Sandhills are represented. We spatial scales than sites where fire has been note the absence of projects in the middle eliminated; and intermediate-to-wet sites sup- Atlantic Coastal Plain, where few examples port more species than very dry sites (Peet this of remnant vegetation remain, in the moun- volume). tain longleaf pine communities of the Blue In spite of these general patterns, there is Ridge and Cumberland Plateau, and in the considerable compositional variation from one longleaf pine–bluestem communities. Infor- part of the region to another. Most herbaceous mation presented in this chapter draws on species have geographic ranges that are much projects conducted by researchers and restora- smaller than that of longleaf pine. Species that tion practitioners in Florida, South Carolina, have more or less restricted geographic ranges and Georgia. Although many of the same are known as endemic species, and the longleaf issues exist in the underrepresented areas, pine ecosystem has many subregional and lo- unique species, habitats, spatial contexts, and cal endemic species (Estill and Cruzan 2001; historical land uses are bound to generate LeBlond 2001; Sorrie and Weakley 2001). As some restoration challenges that we have not the geographic limit of a species’ range is addressed. reached, it drops out of the local flora but may be replaced by an ecologically similar species. This results in changing species composition in Ground Layer Vegetation the ground layer. Species with very small geographic distributions (narrow endemics) are in Longleaf Pine prone to extinction and include some of the Landscapes: An Overview ground layer species that are federally listed as Endangered or Threatened (Walker 1998). Be- Throughout the range of longleaf pine the cause there are important differences among general picture of a frequently burned high- sites, describing the ecologically appropriate quality natural area shows a predominantly composition for restoration must be done care- herbaceous ground layer dominated by grasses fully. with a diverse mixture of forbs. Woody species, if present, are short and inconspicuous. Most of the common species are sun-loving perennials with an ability to resprout after fire. Fire typi- Reference Models and cally stimulates the flowering and seed produc- Goals for Ground Layer tion by many characteristic species, and there are apt to be species flowering at most any time Restoration practitioners use “reference mod- during the growing season. els” to describe the ecological potential for a Grasses, legumes, and composites are the project site. A reference model is a description most common plant families in these burned of the restoration site as it may have looked and habitats (Harcombe et al. 1993; Peet and Al- functioned in the past, before negative changes lard 1993; Drew et al. 1998). Other com- had occurred. Ideally the description answers mon families include the sedges, especially questions about composition, structure, and 10. Restoring the Ground Layer 299 function. Intact remnant patches of the target same site. Desirable historical information in- ecosystem, such as nearby “natural areas,” are cludes historical photographs, written descrip- sometimes identified as “reference sites.” They tions, plant and animal species lists, frequency are selected to match the restoration project of burn in the area and under what conditions, site with respect to geography and physical en- and/or reports of significant disturbances or vironment and are believed to represent the past land uses. historic or contemporary potential conditions. Practitioners sometimes use historical or Nearby environmentally similar sites are the contemporary information from other sites, or best reference sites, but the condition of any from less specific geographic areas. Though proposed reference site may have been influ- such information may be useful, it is im- enced by unknown stochastic events in the portant to remember that information about past (such as extreme weather or disturbance places is generally place- and time-specific. The events) so that its condition may not represent more distant or more general the information the project site potential. source, the less likely it will accurately repre- Besides using reference sites, restoration sent a specific project site and the less useful ecologists use other kinds of “reference infor- it will be for setting feasible objectives (Table mation” to develop reference models (Table 1; 1). Egan and Howell (2001) recommend that White and Walker 1997). Ideally the project restorationists use a combination of site anal- planner would conduct a site assessment to ysis (same time, same place), historic infor- gather current information about the site to mation from the project site (different time, be restored, including a description of the un- same place), and information from contem- derlying environmental conditions, and search porary reference sites (same time, different for accurate historical information about the place).

TABLE 1. Examples of reference information that can be used to develop a reference model.a Time/space Restoration project site Different site or general location Contemporary (Site analysis) (Reference site if it matches the (Observed directly; change can r conditions and geography of Physical environment. Examples: be monitored) site to be restored) soil type, fertility, hydrology, r r topographic position, etc. Same as for site analysis data Biotic environment. Examples: The nearer the site to the (1)canopy—composition, age, size project site, the more likely class distribution, origin; (2) other that information can be used vegetation—composition, species r directly in reference model abundance, presence of exotic General location information r species Examples: county species Disturbance evidence. Examples: lists, herbarium records ﬁre scars, plow lines

HIGH VALUE for reference model HIGH-MODERATE VALUE Historical (Site history) (Historical information from (Snapshot of past; cannot r different sites) observe change or know effect r Photographs, with dates r of stochastic events in the past) Written descriptions of physical r Similar to site history data and biotic conditions, past land Fire scar data in general uses or disturbances (sources: r landscape or region deeds, explorers’ accounts, diaries r Regional land use history and letters of previous owners) Pollen data (prehistorical) HIGH VALUE LOW VALUE a Reference information can be classiﬁed into four categories based on geographic source of data (the site to be restored versus a different or general location) and whether the data represent current (contemporary) or historical conditions. Modiﬁed from White and Walker 1997 with permission from Blackwell. 300 IV. Restoration

A variety of restoration goals may be com- from reference conditions in different ways patible with the site’s ecological potential, for and thus represent somewhat distinct chal- example, establishing fine fuels to facilitate fire lenges to restoration (Fig. 1). These largely an- use in timber management, improving habitat thropogenic disturbances have generated very for a bobwhite quail, or creating an aestheti- different starting points in terms of physical cally desirable setting. Goals like these exam- conditions and especially of biotic legacies, ples may be viewed as restoring a subset of which are the remnant components of the the composition, structure, and function that undisturbed longleaf ecosystem. Effort needed historically characterized a site, in contrast to to restore a site will vary inversely with the the ambitious goal of restoring the entire com- amount of biotic legacy remaining (Fig. 2). plement of species and their interactions. In practice, such “partial restoration” goals are far more likely to be achieved than “com- Altered Fire Regimes plete restoration of biodiversity” (Lockwood The historical fire regime has been described and Pimm 1999). The feasibility of project goals as one of frequent, low-intensity surface fires. must be examined in light of the ecological ca- The extent of individual fire events and return pability of the restoration site including lim- interval are likely to have varied with topogra- itations imposed by spatial scale and context phy, thus among different parts of the longleaf (White 1996), as well as the resources that are range (Frost 1998). It is assumed that most available to do the work and to maintain it acres of longleaf pine habitat ignited by light- (White and Walker 1997; Ehrenfield 2000). ning burned during the early to mid-growing season, but Native American ignitions spanned the seasons (Robbins and Myers 1992). Over Recent Land Uses and the last 60 years land managers have reduced the spatial extent of fires, shifted the predomi- Legacies: Starting Points nant season to winter burning (which may be for Restoration associated with lower intensity fires owing to high fuel moistures and low air temperatures), Altered fire regimes, plantation establishment, and reduced fire frequency or eliminated fires and conversion of forest lands to agriculture altogether. These practices are associated with have resulted in loss of the ground cover di- increased densities and expanded distributions versity throughout the longleaf pine range of woody species (Platt et al. 1991; Robbins and (Wear and Greis 2002). Ongoing and com- Myers 1992; Waldrop et al. 1992; Streng et al. pleted restoration projects that we reviewed 1993; Glitzenstein et al. 1995; Gilliam and Platt all fall into one of these recent land use history 1999; Drewa et al. 2002) and decreased abun- classes. Of these classes pine plantations are dance of herbs (Walker and Peet 1983; Peet the most heterogeneous. They differ in canopy and Allard 1993). species (primarily loblolly or slash pine), in age, The hardwood component increases with and in methods of establishment. Also, pre- fire exclusion or reduced fire frequency, but planting site histories vary, most significantly the specific composition varies both geograph- in whether they have a history of modern cul- ically and with site conditions within a land- tivation in contrast to continuously forested or scape (Gilliam et al. 1993; Harcombe et al. lightly cultivated. Finally, they may have expe- 1993; Liu et al. 1997; Gilliam and Platt 1999; rienced a period of fire suppression. As a result Varner et al. 2003). The losses in the ground of diverse management histories, plantations cover after long periods without fire (more may resemble both agricultural sites and sites than two decades) are so profound that stud- with altered fire regimes. ies of extant old growth with significant fire Although these land use groups do not rep- exclusion focus almost exclusively on the resent mutually exclusive conditions, we think woody species component (e.g., Gilliam and it useful to consider them because they differ Christensen 1986; Gilliam et al. 1993; Gilliam 10. Restoring the Ground Layer 301

FIGURE 1. Different starting conditions for longleaf pine community restoration: (a) Xeric site where fire was excluded for about 30 years; (b) slash pine plantation on a mesic site once occupied by longleaf pine. Note increased turkey oak with fire exclusion, in contrast to absence of hardwoods in the plantation where hardwoods were controlled. Both have abundant pine straw or leaf litter, but lack a diverse herb layer. and Platt 1999; Varner et al. 2000). Descrip- an old-growth remnant in the North Carolina tions of these sites note low richness and sparse sandhills: “sparse and relatively species-poor, cover of herbaceous species, as in the Gilliam typical of pine forest herb layers under chronic et al. (1993) description of the Boyd Tract, no-fire conditions.” The most abundant species 302 IV. Restoration

FIGURE 2. Relationship between time and resources needed for restoration and the abundance of remnant biota (biological legacy) on the site to be restored. Ovals indicate the relative position of common starting conditions as deﬁned by land use history. Sites subjected to modern agricultural methods (repeated machine tilling) are distinct from sites that have remained in forest or escaped intense agriculture. Nonforest sites require planting trees, and all previously tilled sites are likely to require species additions. Infrequently burned forests may need hardwood removal or other canopy manipulations. The abundance of remnant biota on forested sites varies inversely with the intensity of any site preparation used for pine regeneration.

in the ground layer were seedlings of pine and dent at small spatial scales (less than or equal oaks and other hardwoods (such as flower- to 1.0 m2 plots) and less evident at larger ing dogwood, mockernut hickory, and black scales (greater than 600 m2). This pattern has gum). After 45 years of fire exclusion in an been shown both in mesic productive savan- old-growth site in Escambia County, Alabama, nas (Walker and Peet 1983; Glitzenstein et al. a single herb, Acalypha virginica, was present 2003) and in xeric sites (Walker 1998). Species prior to restoration fire treatments (Varner retained at larger scales may provide on-site et al. 2000). Results of fire frequency experi- seed sources for restoring the ground layer via ments, generally sampled through several fires natural dispersal and establishment. How long over a decade or less, indicate that increased species will persist is not known; if retention is woody species dominance can occur in rela- short-lived, opportunities for restoring resid- tively short periods of time (Mehlman 1992; ual populations with fire alone will diminish Beckage and Stout 2000; Glitzenstein et al. with time. 2003). Rates of species loss associated with re- Reduced fire frequency leads to scale- duced fire frequency vary with plant groups. dependent decreases in herbaceous species Among the species most likely to be lost or richness: decreases in richness are most evi- significantly reduced in mesic to wet longleaf 10. Restoring the Ground Layer 303 pine savannas are the dominant rhizomatous of flowering (Platt et al. 1988; Streng et al. and bunch grasses (Walker and Peet 1983; 1993; Brewer and Platt 1994) and that domi- Glitzenstein et al. 2003), species present as nant grasses flower only infrequently without basal rosettes (many composites), many sedges growing season burning (Robbins and Myers and other small monocots, and insectivorous 1992), there have been no convincing changes species. Mehlman (1992) found predictable in abundance and composition directly related patterns of species loss, or conversely of persis- to season of burning (Streng et al. 1993; Brock- tence, with fire exclusion in drier upland lon- way and Lewis 1997). But, fire season may gleaf pine forests. He identified more species affect the herbaceous community indirectly persisting in high-frequency than in low- through changes in canopy structure and con- frequency, burned or unburned sites. While sequent changes in the environment (reduced all three groups of stands included ruderal resource availability, particularly light and wa- and “climax” longleaf pine associates, legumes ter) for herbs (Harrington and Edwards 1999). were significantly associated only with burned In summary, fire frequency is likely to be groups, and only woody species were signifi- more important than season of burning for cantly associated with fire exclusion. Based on maintaining longleaf pine communities; sites work in prairies, which resemble longleaf pine with a history of frequent dormant season fire ground cover vegetation in their dominance by are likely to have retained most of the species bunch grasses, and abundance of composites found in a nearby reference site. If a substan- and legumes, Leach and Givnish (1996) con- tial period of fire exclusion has occurred, how- firmed higher than expected losses of nitrogen- ever, it is not likely that simply restoring the fixing legumes and small-seeded species, and fire regime will restore the herb layer. that losses were more pronounced on more productive sites. Regionally rare species were lost at a rate more than twice the average for Plantation Establishment and all species. We expect similar patterns for longleaf pine savannas. We do not know whether Management species associated with high fire frequencies The condition of the ground layer in planta- persist as a result of fire-associated vigor, or tions varies with land use history, site prepa- if persistence requires continued seedling es- ration methods, stand age, treatments applied tablishment in fire-created “safe sites” (sensu during stand development, and site type. A Harper 1977). history of machine tilling is likely to have the The most obvious and consistent effects greatest adverse impacts on the ground layer. of season of burning are effects on woody For example, Hedman et al. (2000) showed stems, with growing season fires being more that as little as 2 years of cultivation prior to site effective at reducing both size and density preparation and planting resulted in sites with compared to dormant season burning (Rob- reduced species richness and cover compared bins and Myers 1992 and references therein). to reference longleaf pine stands in southern Drewa et al. (2002) reported shrubs sprouted Georgia. Effects of other factors (stand age, more vigorously following dormant season canopy composition, site preparation, recent fires than growing season fires, and further, fire history) were small by comparison. that repeated growing season fires reduced the Site preparation effects vary with the type of size but not the number of stems of estab- method and with intensity. Mechanical meth- lished shrubs. Others have suggested that trees ods include treatments such as drum chopping once established are not easily removed by (crushing with a roller) and leaving the vege- fire (Rebertus et al. 1989; Platt et al. 1991; tation, or shearing (cutting at the ground level) Glitzenstein et al. 1995), even after 30 years of and piling the organic materials. Intensity may annual or biennial summer burning (Waldrop be increased, for example by weighting the et al. 1992). chopper and rolling the site more than once, Despite reports that growing season burning for more complete competition control. Me- stimulates flowering and increases synchrony chanical methods generally reduce the cover 304 IV. Restoration of both woody and herbaceous species directly, a surprisingly large number of the species but many individuals survive to resprout. All found in remnant forests on similar site types mechanical methods expose some mineral (Hedman et al. 2000; Smith et al. 2001). A soils, and may inadvertently redistribute top- study of xeric communities in the Carolina soil with nutrients and organic matter. Sandhills National Wildlife Refuge showed Herbicides (chemical methods) can be used that 40-year-old plantations had nearly iden- to target specific plant groups, and are effec- tical species-presence lists as remnant longleaf tively used to reduce woody species with pre- pine stands (Walker and van Eerden unpub- sumably low impacts on herbaceous species. lished data). Importantly, however, Aristida Treatments can be broadcast or applied to indi- stricta, the dominant bunch grass, and Gaylussa- vidual stems, further increasing the specificity cia dumosa, the second most abundant ground of applications. Chemical treatments do not cover species in remnant sites, were essen- disturb soil, reducing opportunities for weeds tially eliminated from plantations. Compared to become established. Both mechanical and to the xeric sandhills sites, more productive chemical methods are often coupled with fire, sites are likely to lose herb species and cover which maintains pine dominance and benefits relatively quickly, and to provide fertile ground herbaceous species, especially grasses. for weedy species (Smith et al. 2001). Regardless of the method, the general objec- In summary, except on old agricultural sites, tive of site preparation is to favor the establish- plantations are likely to support many char- ment and early growth of planted pines, and acteristic native species, and thus might be often results in increased herbaceous cover in restored without species additions. However, the first few years. Residual herbaceous species the loss of grasses and dominant ground cover may increase, and exposed mineral soil pro- species may limit the effectiveness of fire as a vides a seed bed for both weedy and desirable restoration tool. We do not know how long climax species (species of undisturbed longleaf populations of nonweedy species can persist communities). Because many climax species in longleaf pine plantations; thinning may in- do not have adaptations for rapid dispersal and crease their longevity, but early postestablish- establishment, the short-lived flush of herba- ment stands are most likely to have resid- ceous growth following site preparation is rel- ual populations to “rescue.” In this way they atively enriched with ruderal species and de- resemble sites where reduced fire frequency pleted of climax herbs (Swindel et al. 1986; was the primary disturbance. In other planta- Glitzenstein 1993). tions desirable trees may be present, but char- Herbaceous cover and richness tend to de- acteristic herbs missing, indicating the need cline with plantation age, and without inter- for modifying canopy structure and adding vening fire treatments herb cover can decline characteristic herbaceous species. The need to significantly by age six (Zutter and Miller 1998) establish or augment herbaceous species pop- while woody species increase. Additional sil- ulations makes them similar to conditions in vicultural treatments may reduce herbaceous agricultural sites. species, for example the reduction of pineland threeawn with fertilization (White 1977); or invigorate the herb layer, as by thinning (Gre- Agricultural Sites len and Enghardt 1973; Means 1997; Harring- Established pastures and recently cultivated ton and Edwards 1999). Over a range of site fields present a predictable condition nearly conditions, high herbaceous cover has been as- devoid of any vestiges of the former ground sociated with frequent burning and inversely cover. An agricultural history can have long- related to basal area of the canopy trees, sug- lasting impacts on the vegetation, soils, and mi- gesting the potential benefits of burning and croorganisms of other forest types (Foster et al. thinning in plantations to restore the ground 2003), but there is little information available layer (Hedman et al. 2000). The ground layer about the long-term effects of agricultural use in plantations (on untilled sites) often includes on longleaf pine systems and how those effects 10. Restoring the Ground Layer 305 impact restoration efforts. Much of what is Restoration Tasks known comes from project reports and proceedings of regional restoration conferences, Based on conditions described in the previous mainly from sites where the primary agricul- sections, we recognize some general conditions tural use was management of improved pas- requiring management action (Table 2). The ture. These sites are characterized by the com- canopy may be dominated by species other plete absence of longleaf pine, a depauperate than longleaf pine, and at altered densities to nonexistent native species pool, and domi- (often greater than the reference model) and nation by cultivated grasses such as bahia (Pas- distributions (more regular in pine plantations palum notatum) or early successional old-field than reference conditions). Similarly, the com- weeds. Consequently, most of the time, energy, position and structure of the ground layer and money invested in restoring such sites go vegetation may be changed, including species to eliminating the nonnative and undesirable richness and relative abundance of species. vegetation and establishing new populations Common or rare species may be absent; na- of longleaf pine ground cover species. tive ruderal species and exotics species may

TABLE 2. Summary of ecosystem changes that may have to be treated to achieve restoration goals. The necessity to treat any of these depends on speciﬁed project goals.

Does it matter for Alternatives: general biodiversity conservation Condition treatments Can fire alone fix it? and sustainability? Woody species Canopy/subcanopy Prescribed fire Depends: yes, if fuels Yes density higher Mechanical treatments adequate and long than reference Chemical treatments time; no, if no fuels and short time constraint Canopy composition Regenerate to longleaf No Yes, for long-term success altered pine; approaches may vary from clearcutting through progressive thinning and patch regeneration Herbs Absence or scarcity Prescribed fire No; possibly increase Yes of dominant or Direct seeding sparse population, common species Plant plugs but probably take a long time Absence or scarcity Prescribed fire Direct Probably not Sometimes; depends on of rare species seeding Plant plugs objectives for site Presence of Prescribed fire Depends on identity of Sometimes; depends on persistent weeds Hand/mechanical weeds and available nature of “weed,” but (natives) “weeding” Chemical time; some are really probably not treatments difficult Presence of exotic Prescribed fire Depends, but for noted Sometimes; depends on species Hand/mechanical species in longleaf nature of “weed,” but “weeding” Chemical pine systems, they probably should be treatments seem to tolerate fire remedied Site conditions Hydrology altered Restore drainage No Yes Soil structure/fertility Burn off excess organic Maybe Yes altered capital 306 IV. Restoration be present. Finally, the physical condition of shrub component. Fire, mechanical methods the site itself may have been altered, for ex- (e.g., felling, girdling, drum chopping, shear- ample by attempts to drain wet sites or to alter ing), and chemical (herbicide) methods are microsites for pine seedling establishment. All used. All of these treatments, alone or in com- changes in structure and composition may be binations, both reduce and control trees and combined with the elimination of fire. shrubs, and affect existing ground layer vege- It is not necessary to remedy all altered con- tation to varying degrees. ditions in order to achieve some restoration Fire is promoted as a “natural” method goals. The addition of rare species may be op- with positive benefits and most restoration tional, and should be pursued if restoring the practitioners and researchers concur that in biodiversity in a nature preserve is the goal; or some cases fire alone may restore canopy if their establishment supports a rare species structure and favorable conditions for ground conservation goal, and then only if postrestora- cover recovery, but that restoration will re- tion management can maintain high-quality quire multiple fires over relatively long times habitat conditions (Gordon 1994). The pres- (Robbins and Myers 1992; Waldrop et al. 1992; ence of native weeds may go untreated if they Glitzenstein 1993; Streng et al. 1993). Fac- are not aggressively displacing desired climax tors that limit the capacity for fire to restore species (D’Antonio and Meyerson 2002), but structure include a lack of fine fuels, pres- we favor eliminating exotic species where pos- ence of ladder fuels that may promote crown sible as ecological ramifications may not be damage, and thick duff that resists burning known at this time. when moist and kills trees when it does burn. Except for restoring altered physical con- The problem is particularly vexing when the ditions, such as site hydrology changed by site contains desirable old trees with heavy drainage ditches, prescribed burning is essen- duff accumulations at their bases (Varner et al. tial for rectifying and maintaining the restored 2000; Kush et al. 2004). Compared to the pre- condition of nearly all aspects of community sumed historical fire regime, a prescribed fire change (Table 2). Additional possible treat- regime for restoration may differ in seasonal- ments can be grouped based on two overall ity, frequency, and intensity, or be combined goals: (1) restore canopy structure to a condi- with pretreatments to protect desirable bio- tion that promotes ground layer establishment logical legacies like remnant old trees or trees and vigor and (2) start new populations or with red-cockaded woodpecker cavities (see augment existing populations of native ground Box 10.1). An initial series of cool, winter layer species. This chapter focuses on actions burns may effectively reduce duff accumula- required to restore the ground layer vegetations and protect old trees in fire-suppressed tion, and does not address establishing the stands (Kush et al. 2004). longleaf pine. However, approaches to restor- The effectiveness of fire for changing canopy ing the longleaf component can affect ground structure can be enhanced by combining burn- layer development, and successful restoration ing with mechanical and/or chemical treat- will require coordinating longleaf and ground ments (Tanner et al. 1988; Outcalt 1994; layer restoration. Walker and van Eerden 1998; Provencher et al. 2001; Kush et al. 2004; Walker et al. 2004). In general, mechanical or chemical treatments Changing Canopy reduce hardwoods and subsequent fires consume fuels and maintain hardwoods as basal Structure to Enhance sprouts. Ground Layer Vegetation The best-documented study of the effects of treatments to restore canopy structure in Aside from establishing longleaf pine in the a longleaf pine ecosystem was conducted canopy, the most common objective for canopy in a large-scale experiment in the sandhill management is to reduce a hardwood and communities at Eglin Air Force Base in the 10. Restoring the Ground Layer 307

Florida Panhandle (Provencher et al. 2001). story improvement, and were judged to be ef- The study plots were second-growth longleaf fective for restoring community structure in pine stands that had a long history of fire sup- the long term only if followed by prescribed pression. Consequently, the midstory had be- burning. come dominated by a variety of oak species Among mechanical options, drum chopping and there was a very sparse understory, with has been widely applied to reduce hardwoods, mats of hardwood leaf litter interspersed with especially small oaks (Quercus laevis, Q. in- bare ground. The goal of the study was to cana, Q. margaretta), and other woody species use management techniques commonly em- such as saw palmetto (Serenoa repens) (Tan- ployed to reduce hardwood midstory in lon- ner et al. 1988). Light chopping treatments gleaf pine systems, and to document their (single passes with an empty drum chopper) effects on both target (oak) and nontarget have short-lived impacts on dominant bunch (herbaceous) species, thereby testing the hy- grasses in dry sites (Aristida beyrichiana in pothesis that restoration of the habitat struc- Florida flatwoods [Grelen 1959] and A. stricta ture would be sufficient to return the under- in South Carolina sandhills [Walker and van story vegetation to reference conditions. Three Eerden 1998; Walker et al. 2004]), but more hardwood reduction treatments were used: intensive treatments are likely to substantially spring burning, application of a hexazinone reduce or eliminate the dominant wiregrass herbicide, and mechanical felling/girdling of (Grelen 1962; Moore 1974). hardwoods. The herbicide and felling/girdling Mechanical treatment effects vary with sea- treatments were followed by fuel reduction son of treatment, and when applied in the burns in the year after treatment. These plots same year as prescribed fire. In a field experi- were compared to both untreated controls ment at the Carolina Sandhill National Wildlife and reference plots to determine the effect Refuge prescribed fire (growing season, dor- of oak reduction treatments on herbaceous mant season, and no-burn treatments) was species richness and densities. They predicted combined with light drum chopping (growing an increase in plant species richness and in season, dormant season, and no-chop treat- densities of herbaceous plants that qualita- ments), and treatment effects on wiregrass and tively tracked increasing levels of hardwood turkey oak recovery were evaluated. Wiregrass reduction. recovered to pretreatment levels within two All treatments were effective for reducing growing seasons in all growing season burn oaks; however, 4 years after treatment, results treatments, regardless of chopping treatment; suggested that fire alone was the least effec- dormant season burn plots recovered more tive hardwood reduction method, but yielded slowly, the impact exacerbated by extreme the greatest increases in ground cover species drought (Walker et al. 2004; Fig. 3). The plots richness and densities. Brockway and Outcalt showing the slowest recovery were chopped (2000) similarly reported that hexazinone fol- after dormant season burning, perhaps as a re- lowed by burning more effectively reduced sult of chopping without a layer of pine straw turkey oak and shrub density and enhanced to protect wiregrass roots and crowns. Other ground layer recovery in an oak-dominated tools that cut or crush understory trees with- site than did burning alone. Provencher et al. out intense ground disturbance are available (2001) concluded that if gradual reduction of and we would expect similar effects. Chem- hardwood densities were acceptable, fire was ical applications can effectively reduce hard- an effective and cost-efficient means of hardwoods within pine stands on a range of site wood control and would benefit the ground conditions. Hexazinone formulations are of- layer vegetation. Chemical and mechanical ten used on oaks and shrubs typically found control were recognized to be viable options on mesic to dry sites; treatments for mesic for situations where hardwood reduction is sites are more variable including glyphosate, needed immediately, but they cost up to eight triclopyr, and 2,4-D formulations (Litt et al. times more than burning, showed less under- 2001). 308 IV. Restoration

FIGURE 3. Basal area of wiregrass (Aristida stricta) before and following experimental burning and drum chopping at the Carolina Sandhills National Wildlife Reserve, South Carolina. Study sites were longleaf pine saw timber or pine-scrub oak stands on Alpin soils. Three burn and three chopping (single pass, unweighted drum) treatments were deﬁned by season: dormant season (DS), growing season (GS), and not treated (that is, not burned [NB] or not chopped [NC]). Treatments: Uppercase letters indicate burn season; lowercase indicate chopping season. Measurements shown are pretreatment, after one growing season, and after two growing seasons. NB treatments did not change through time; GS burns recovered to pretreatment levels; DS burns had not recovered to pretreatment levels in two seasons (see Walker et al. 2004).

Although target woody species are success- herbicide, or combination of herbicides, used. fully controlled with chemical applications, Hexazinone herbicides were most widely used the effects on nontarget species of the ground as they are especially effective against com- layer are not fully understood. A recent review mon midstory hardwood species such as oaks, of herbicide application studies to determine sweetgum, and sumacs. effects on native, nontarget species (Litt et al. In flatwoods habitats, all herbicides used re- 2001 and references therein) found that ex- duced species richness and cover of herbaceous tremely variable treatments and their applica- and woody ground layer plants. Decreases tion mostly in plantations rather than natural ranged from 5.1% in herbaceous species rich- stands make it difficult to evaluate herbicide ness compared to control using a form of hex- use for restoration of ground layer vegetation. azinone, to 71.8% in total species richness us- The effects of herbicide use on the ground ing a mixture of three herbicides. The only layer varied with habitat and with the specific study to document vegetative cover reported 10. Restoring the Ground Layer 309 declines in the cover of both herbaceous application rate used. Even studies using the (27.2%) and woody (58.6%) vegetation after same herbicide have shown a wide range of re- hexazinone application. sponses, and responses within the same study In sandhills habitats, the effects were more can vary widely from year to year. At this varied. Woody cover and density were reduced point in time, it is difficult, if not impossi- 10.3% to 55.9% by hexazinone application, ble, to draw any general conclusion about the but woody biomass increased 105.3% with use effects of herbicide use on wiregrass, simply of 2,4-D. Graminoid density and cover gen- because the little information that we do have erally increased with hexazinone application, shows no consistent pattern. The same is true but the response of nongraminoid herbaceous for any individual species of interest, although plant cover ranged from a 49.8% increase to it is important to note that responses of three a 33% decrease with hexazinone use. The threatened species in Mississippi to herbicide response of ground layer species richness to use were all negative. herbicide use in sandhills was dependent on In summary, herbicide use is generally suc- the type of herbicide used and the application cessful in reducing mid- and understory hard- rate. Total species richness increased anywhere woods in all systems; however, there remain from 6.4% to 81%, while herbaceous species significant unknowns about impacts on na- richness was shown to increase 55.2% in the tive species, especially those in the herbaceous one study for which it was reported. ground layer. Additional well-designed stud- Among pine plantation studies, treatments ies in natural systems would provide much were especially variable with respect to both needed information and would be advised be- herbicide or herbicide combination used and fore large-scale application of herbicides is used application rate, making it difficult to de- as a restoration method, especially in those ar- scribe any general response patterns. Her- eas with remnant native plant populations. bicide treatments generally increased herba- Most studies of restructuring the canopy to ceous species richness (10.5% to 84.7%), restore diverse ground layer vegetation have and reduced woody species richness to vary- been conducted in comparatively dry sites. ing degrees, never exceeding a 17.2% de- It seems likely that higher productivity sites cline. Graminoid species richness increased by might differ in the following ways: need for 30.8% in one study, decreased by 16.7% in more frequent retreatment; need for more in- another, and showed intermediate responses tensive initial treatments relative to the period in the rest. Herbicides tended to decrease total of fire exclusion; greater challenge from exotic species richness in plantations, with declines species; more profound species losses because as much as 11.2% reported. However, triclopyr mesic sites will develop more intense competi- and glyphosate herbicide application increased tive species interactions, mesic sites have more total species richness by 10.9 and 8.7%, re- species to lose, and mesic sites have more rel- spectively. In plantation studies competition atively rare species which are prone to elimi- control was the motivation for herbicide use; nation (Leach and Givnish 1996). As a result however, differences between control of de- of more species losses, we suggest that mesic sirable ground layer plants and weeds are not sites are more likely to require species reintro- reported. Thus, it is impossible to determine ductions. the contribution of each group to the reported changes in species richness. Very few studies have reported herbicide ef- Plantation Restoration fects on individual species of concern, such as wiregrass or other herbaceous species. With re- Strategies spect to wiregrass, study results range from Several research groups have proposed strate- increases of up to 7480% to decreases of gies for restoring plantations. Based on the re- 142%, depending on the specific chemical and sults from an experiment to study the relative 310 IV. Restoration importance of light and soil water availability plantations as well as for restoring longleaf to and litterfall in limiting herbaceous density and sites currently planted in other pines. Methods cover in longleaf pine plantations, Harrington such as these could be especially helpful to land and Edwards (1999) concluded that conven- managers with responsibilities to recover red- tional silvicultural treatments, including thin- cockaded woodpecker populations challenged ning, herbicides, and prescribed fire, can be to restore both the canopy and ground layer of used to create a stand structure that favors existing plantations (U.S. Fish and Wildlife Ser- herb layer diversity and production. Although vice 2001). A gradual conversion and restora- the study was conducted in sites where the tion would retain the value of the plantation herbaceous composition differed substantially as woodpecker foraging habitat, while devel- from an undisturbed ground layer, they sug- oping future habitat. gest similar conditions would favor climax species. That assertion may be generally true, but we suspect that once established, climax Altering Species species would respond more slowly than oldfield species to changing conditions. Harring- Composition ton and Edwards (1999) caution that prescribed fire (every 2 to 3 years) and thinning Species composition may differ from refer- must be applied periodically to maintain an ence conditions by the absence of common open structure that favors herbs. Alternatively, or rare species, or the presence of weedy na- managing the stand for herbaceous layer di- tives or exotic species. In this section we focus versity and productivity could begin with site on starting new populations of native species, preparation, using herbicides (for control of although exotic species effects can pose sig- woody species) and prescribed fire to bene- nificant problems for ecological restoration in fit pine seedlings and existing herbs. Missing the longleaf pine system as elsewhere (Hobbs herbaceous species may be added at this time and Humphries 1995; D’Antonio and Meyer- to restore the composition of the herbaceous son 2002). For example, cogon grass (Imper- community. (See related information in the ata cylindrica), a well-studied exotic rhizoma- section on Direct Seeding.) tous grass invasive in the southern part of Kirkman and Mitchell (2002) describe a pro- the longleaf pine range, can displace native gressive thinning strategy to restore even-aged grasses and alter the fire regime because it slash pine plantations to multi-aged longleaf burns more intensively than the native bunch pine communities with diverse ground cover. grasses (Lippincott 2000; Jose et al. 2002). By The work is being conducted in an upland changing the fire regime, cogon grass has the site in southwest Georgia and in a flatwoods potential to alter patterns of species recruit- site in the Florida Panhandle. Gradual thin- ment and persistence through time. ning leaves pines producing litter to support Exotic species clearly challenge restoration surface fires and providing for future timber efforts, but an exhaustive treatment of the harvest, while creating conditions that favor topic is beyond the scope of this chapter. herbaceous species. This research group is in- For more details we refer the reader to the vestigating the effects of gap size, and of differ- rapidly expanding literature on exotic species, ent methods (including herbicide and mowing including excellent sources of information treatments) to control woody species growth for identification and control of exotic plant and promote a grassy ground layer. Treat- species (e.g., Miller 2003). Especially helpful ments also include seeding Aristida beyrichiana are websites devoted to management of non- in experimental gaps. Researchers will monitor native plants, including a site with informa- herb layer development with burning to de- tion from U.S. federal and state governments termine the need for additional species intro- (http://www.invasivespecies.gov), and from ductions. No results are published yet, but this the Nature Conservancy’s Invasive Species Ini- approach has promise for restoring longleaf tiative (http://tncweeds.ucdavis.edu). 10. Restoring the Ground Layer 311

TABLE 3. Comparison of direct seeding and outplanting options. Direct seeding Outplanting Advantages Advantages Economical ($3K/acre) Can choose individual target species Simultaneously introduce multiple species No need to disrupt existing conditions known to co-occur No special planting tools Can create custom seed mixes by varying Can be done on slopes where seeding equipment cannot be timing and methods of collection used safely Can be done concurrent with site Conducive to volunteer assistants preparation Good success for many species Can be done in winter, before frost, when Reduced susceptibility to drought at early stages competition for labor is lower Appropriate for rare species Mechanized approaches can treat large areas Few seeds are needed to ensure establishment objectives Genetically diverse seeds can be used so that Stock can be propagated any time when seed is available site conditions “select” most suitable Shorter period of competition control needed in many cases individuals Disadvantages Disadvantages Unreliable establishment requires large seed Expensive (up to $10K/acre) supplies Introduce only one species at a time Not as useful for rare species Available stock may be limited by the need for Special care needed to create seed mixes hand-collecting seed and size of nursery Seeding rates difﬁcult to determine to ensure Germination and initial establishment in greenhouse outcome conditions; may favor genotypes less suitable for future Competition control essential establishment in ﬁeld conditions

Options for Starting New seed collection, site preparation, and seeding. Populations Machine planting options for direct seeding make it possible to treat large areas, and when Options for starting new populations include seed mixes (mixed species, or mixed collec- direct seeding, out-planting nursery stock, or tions from more than one site) are used, estab- transplanting wild stock (Guerrant 1996). In lished individuals will represent genotypes that the context of biodiversity conservation the are successful as seedlings in field conditions latter approach is generally regarded as a rather than greenhouse conditions. Outplant- last resort, reserved for rescuing native plants ing approaches allow for selecting individual from sites destined for destruction, and will species (e.g., rare species), controlling the ge- not be addressed further. Both direct seed- netic composition of the new population, and ing and outplanting nursery stock have been for establishing plant cover quickly, but may be used successfully in longleaf pine restoration best suited for small areas. Native seed is not projects and have advantages and disadvan- available commercially, so seed may have to be tages (Table 3). Economic considerations give supplied to a grower for seedling production by direct seeding a clear advantage over plant- special order. ing plugs. Costs for using plugs, which in- Key issues associated with seeds include clude seed collection, nursery personnel, site what species to plant; where, how, and when preparation, and planting, can run from $3000 to collect seed; how to clean and store native (van Eerden, unpublished report to North Car- seed; seed viability, germination requirements, olina Department of Agriculture) to as high as and factors that affect seedling establishment. $12,000/acre (Seamon, in Disney Wilderness Preserve 2000); cost estimates for direct seeding were estimated at $155–650 and $300– Species Selection 400/acre at the same sites, respectively, and in- Criteria for species selection include: (1) the clude maintenance of the seed collection site, species’ habitat is similar to the restoration site 312 IV. Restoration conditions; (2) the restoration site is within the abundant flowering does not necessarily pre- natural distribution range of the target species; dict abundant seed production. (3) the species is needed to meet the project There are very few direct measurements of goal. For examples, restoring fine fuel produc- the magnitude of seed production in natu- tion, restoring the diversity of vascular plants ral populations of longleaf pine associates. In to a site, and restoring habitat for a rare butter- Pityopsis graminifolia (Brewer and Platt 1994) fly require different suites of species. In order and Aristida stricta (van Eerden 1997) viable to maintain a restoration project, burning must seed production (seeds/plant) following grow- be possible, and we recommend fuel producing season burns was significantly greater than tion, through the establishment of dominant after dormant season fires, while Hiers et perennial grasses and retention of onsite fuel al. (2000) reported that effects of burn sea- sources (e.g., pine straw), as an objective for all son on seed production in legumes varied restoration projects. If restoration goals do in- with species. Greater losses of some legume clude the introduction of uncommon species, species’ seed to predators occurred after win- there is as yet no general consensus about ter compared to growing season fires, but whether to add these species at the beginning that also varied among species (Hiers et al. of the restoration process (Weber 1999), or to 2000). wait to add them until after a matrix of dominant native species is established (Packard and Genetic Considerations Mutel 1997). Gordon (1994) developed a dichotomous key to support or guide manage- The genetic composition of populations at the ment decisions to introduce (or not) a native seed donor site can affect the success of the species. Although this tool highlights issues as- new population by providing genotypes suit- sociated with individual species, especially “at- able for the environmental conditions at the risk” species, some of them are directly applica- restoration sites. Donor composition can also ble to restoration, such as considering genetic affect the genetic structure of residual popula- and environmental suitability of the donor site, tions in or near the restoration site by introduc- considering impacts on any remnant popula- ing new genes and creating novel genotypes tions on the recipient site, and the potential via genetic recombination. To minimize po- for managing the site after the species intro- tentially adverse consequences, collection sites ductions. should be as near as possible and as similar as possible with respect to physical environment to the planting site. Seed Sources Widespread species, such as some of the dominant grasses, may harbor considerable ge- Abundant seed production is expected in sites netic diversity across their ranges (Hamrick et with abundant flowering, often resulting from al. 1991; Millar and Libby 1991). Based on burning in the current or previous growing morphological, geographic, and ecological fac- season (Platt et al. 1988; Robbins and My- tors, Peet (1993) divided A. stricta (sensu Rad- ers 1992; Streng et al. 1993). However, viable ford et al. 1968) into a more northerly species, seed production in native plant populations A. stricta, and more southerly taxon, A. beyrichi- is a complex process dependent on success- ana. Aristida stricta and A. beyrichiana dominate ful pollination, fertilization, seed development, the ground cover in many longleaf pine com- and seeds escaping predation or destruction munities, and despite disagreement as to the by pathogens. In theory, all of these processes taxonomic status (Walters et al. 1994; Kesler could be affected through various mechanisms et al. 2003), the bunchgrass is undoubtedly by season of burning. Independent of recent variable within its geographic range. Further, fire history, year-to-year and site-to-site vari- species with wide habitat tolerances within the ation in seed production is typical of natu- same landscape may exhibit ecotypic differen- ral plant populations (Fenner 1985). Thus, tiation, as Kindell et al. (1996) demonstrated 10. Restoring the Ground Layer 313 for A. beyrichiana. They reported a differential plant species (Center for Plant Conservation performance of seedlings from different habi- 1991). tats (xeric sandhills versus mesic flatwoods in north Florida) in common gardens and recip- rocal plantings, such that individuals grew bet- Seed Collection and Handling ter in sites similar to their habitat of origin. Timing Brewer (1995) showed that for the widespread Because each species has a specific phenol- Pityopsis graminifolia, individuals from differ- ogy, there is no one best time to collect seeds. ent locations, with potentially different eco- Plants that are ready for harvest have full- logically limiting conditions, responded differ- sized seeds with seed coats changing color, ently to fire. usually from green to a darker hue, and dry Matching environmental conditions of stems (Apfelbaum et al. 1997). Baskin and donor and recipient sites may be more im- Baskin (1998) recommend harvesting when portant for species that have limited potential the seeds would naturally disperse. Not only for gene flow among populations, including are there differences among species, but seeds shorter-lived rather than long-lived perenni- of the same species can mature both at differ- als, animal rather than wind pollinated, and ent times across the range of longleaf pine sys- species with no adaptations for widespread tems due to differences in climate and topog- dispersal (Hamrick et al. 1991). In such raphy, and at different times from year to year species, populations tend to be genetically due to variations in weather. Seed maturation distinct (compared to species with ample gene may also be affected by season of burning, but flow among populations), and consequently effects likely vary with species. Some guide- more finely adapted to local environmental lines for seed collection have been published, conditions. If no good collection site match such as those by Pfaff et al. (2002), which list is available, Huenneke (1991) suggests that collection dates for several species of grasses collecting from multiple suitable sites may and forbs. But, because of the variations listed be advantageous in producing a genetically above, these types of recommendations should diverse propagule mix, and thereby increas- be used with caution and paired with direct ob- ing the likelihood of including a suitable servation of the maturity of the plants. More environmental match to the restoration site. is known about the timing of seed collection Removing seed from a donor site may have for wiregrass than for most other species. Al- adverse effects at the donor site, particularly if though wiregrass seeds seem to ripen in midfall the persistence and structure of the commu- (October) there is an “after-ripening” effect, nity rely on frequent establishment from sex- such that seeds collected later in the fall and ual reproduction, or if collected species pro- into winter have higher germination rates. van vide critical food resources for native fauna. Eerden (1997) found that Aristida stricta seeds Because most species are perennial in this sys- collected in December had higher germination tem, and there have been few observations of rates than seeds collected from the same North seedling establishment, effects of periodic seed and South Carolina sandhill sites in Novem- removal on donor site composition are not ex- ber. Similar results were found for A. beyrichi- pected to be significant. Out of concern for po- ana collected in Georgia sandhills (Walker and tential adverse effects of collection, collectors Silletti unpublished data). often follow informal “rules” such as: take less than 50% of a strong perennial or less than Seed Harvest, Cleaning, Storage 10% of an annual; take only what you are prepared to handle responsibly; avoid tram- Methods for collecting, cleaning, and storing pling; collect as close to the restoration sites as seed for prairie restorations (Apfelbaum et al. is practical (Apfelbaum et al. 1997). The Cen- 1997; Clinebell 1997) are mostly applicable ter for Plant Conservation developed collection for seed handling for longleaf pine restoration guidelines for preserving the diversity of rare projects (Glitzenstein et al. 2001). Generally, 314 IV. Restoration seed can be collected by hand, which is espe- stripping individual seed heads by hand, to col- cially useful for rare or infrequent species or lecting entire infructescences with clippers, to when individual species are needed to enrich collecting small seeds with a hand-held vac- an existing site, or mechanically, which is very uum. effective for collecting seed mixtures (Fig. 4). Several types of seed harvesting ma- Hand collection methods vary from simply chines are available, but often prohibitively

FIGURE 4. Bulk seed collection with an ATV mounted seed stripper in a remnant upland site (a) and emptying the collection hopper into a storage bin (b). Seed collections include seed from all species in fruit at the collection time, e.g., common large grasses and composites. (Photo courtesy of Lin Roth.) 10. Restoring the Ground Layer 315 expensive. Sharing the cost of equipment may be useful for seed collected in the fall and provide a feasible option, but requires coop- intended for planting the following season eration in collection efforts. Green silage cut- (Glitzenstein et al. 2001; Pfaff et al. 2002). ters harvest and collect all aboveground plant However, Pittman and Karrfalt (2000) report material. The resulting “green” harvest con- that viability of wiregrass seed drops rapidly tains seeds as well as other vegetation and after 8 months of storage at ambient temper- must be distributed quickly so as to avoid seed- atures, and they used annually collected seed destroying mildew as the plants decompose. for seedling production. Pull-type and front-end mounted seed strip- pers harvest seeds plus accessory plant parts (mostly dry) with rotating brushes. Seed strip- Factors Affecting Germination per types, available in sizes suitable for mount- ing on four-wheelers and larger, have been and Establishment used in longleaf pine systems, but smaller mod- Properly collected seeds of many longleaf pine els are most convenient for harvesting seed associates readily germinate without elaborate from sites with trees. Pfaff et al. (2002) provide pretreatments. Germination rates across com- more details on seed harvester equipment and mon plant families are similar and highly vari- sources. able, ranging from zero to greater than 80% Seed harvests are processed to varying de- in laboratory, greenhouse, or outdoor trays grees depending on how long they are to be exposed to ambient environmental variations stored, how pure the seed must be, and such (Pfaff and Gonter 1996; van Eerden 1997; practical considerations as how much space is Glitzenstein et al. 2001; Pfaff et al. 2002). Re- available for storing (see Apfelbaum et al. 1997 sults of a study of 42 species characteristic of for details; Baskin and Baskin 1998). Except Atlantic coastal plain savannas indicate that for fleshy fruits, such as blueberries and huck- germination rates within a species vary from leberries common to the longleaf system, har- site to site and year to year, but are not re- vests are usually dried before processing fur- lated to time of burning or time since burning ther. [Fleshy fruits require special handling to (Glitzenstein et al. 2001). In that study, most separate seed and pulp. See Phillips (1985) for trials exceeded 30% germination. Glitzenstein suggestions.] Experience supports that collect- and colleagues compared germination rates in ing seeds on low-humidity days and spread- laboratory trials with germination in flats ex- ing them out of the weather in a warm place posed to outdoor conditions and found that for provides adequate drying. Collections gener- some species, especially fall-seeding compos- ally include a variety of other plant parts and ites, field germination exceeded lab trials. are cleaned in several stages (threshing, scalp- Several treatments have been reported to ing, final cleaning), depending on the desired increase germination rates in some common final condition. Methods can be simple such longleaf pine associates. Cold stratification in- as hand sorting, to more complex screening, creases germination in fall-fruiting composites and milling of various forms. (See Packard and such as Liatris spp. (Pfaff and Gonter 1996), Mutel 1997 for details and references.) and perennial grasses including Andropogon Seeds of many longleaf-associated native gerardii, Schizachyrium scoparium, Ctenium aro- species stored indoors in paper or grass seed maticum, Erianthus giganteus, Aristida beyrichi- bags retain viability for at least a year. ana (Glitzenstein et al. 2001), and A. stricta Glitzenstein et al. (2001) report acceptable ger- (van Eerden 1997). A period of after-ripening minability for 2 years, but much reduced via- reportedly benefits germination rates in Aris- bility and deformed seedlings after 2 years of tida beyrichiana, with germination increasing storage at room temperatures. Storing seeds for up to 5 months in dry storage (Pittman in dry unheated areas (e.g., unheated storage and Karrfalt 2000). Finally, heat treatments, shed) will expose seed to temperature vari- which can be as simple as pouring boiling ations similar to field conditions, and may water over seeds and allowing them to cool 316 IV. Restoration slowly, increase germination in various legume data; Fig. 5). The effects of pine straw litter species (Cushwa et al. 1968; Pfaff and Gonter on the germination and establishment of ten 1996; Baskin and Baskin 1998). Testing ger- common sandhills species were monitored for minability is the most reliable basis for calcu- one growing season. Germination of common lating seeding rates and for determining timing grasses and composites benefited from light lit- of harvest; but as a general rule, mature seeds ter. Heavy litter seemed to benefit germination should be planted at the same time they are among legumes, but overall germination rates naturally dispersed (Baskin and Baskin 1998). were low and at the end of the first season Delays in planting some species result natu- only one seedling of Baptisia cinerea remained rally in induced dormancy (Baskin and Baskin in a no-litter plot. In summary, available in- 1998). Consult references in this section for formation suggests that planting in an appro- more information about native seed germina- priate site with respect to a moisture gradient, tion and growing native species. with low competition, and low litter loads fa- Although seedlings of many native species vors seedling establishment under field condi- have been established and grown under green- tions. Similar conditions may be achieved via house and nursery conditions (Pfaff and site preparation for direct seeding projects. Gonter 1996; Glitzenstein et al. 2001; Da- gley et al. 2002; Pfaff et al. 2002) and even commercially produced (Pittman and Karrfalt Site Preparation and Sowing 2000), there is little information about factors that affect seedling establishment either Treatments from naturally dispersed seed in intact lon- The challenges for controlling competition gleaf communities or from seed introduced vary markedly with site history. As a general into field conditions. Experimental results sug- rule, in previously forested sites where ex- gest that seedling establishment in intact lon- otic herbaceous species are not dominant in gleaf pine communities is rare, and that gen- the ground layer, site preparation suitable for eral failure of seedling establishment can be at- planting longleaf pine seedlings will also fa- tributed to competition from established dom- vor ground cover establishment, as long as a inant species (Brewer et al. 1996; van Eerden sufficient amount of bare soil is available for 1997; Glitzenstein et al. 2001); higher estab- plant establishment. Experimental evidence lishment in mesic compared to xeric sites sug- suggests that more complete competition con- gests that competition for water may be the trol is likely to benefit seedling establishment. specific cause of mortality (van Eerden 1997; However, we observe acceptable wiregrass es- Glitzenstein et al. 2001). tablishment (about two clumps per square In a garden experiment using a variety of meter; from about 15.4 kg cleaned seed/ha; species, Glitzenstein et al. (2001) found that Walker and Silletti unpublished data) from each species was most successful in soil and broadcast seeding soon after planting trees in drainage conditions that most closely matched sandy sites where piling harvest slash left a the environments where it grows naturally. mosaic of bare soil, litter, and residual plant Thus, matching species and probably match- cover. Using a cultipacker to press seed into the ing seed source habitats for species found on a soil did not increase establishment success on broad environmental gradient (especially soil these uneven forest site surfaces. Acceptable moisture in the longleaf pine system) will most establishment was similarly achieved on mesic surely enhance establishment success. savanna sites at Fort Stewart, Georgia (Dena The presence of litter likely affects ger- Thomson personal communication). mination and establishment (Fowler 1986; Pasture lands and abandoned agricultural Facelli and Pickett 1991) and species-specific fields present the dual challenges of remov- responses to experimental litter treatments ing existing vegetation, including nonnative were observed at the Carolina Sandhills Na- perennial pasture grasses, and reducing the tional Wildlife Refuge (Walker unpublished numbers of weed seeds present in the soil 10. Restoring the Ground Layer 317

FIGURE 5. Effect of pine straw litter on seedling germination and establishment in selected common sandhills species. Small garden plots at the Carolina Sandhills National Wildlife Refuge were treated with a low litter (comparable to litter deposited in the ﬁrst year after burning) and heavy litter (twice the low level) application. Bars represent the mean of ﬁve replicates per species per treatment. Seedlings were counted in the same plots in four successive months during a single growing season. Number of seeds varied among species, but was constant within a species; thus internal comparisons can be made, but conclusions about species differences cannot be drawn. (Unpublished data).

(the seed bank). Recent projects have demon- ment protocol, additional weeds can be con- strated that removal of unwanted species takes trolled through periodic spot application of at least 1 year of treatment before planting na- herbicide and eventually reduced as they are tive species. The generally recommended pro- outcompeted by natives. tocol requires herbicide to remove all vege- Comprehensive studies of sowing treatment tation from the site (Disney Wilderness Pre- effects on establishment of wiregrass and serve 2000) followed 3 to 4 weeks later with other species were conducted at Apalachicola disking to expose weed seeds allowing them Bluffs and Ravines Preserve in north Florida to germinate. Disking is repeated every 4 to (Hattenbach et al. 1998; Seamon 1998; Cox et 6 weeks for about 6 months prior to planting. al. 2004). They examined the effects of eight Immediately before sowing desired species, the treatments in a three-factor experiment: sow- soil is compacted by rolling, and a ﬁnal her- ing native seed alone or with winter rye as bicide treatment is applied 2 to 3 weeks be- a cover crop, rolling the seed in after sow- fore planting. Variations on this protocol were ing, or not and adding supplemental water for shown to effectively prepare sites once domi- the ﬁrst 4 months after sowing or not. They nated by pasture grasses or with cogon grass found that neither supplemental water nor (Imperata cylindrica). If populations of native sowing an annual cover crop increased ground plants are quickly established after this treat- layer species richness or density. Rolling seeds 318 IV. Restoration

FIGURE 6. Standard hay blower distributing wiregrass seed at Fort Gordon, GA. This equipment is effective and widely available. in immediately after sowing, however, signifi- ficient distribution of native seed, especially cantly increased wiregrass establishment and wiregrass, is a standard hayblower (Fig. 6), survival, as others have reported (Pfaff and which allows for relatively even distribution Gonter 1996; Bissett 1998). of seed over a large area with some control Additions of fertilizer and mulch to sites after over seed placement (Disney Wilderness Pre- sowing are not recommended. Neither treat- serve 2000; Jones and Gordon unpublished re- ment significantly increases establishment of port to Florida Department of Transportation). native species under most conditions, but both See Pfaff et al. (2002) for a more detailed dis- favor the growth of native and exotic weeds cussion of seeding methods. that tend to outcompete natives (Bissett, in The question of at what rate to spread Florida Institute of Phosphate Research 1996; seed (or seed bearing material) highlights the Clewell, in Florida Institute of Phosphate Re- lack of consensus in longleaf pine restoration search 1996; Pfaff and Gonter 1996; Jones and projects. At a discussion of restoration meth- Gordon unpublished report to Florida Depart- ods at the Disney Wilderness Preserve Con- ment of Transportation; Jenkins et al. 2004). ference on Uplands Restoration, the range of seeding rates for seed stripper collected seed Direct Seeding: How Much, was 25 kg material/ha. The participants agreed on a recommendation of at least 56 kg/ha How, When? of material that is 10–11% wiregrass seed by Several options for sowing seed including weight, and 8–10% other seed by weight. Their hand broadcasting, hydroseeders, cultipack- goal was at least three established wiregrass ers, fluffy-seed drills, and fertilizer spreaders clumps per square meter. For material col- have been tested with varying results. The de- lected with a green silage cutter, they esti- vice most often recommended for quick, ef- mated that approximately 1500 kg/ha would 10. Restoring the Ground Layer 319 yield the desired establishment rates. One es- 1997; Outcalt et al. 1999) and competition timate for distributing clean seed was 2.2– (Mulligan and Kirkman 2002a). There is no 2.8 kg/ha (Disney Wilderness Preserve 2000). consensus on the best time of year to plant Other rates found in the literature include 133 seedlings; reported planting times ranged from kg/ha (Seamon 1998) for stripped material, April (Outcalt et al. 1999) to November (van 4.4–8.8 kg/ha of hand-collected and cleaned Eerden 1997, unpublished report to the North seed (van Eerden unpublished report to North Carolina Department of Agriculture) for wire- Carolina Department of Agriculture), 3.3–4.4 grass and plantings of other species occurred kg seed/ha (Pfaff et al. 2002), and 91 kg/ha throughout the year (Glitzenstein et al. 2001). (Hattenbach et al. 1998) which yielded 5 to Seedling density is also species dependent, but 7 plants/0.5 m2. Clearly, further experimental for wiregrass three plugs per square meter trials of seeding rates and resulting yields for is commonly used for experimental purposes different materials are needed. (Mulligan and Kirkman 2002a). Adequate soil moisture during early estab- In general, survivorship and growth of out- lishment is essential for native plant species. planted plugs in field situations are high, with Planting should occur just prior to the sea- survivorship rates of 90% (Glitzenstein and son of most reliable moisture (Pfaff and Gonter Streng, in Florida Institute of Phosphate Re- 1996); in most cases this is during the win- search 1996) and 60% (Outcalt et al. 1999) ter rainy season, from November to February after one growing season, and 80% after two (Pfaff and Gonter 1996; van Eerden 1997). In (Glitzenstein and Streng, in Florida Institute projects where soil moisture stress is severe, of Phosphate Research 1996). Both survivor- and the site is small enough to make it manage- ship and growth are reduced by competition able, irrigation can be applied for the first 3 to 4 from neighboring plants (Outcalt et al. 1999; months after planting to increase seedling es- Mulligan and Kirkman 2002a). Regarding un- tablishment (Jones and Gordon unpublished derplanting seedlings to restore plantations, report to Florida Department of Transporta- seedling performance is likely to be maximized tion; Jenkins et al. 2004). when planted in large canopy openings with minimal root competition from both woody and other herbaceous species (Dagley et al. Seedling Plugs: How Many, 2002) and low inputs of pine straw litter (van When, Where? Eerden 1997; Dagley et al. 2002). Seedlings for outplanting are best grown under Post-planting Management conditions that ensure adequate growth and survival, while maintaining an environment Prescribed fire is essential to encourage flow- that is stressful enough to select for stress- ering in many species and to control the tolerant plants and natural root-to-shoot ra- growth of woody and exotic species. Clewell tios. Glitzenstein et al. (2001) discuss consid- (in Florida Institute of Phosphate Research erations for cultivation including germination 1996) advocates burning as soon as the site and growth media (this can include horticul- is able to carry a fire; however, burning too tural media or soil taken from the restoration early can kill young wiregrass plants and slow site, which has the added advantage of provid- the growth of those that survive (Outcalt et al. ing mycorhizal innoculum), watering regime, 1999; Mulligan and Kirkman 2002b). Addi- and overwintering of seedlings, all of which tionally, reports indicate that wiregrass seeds prepare seedlings for successful outplanting. may remain dormant for a year after sowing The results of several studies suggest that, at and germinate in the second season (Seamon least for wiregrass, plugs should be at least 1998; Mulligan and Kirkman 2002b; Cox et al. 6 months old before they are outplanted be- 2004). It has therefore been recommended cause younger, smaller seedlings are more sus- that new plants be given at least one to two ceptible to the effects of drought (van Eerden complete growing seasons and as long as four 320 IV. Restoration to five seasons (Outcalt et al. 1999) before pre- the near future (for example, see Box 10.2). scribed fire is introduced. After that, a 2- to We expect that continued knowledge devel- 3-year burn cycle has been suggested, as com- opment would benefit from increased collabo- petition begins to negatively impact wiregrass ration and coordination among research and plants after 3 years (Glitzenstein and Streng, in restoration trials, and further that a widely Florida Institute of Phosphate Research 1996). accessible outlet for developing information will generate even more landowner interest in ground cover restoration. Population Establishment: Does In addition to the need for increased com- It Work? munication, we have identified some specific information gaps. Restoration research Population establishment can be considered a or adaptive restoration projects conducted in success when it results in a self-maintaining other locations within the range of longleaf population with sufficient genetic diversity for pine would advance the restoration cause, as long-term persistence (Pavlik 1996). While well as contribute to understanding the nat- few studies have been in place long enough ural variability of the ecosystem. In the ab- to reach this ultimate goal, there are many sence of more specific information, research encouraging results thus far. Seedling recruit- projects designed to understand the variation ment has been observed in populations of both in ecosystem functions across gradients, espe- outplanted wiregrass plugs (Mulligan et al. cially a productivity gradient, may be useful 2002) and in plots that were direct-seeded for targeting the most difficult and most press- (Bissett, in Florida Institute of Phosphate Re- ing restoration needs. There is always a need search 1996). Glitzenstein et al. (2001) mon- for more species-specific information about the itored six species of outplanted grasses and biology and habitat requirements of both com- forbs, including both wiregrass and the rare mon and rare species, especially regarding re- forb Parnassia caroliniana, for 5 years and productive biology. Because species reintro- are optimistic about their chances of long- ductions are needed for many sites, a more term success. At the Nature Conservancy comprehensive understanding of population Apalachicola Bluffs and Ravines Preserve seeds processes in experimental as well as natural were collected from 4- and 5-year-old direct- species matrices is essential. Information about seeded populations, and a direct seeding study persistent native seed banks is scarce (Cohen for the Florida DOT (Jones and Gordon un- 1998; Jenkins 2003), but would be especially published report to the Florida Department of useful in developing restoration protocols. Fi- Transportation) resulted in a stand that was nally, while there are suggestions that small within the natural range of species cover and fragments of this diverse herbaceous commu- able to carry fire in 3 years. nity can persist (Heuberger and Putz 2003), the effects of fragmentation and isolation on the persistence of the ground cover of longleaf Filling Information Gaps: are not well known; such knowledge could ensure that feasible restoration goals are estab- Adaptive Management and lished and that restoration resources are tar- Research geted where they can be successful.

Knowledge about restoring the ground layer in longleaf pine communities has increased sub- References stantially in recent years, and results from es- Apfelbaum, S.I., Bader, B.J., Faessler, F., and tablished projects promise a bright future for Mahler, D. 1997. Obtaining and processing restoration. Restoration projects and research seeds. In The Tallgrass Restoration Handbook, eds. efforts underway in various places through S. Packard and C.F. Mutel, pp. 99–126. Washing- the region will yield still more information in ton, DC: Island Press. 10. Restoring the Ground Layer 321

Baskin, C.C., and Baskin, J.M. 1998. Seeds: Ecology, Dagley, C.M., Harrington, T.B., and Edwards, M.B. Biogeography, and Evolution of Dormancy and Germi- 2002. Understory restoration in longleaf pine nation. San Diego: Academic Press. plantations: Overstory effects of competition and Beckage, B., and Stout, J.I. 2000. Effects of repeated needlefall. In Eleventh Biennial Southern Silvicul- burning on species richness in a Florida pine sa- tural Research Conference, ed. K.W. Outcalt, pp. vanna: A test of the intermediate disturbance hy- 487–489. Asheville, NC: USDA Forest Service, pothesis. J Veg Sci 11:113–122. Southern Research Station. Bissett, N.J. 1998. Direct seeding wiregrass, Aristida D’Antonio, C., and Meyerson, L.A. 2002. Exotic beyrichiana, and associated species. In Proceedings plant species as problems and solutions in ecolog- of the Longleaf Pine Ecosystem Restoration Symposium, ical restoration: A synthesis. Restor Ecol 10:703– The Longleaf Alliance Report No. 3, pp. 59–60. 713. Fort Lauderdale: The Longleaf Alliance. Disney Wilderness Preserve. 2000. Proceedings of the Brewer, J.S. 1995. The relationship between soil fer- Upland Restoration Workshop. Kissimmee, FL. tility and fire-stimulated floral induction in two Drew, M.B., Kirkman, L.K., and Gholson, populations of grass-leaved golden aster, Pityopsis A.K. 1998. The vascular flora of Ichauway, graminifolia. Oikos 74:45–54. Baker County, Georgia: A remnant longleaf Brewer, J.S., and Platt, W.J. 1994. Effects of fire sea- pine/wiregrass ecosystem. Castanea 63:1–24. son and herbivory on reproductive success in a Drewa, P.B., Platt, W.J., and Moser, E.B. 2002. Fire clonal forb, Pityopsis graminifolia. J Ecol 82:665– effects on resprouting of shrubs in headwaters 675. of southeastern longleaf pine savannas. Ecology Brewer, J.S., Platt, W.J., Glitzenstein, J.S., and 83:755–767. Streng, D.R. 1996. Effects of fire-generated gaps Egan, D., and Howell, E.A. 2001. The Historical Ecol- on growth and reproduction of golden aster (Pity- ogy Handbook. Washington, DC: Island Press. opsis graminifolia). Bull Torrey Bot Club 123:295– Ehrenfeld, J.G. 2000. Defining the limits of restora- 303. tion: The need for realistic goals. Restor Ecol 8:2–9. Brockway, D.G., and Lewis, C.E. 1997. Long-term Estill, J.C., and Cruzan, M.B. 2001. Phytogeography effects of dormant-season prescribed fire on plant of rare plant species endemic to the southeastern community diversity, structure and productivity United States. Castanea 66(1–2):3–23. in a longleaf pine wiregrass ecosystem. For Ecol Facelli, J.M., and Pickett, S.T. A. 1991. Plant litter: Its Manage 96:167–183. dynamics and effects on plant community struc- Brockway, D.G., and Outcalt, K. W. 2000. Restoring ture. Bot Rev 57:1–32. longleaf pine wiregrass ecosystems: Hexazinone Fenner, M. 1985. Seed Ecology. London: Chapman & application enhances effects of prescribed fire. For Hall. Ecol Manage 137:121–138. Florida Institute of Phosphate Research. 1996. Pro- Center for Plant Conservation. 1991. Genetic sam- ceedings of the Ecosystem Restoration Workshop, pling guidelines for conservation collections of Sponsored by the Florida Institute of Phosphate endangered plants. In Restoring Diversity, eds. D.A. Research and the Society for Ecological Restora- Falk, C.I. Millar, and M. Olwell, pp. 225–238. tion. Lakeland, FL. Washington, DC: Island Press. Foster, D., Swanson, F., Aber, J., Burke, I., Brokaw, Clinebell, R.R., II. 1997. Tips for gathering indi- N., Tilman, D., and Knapp, A. 2003. The impor- vidual species. In The Tallgrass Restoration Hand- tance of land-use legacies to ecology and conser- book, eds. S. Packard and C.F. Mutel, pp. 127–134. vation. BioScience 53:77–88. Washington, DC: Island Press. Fowler, N.L. 1986. Microsite requirements for ger- Cohen, S. 1998. Assessment of the restoration po- mination and establishment of three grass species. tential of severely disturbed longleaf pine sites. Am Midl Nat 115:131–145. Master’s thesis. North Carolina State University, Frost, C. 1998. Presettlement fire frequency regimes Raleigh. of the United States: A first approximation. Tall Cox, A.C., Gordon, D.R., Slapcinsky, J.L., and Timbers Fire Ecol Conf Proc 20:70–81. Seamon, G.S. 2004. Understory restoration in Gilliam, F.S.,and Christensen, N.L. 1986. Herb-layer longleaf pine sandhills. Nat Areas J 24:4– response to burning in pine flatwoods of the lower 14. coastal plain of South Carolina. Bull Torrey Bot Club Cushwa, C.T., Martin, R.E., and Miller, R.L. 1968. 113:42–45. The effects of fire on seed germination. J Range Gilliam, F.S., and Platt, W.J. 1999. Effects of long- Manage 21:250–254. term fire exclusion on tree species composition 322 IV. Restoration

and stand structure in an old-growth Pinus palus- Harper, J. 1977. Population Biology of Plants. New tris (longleaf pine) forest. Plant Ecol 140:15–26. York: Academic Press. Gilliam, F.S., Yurish, B.M., and Goodwin, L.M. Harrington, T.B., and Edwards, M.B. 1999. Under- 1993. Community composition of an old growth story vegetation, resource availability, and litter- longleaf pine forest: Relationship to soil texture. fall responses to pine thinning and woody vege- Bull Torrey Bot Club 120:287–294. tation control in longleaf pine plantations. Can J Glitzenstein, J.S. 1993. Panel discussion: Silvicul- For Res 29:1055–1064. tural effects on groundcover plant communities Hattenbach, M.J., Gordon, D.R., Seamon, G.S., and in longleaf pine forests. In Tall Timbers Fire Ecology Studenmund, R.G. 1998. Development of direct- Conference No. 18, ed. S.M. Hermann, pp. 357–370. seeding techniques to restore native groundcover Tallahassee: Tall Timbers Research Station. in a sandhill ecosystem. In Proceedings of the Lon- Glitzenstein, J.S., Platt, W.J., and Streng, D.R. 1995. gleaf Pine Ecosystem Restoration Symposium, The Lon- Effects of fire regime and habitat in three dynam- gleaf Alliance Report No. 3, pp. 64–70. Fort Laud- ics in north Florida longleaf pine savannas. Ecol erdale: The Longleaf Alliance. Monogr 65:441–476. Hedman, C.W., Grace, S.L., and King, S.E. 2000. Glitzenstein, J.S., Streng, D.R., Wade, D.D., and Vegetation composition and structure of southern Brubaker, J. 2001. Starting new plants of lon- coastal plain pine forests: An ecological compari- gleaf pine ground-layer plants in the outer coastal son. For Ecol Manage 134:233–247. plain of South Carolina, USA. Nat Areas J 21:89– Heuberger, K.A., and Putz, F.E. 2003. Fire in the 110. suburbs: Ecological impacts of prescribed fire in Glitzenstein, J.S., Streng, D.R., and Wade, D.D. small remnants of longleaf pine (Pinus palustris) 2003. Fire frequency effects on longleaf pine (Pi- sandhill. Restor Ecol 11:72–81. nus palustris Miller) vegetation in South Carolina Hiers, J.K., Wyatt, R., and Mitchell, R.J. 2000. The and northeast Florida, USA. Nat Areas J 23:22– effects of fire regime on legume production in lon- 37. gleaf pine savannas: Is a season selective? Oecolo- Gordon, D.R. 1994. Translocation of species into gia 125:521–530. conservation areas: A key for natural resource Hobbs, R., and Humphries, S.E. 1995. An integrated managers. Nat Areas J 14:31–37. approach to the ecology and management of plant Grelen, H.E. 1959. Mechanical preparation of pine invasions. Conserv Biol 9:761–770. planting sites in Florida sandhills. Weeds 7:184– Huenneke, L.F. 1991. Ecological implications of ge- 188. netic variation in plant populations. In Genetics Grelen, H.E. 1962. Plant succession on cleared sand- and Conservation of Rare Plants, eds. D.A. Falk and hills in northern Florida. Am Midl Nat 67:36–44. K.E. Holsinger, pp. 31–44. Oxford: Oxford Uni- Grelen, H.E., and Enghardt, H.G. 1973. Burning and versity Press. thinning maintain forage in a longleaf pine plan- Jenkins, A.M. 2003. Seed banking and vesicular- tation. J For 71:419–425. arbuscular mycorrhizae in pasture restoration Guerrant, E.O., Jr. 1996. Designing populations: in central Florida. Master’s thesis. University of Demographic, genetic, and horticultural dimen- Florida, Gainesville. sions. In Restoring Diversity, eds. D.A. Falk, C.I. Mil- Jenkins, A.M., Gordon, D.R., and Renda, M.T. lar, and M. Olwell, pp. 171–208. Washington, DC: 2004. Native alternatives for nonnative turf- Island Press. grasses in central Florida: Germination and re- Hamrick, J.L., Godt, M.J.W., Murawski, D.A., sponses to cultural treatments. Restor Ecol 12:190– and Loveless, M.D. 1991. Correlations between 199. species traits and allozyme diversity: Implications Jose, S., Cox, J., Miller, D.L., Shilling, D.G., and Mer- for conservation biology. In Genetics and Conser- ritt, S. 2002. Alien plant invasions: The story of vation of Rare Plants, eds. D. A. Falk and K.E. cogongrass in southeastern forests. J For 100:41– Holsinger, pp. 75–86. Oxford: Oxford University 44. Press. Kesler, T.R., Anderson, L.C., and Hermann, S.M. Harcombe, P.A., Glitzenstein, J.S., Knox, R.G., 2003. A taxonomic reevaluation of Aristida stricta Orzell, S.L., and Bridges, E.L. 1993. Vegetation of (Poaceae) using anatomy and morphology. South- the longleaf pine region of the west gulf coastal east Nat 2:1–10. plain. In Tall Timbers Fire Ecology Conference No. 18, Kindell, C.E., Winn, A.A., and Miller, T.E. 1996. ed. S.M. Hermann, pp. 83–104. Tallahassee: Tall The effects of surrounding vegetation and trans- Timbers Research Station. plant age on the detection of local adaptation in 10. Restoring the Ground Layer 323

the perennial grass Aristida stricta. J Ecol 84:745– Mulligan, M.K., and Kirkman, L.K. 2002b. Burn- 754. ing influences on wiregrass (Aristida beyrichiana) Kirkman, L.K., and Mitchell, R.J. 2002. A forest restoration plantings: Natural seedling recruit- gap approach to restoring longleaf pine-wiregrass ment and survival. Restor Ecol 10:334–339. ecosystems (Georgia and Florida). Ecol Restor Mulligan, M.K., Kirkman, L.K., and Mitchell, R.J. 20:50–51. 2002. Aristida beyrichiana (wiregrass) establish- Kush, J.S., Meldahl, R.S., and Avery, C. 2004. A ment and recruitment: Implications for restora- restoration success: Longleaf pine seedlings estab- tion. Restor Ecol 10:68–76. lished in a fire-suppressed, old-growth stand. Ecol Outcalt, K.W. 1994. Seed production of wiregrass Restor 22: 6–10. in central Florida following growing season pre- Leach, M.K., and Givnish, T.J. 1996. Ecological described burns. Int J Wildland Fire 4:123–125. terminants of species loss in remnant prairies. Sci- Outcalt, K.W., Williams, M.E., and Onokpise, O. ence 273:1555–1558. 1999. Restoring Aristida stricta to Pinus palustris LeBlond, R.J. 2001. Endemic plants of the Cape Fear ecosystems on the Atlantic coastal plain, USA. Arch region. Castanea 66: 83–97. Restor Ecol 7:262–270. Lippincott, C.L. 2000. Effects of Imperata cylindrica Packard, S., and Mutel, C.F. 1997. The Tallgrass (cogongrass) invasion on fire regime in Florida Restoration Handbook. Washington, DC: Island sandhill (USA). Nat Areas J 20:140–149. Press. Litt, A.R., Herring, B.J., and Provencher, L. 2001. Pavlik, B.M. 1996. Defining and measuring success. Herbicide effects on ground-layer vegetation in In Restoring Diversity, eds. D.A. Falk, C.I. Millar, southern pinelands, USA: A review. Nat Areas J and M. Olwell, pp. 127–156. Washington, DC: Is- 21:177–188. land Press. Liu, C., Harcombe, P.A., and Knox, R.G. 1997. Ef- Peet, R.K. 1993. A taxonomic study of Aristida stricta fects of prescribed fire on the composition of and A. beyrichiana. Rhodora 95:25–37. woody plant communities in southeastern Texas. Peet, R.K., and Allard, D.J. 1993. Longleaf pine veg- J Veg Sci 8:495–504. etation of the southern Atlantic and eastern Gulf Lockwood, J.L., and Pimm, S.L. 1999. When does Coast regions: A primary classification. In Tall Tim- restoration succeed? In Ecological Assembly Rules, bers Fire Ecology Conference, no. 18 ed. S.M. Her- eds. E. Weiher and P. Keddy, pp. 363–392. Cam- mann, pp. 45–81. Tallahassee: Tall Timbers Re- bridge: Cambridge University Press. search Station. Means, D.B. 1997. Wiregrass restoration: Probable Pfaff, S., and Gonter, M.A. 1996. Florida native plant shading effects in a slash pine plantation. Restor collection, production, and direct seeding tech- Manage Notes 15:52–55. niques: Interim report. Brooksville, FL: USDA, Mehlman, D.W. 1992. Effects of fire on plant com- Natural Resources Conservation Service, Plant munity composition of North Florida second Materials Center. growth pineland. Bull Torrey Bot Club 119:376– Pfaff, S., Gonter, M.A., and Maura, C. 2002. Florida 383. Native Seed Production Manual. Brooksville, FL: Millar, C.I., and Libby, W.J. 1991. Strategies for con- USDA, Natural Resources Conservation Service, serving clinal, ecotypic, and disjunct population Plant Materials Center. diversity in widespread species. In Genetics and Phillips, H.R. 1985. Growing and Propagating Wild Conservation of Rare Plants, eds. D.A. Falk and K.E. Flowers. Chapel Hill: University of North Carolina Holsinger, pp. 149–170. Oxford: Oxford Univer- Press. sity Press. Pittman, T., and Karrfalt, R.P. 2000. Wiregrass prop- Miller, J.H. 2003. Nonnative Invasive Plants of South- agation. Native Plants J 1:45–47. ern Forests: A Field Guide for Identification and Con- Platt, W.J., Evans, G.W., and Davis, M.M. 1988. trol. GTR SRS–62. Asheville: USDA Forest Service, Effects of fire season on flowering of forbs and Southern Research Station. shrubs in longleaf pine forests. Oecologia 76:353– Moore, W.H. 1974. Some effects of chopping 363. saw-palmetto–pineland threeawn range in south Platt, W.J., Glitzenstein, J.S., and Streng, D.R. 1991. Florida. J Range Manage 27:101–104. Evaluating pyrogenicity and its effects on vegeta- Mulligan, M.K., and Kirkman, L.K. 2002a. tion in longleaf pine savannas. Proc Tall Timbers Competition effects on wiregrass (Aristida Fire Ecol Conf 17:143–161. beyrichina) growth and survival. Plant Ecol 163:39– Provencher, L., Herring, B.J., Gordon, D.R., Rogers, 50. H.L., Galley, K.E. M., Tanner, G.W., Hardesty, J.L., 324 IV. Restoration

and Brennan, L.A. 2001. Effects of hardwood Forest Ecology: Innovative Silvicultural and Vege- reduction techniques on longleaf pine sandhill tation Management, Tall Timbers Fire Ecology Con- vegetation in northwest Florida. Restor Ecol 9:13– ference Proceedings No. 21, eds. W. K. Moser and 27. C.E. Moser, pp. 216–219. Tallahassee: Tall Timbers Radford, A.E., Ahles, H.E., and Bell, R.C. 1968. Man- Research Station. ual of the Vascular Flora of the Carolinas. Chapel Hill: Varner, J.M., III, Kush, J.S., and Meldahl, R.S. 2003. University of North Carolina Press. Vegetation of frequently burned old-growth lon- Rebertus, A.J., Williamson, G.B., and Moser, E.B. gleaf pine (Pinus palustris Mill.) savannas on Choc- 1989. Longleaf pine pyrogenicity and turkey oak colocco Mountain, Alabama, USA. Nat Areas J mortality in Florida xeric sandhills. Ecology 70:60– 23:43–52. 70. Waldrop, T.A., White, D.L., and Jones, S.M. 1992. Robbins, L.E., and Myers, R.L. 1992. Seasonal ef- Fire regimes for pine-grassland communities in fects of prescribed burning in Florida: A review. the southeastern United States. For Ecol Manage Miscellaneous Publication No. 8, Tall Timbers Re- 47:195–210. search, Inc., Tallahassee, FL. Walker, J.L. 1998. Ground layer vegetation in lon- Seamon, G. 1998. A longleaf pine sandhill restora- gleaf pine landscapes: An overview for restoration tion in northwest Florida. Restor Manage Notes and management. In Proceedings of the Longleaf Pine 16:46–50. Ecosystem Restoration Symposium, The Longleaf Al- Smith, G.P., Shelburne, V.B., and Walker, J.L. 2001. liance Report No. 3, pp. 2–13. Fort Lauderdale: Structure and composition of vegetation of lon- The Longleaf Alliance. gleaf pine plantations compared to natural stands Walker, J., and Peet, R.K. 1983. Composition occurring along an environmental gradient at the and species diversity of pine–wiregrass savannas Savannah River Site. In Eleventh Biennial Southern of the Green Swamp, North Carolina. Vegetatio Silvicultural Research Conference, ed. K.W. Outcalt, 55:163–179. pp. 481–486. Ashville, NC: USDA Forest Service, Walker, J.L., and van Eerden, B.P. 1998. Effects of Southern Research Station. drum-chopping on wiregrass and other herba- Sorrie, B.A., and Weakley, A.S. 2001. Coastal plain ceous species in xeric sandhill sites. In Fire in vascular plant endemics: Phytogeographic pat- Ecosystem Management: Shifting the Paradigm from terns. Castanea 66:50–82. Suppression to Prescription, Tall Timbers Fire Ecol- Streng, D.R., Glitzenstein, J.S., and Platt, W.J. 1993. ogy Conference Proceedings No. 20, eds. T.L. Pru- Evaluating effects of season of burn in longleaf den and L.A. Brennan, pp. 118. Tallahassee: Tall pine forests: A critical literature review and some Timbers Research Station. results from an ongoing long-term study. In Tall Walker, J.L., van Eerden, B.P., Robinson, D., Timbers Fire Ecology Conference, ed. S.M. Hermann, and Hausch, M. 2004. Burning and chop- pp. 227–263. Tallahassee: Tall Timbers Research ping for woodpeckers and wiregrass? In Red- Station. cockaded Woodpecker: Road to Recovery, Proceed- Swindel, B.F., Conde, L.F., and Smith, J.E. 1986. ings of the Red-cockaded Woodpecker Re- Successional changes in Pinus elliottii plantations covery Symposium IV, eds.R. Costa and S.J. following two regeneration treatments. Can J For Daniels, pp. 683–685. Savannah, GA, January Res 16:630–636. 27–31, 2003. Hancock House Publishers, Blaine, Tanner, G.W., Terry, W.S., and Kalmbacher, R.B. WA. 1988. Mechanical shrub control on flatwoods Walters, T.W., Decker-Walters, D.S., and Gordon, range in south Florida. J Range Manage 41:245– D.R. 1994. Restoration considerations for wire- 248. grass (Aristida stricta): Allozymic diversity of pop- U.S. Fish and Wildlife Service. 2001. Red-cockaded ulations. Conserv Biol 8:581–585. woodpecker recovery plan (Picoides borealis). At- Wear, D.N., and Greis, J.G. 2002. Southern Forest lanta: U. S. Fish and Wildlife Service. Resource Assessment. General Technical Report van Eerden, B.P. 1997. Studies on the reproductive SRS-53. Asheville, NC: USDA Forest Service. biology of wiregrass (Aristida stricta Michaux) in Weber, S. 1999. Designing seed mixes for prairie the Carollina sandhills. Master’s thesis, University restorations: Revisiting the formula. Ecol Restor of Georgia, Athens. 17:196–201. Varner, J.M., III, Kush, J.S., and Meldahl, R.S. 2000. White, L.D. 1977. Forage production in a five-year- Ecological restoration of an old-growth longleaf old fertilized slash pine plantation. J Range Manage pine stand utilizing prescribed fire. In Fire and 30:131–134. 10. Restoring the Ground Layer 325

White, P.S. 1996. Spatial and biological scales in erence information in restoration ecology. Restor reintroduction. In Restoring Diversity, eds. D.A. Ecol 5:338–349. Falk, C.I. Millar, and M. Olwell, pp. 49–86. Wash- Zutter, B.R., and Miller, J.H. 1998. Eleventh-year ington, DC: Island Press. response of loblolly pine and competing vegeta- White, P.S., and Walker, J.L. 1997. Approximat- tion to woody and herbaceous plant control on a ing nature’s variation: Selecting and using ref- Georgia ﬂatwoods site. South J Appl For 22:88–95. 326 IV. Restoration

BOX 10.1 Underburning can be done using heading, backing, flanking, or spot fires, or a Prescribed Burning for combination of these techniques. Heading fires are fire fronts ignited to spread with the Understory Restoration wind while backing fires are ignited so the Kenneth W. Outcalt fire front spreads against the wind. Flank- ing fires are ignited in a line into the wind Southern Research Station, and thus spread at approximately right an- USDA Forest Service, Athens, gles to wind direction. Spot fires are a series Georgia 30602 of separate ignition points that are allowed to spread in all directions and thus contain Role of Prescribed Burning. Because the long- heading, backing, and flanking fires at each leaf ecosystem evolved with and is adapted spot. Both heading and backing fires can be to frequent fire, every 2 to 8 years, preset as a series of strip fires. Strip heading fires scribed burning is often useful for restor- are used to control how fast the fire spreads ing understory communities to a diverse and thereby the fireline intensity, i.e., the ground layer of grasses, herbs, and small rate of heat energy release (Box A Fig. 1). shrubs. This restoration provides habitat for Placing strips closer together reduces the a number of plant and animal species that rate of spread and intensity. Backing fires are restricted to or found mostly in longleaf have low intensities but move slowly and pine communities. Burning can also be used therefore require a lot of time to burn each to reduce the midstory layer, which catches unit. In addition, backing fires under cer- shed needles and serves as a ladder to carry tain conditions may be quite severe, i.e., understory fires into the crowns of the trees cause much of damage to the site, because resulting in catastrophic wildfires that can of excess duff consumption. Internal fire- kill vast areas of pines. Prescribed burning breaks can be constructed for strip backing also recycles nutrients by releasing those fires to significantly reduce time to complete tied up in litter and duff and significantly the burn. An alternative is to use flanking reduces brown spot needle blight, which at- or spot fires to reduce intensity but speed tacks longleaf seedlings. up the burn without internal fire breaks. Terms and Techniques. Prescribed burn- These techniques require considerable ex- ing is the application of fire by trained perience, especially spot firing as you must professionals following a well-developed continually adjust both the spacing and the plan to obtain desired management objec- timing between spots to obtain the desired tives. Restoration is often done with un- intensity with changing fuel and weather derstory burning or underburning, which conditions (Wade and Lunsford 1989). is prescribed burning under a forest canopy (McPherson et al. 1990). The fuel for these Burning Prescriptions fires is the understory rough that consists of the accumulated living and dead grasses, Sandhills. Prescribed burning can be used forbs and shrubs plus draped needles and for restoration across the range of sites that the litter layer. The litter layer, the top longleaf can occupy from dry sandhills to layer of the forest floor, is composed of wet savannas. Burning prescriptions de- recently fallen and largely intact dead need- pend on the ecosystem type and its cur- les, leaves, twigs, and branches. A duff layer, rent condition. Reduced fire frequency in composed of partially decomposed litter or many xeric and subxeric sandhills longleaf fermentation layer and decomposed hu- areas has resulted in the development of a mus, lies between the litter and mineral midstory layer of native scrub oaks: turkey soil. (Quercus laevis), bluejack (Q. incana), sand 10. Restoring the Ground Layer 327

FIGURE 1. Schematic of typical strip heading fire showing securing downwind side with backfire followed by sequential ignition of strip heading fires with a flanking fire to widen fuel free zone along plow line.

live oak (Q. virginiana var. geminata), and emerge from the roots of many top-killed sand post oak (Q. stellata var. margaretta). stems, but these can be kept in check by Because these sites are very droughty and subsequent periodic burns. nutrient limited, even in the absence of Flatwoods. On flatwoods and wet lowland frequent burning, they do not develop pine types restoration means increasing un- a continuous closed canopy of midstory derstory diversity in longleaf communities hardwoods. Therefore, although greatly re- that have been captured by woody species duced, some of the understory grasses do and in many cases developed a substantial survive. These grasses along with needle midstory layer. The goal is to reduce woody litter from longleaf pines furnish sufficient understory and midstory species and allow fuel to carry at least a patchy prescribed the grasses and forbs to increase and even- burn. Repeated applications of prescribed tually become at least co-dominant. Pre- fires during the growing season, i.e., begin- scribed fire can be used to accomplish this ning in March, in southern latitudes, and transition. Research shows that although ending in July, can be used to restore these growing season burns are sometimes more sites by gradually reducing the density of effective, dormant season burns can also the midstory scrub oaks (Glitzenstein et al. be used to readjust understory composi- 1995) and promoting the growth of un- tion (Waldrop et al. 1987). For areas not derstory grasses and herbs. Managers have burned for 10 years or more, a couple of found that fire causes wounds on the stems dormant season burns should be used to re- of hardwoods, which are enlarged by subse- duce fuel loads before switching to grow- quent fires, and eventually the top breaks or ing season burns. In addition, it is usually the stem is girdled and the top dies. Sprouts best to have these burns close together, i.e., 328 IV. Restoration

2 years or less, to minimize fuel accumu- cambial cells beyond the lethal temperature. lations between burns. Miller and Bossuot Trees suffering from such injury often retain (2000) recommend these initial burns be a healthy-looking green crown for some conducted when the drought index is below time following the burn, but will eventually 250. On sites dominated by saw palmetto die. Longleaf communities needing restora- (Serenoa repens), if burning alone is going to tion burning rarely have many seedlings. If be used for understory restoration, then a significant numbers of seedlings are present, series of closely spaced prescribed burns is however, excessive seedling mortality can required. Frequency of burns is more im- result from burning during the bolting stage. portant than season with annual burns the If burning must be done with seedlings at most effective but biennial burns will re- this stage, then burning should be done in duce palmetto-dominance and increase the the dormant season or early spring prior to herbaceous component. It is important not the candle stage when seedlings would be to miss a burn, as this can result in a signif- most susceptible to fire-caused damage and icant regrowth of palmetto. mortality. Uplands. There also exist upland long- Precautions. In all longleaf types that have leaf types, mostly in Alabama, Mississippi, not been burned for 10 years or more, there Louisiana, and Texas, and montane sites in is an excessive buildup of litter and duff Alabama and Georgia, which have devel- around the base of trees. Reintroduction of oped unnaturally dense hardwood midsto- burning in these stands without excess mor- ries, and suppressed and impoverished un- tality is best accomplished by a series of dor- derstories. Because these are the most pro- mant season burns. Apply burns when only ductive longleaf sites, they change the most the litter is dry enough to burn and the duff rapidly, quickly developing midstory layers is too wet to ignite. On upland or sandhills in the absence of frequent fire. In addi- sites, fast-moving heading or flanking fires tion to a very dense midstory and a shrub- pushed by a good wind are better than slow dominated understory, these sites also ac- backing fires that may dry the duff layer cumulate significant quantities of fuel. As and promote smoldering combustion. Flat- noted for other longleaf ecosystem types, woods sites with palmetto-dominated un- a series of dormant season burns is of- derstories should be burned with heading ten necessary to gradually reduce fuel lev- fires, but will also require a light wind, cool els before switching to a growing sea- temperature, and higher humidity for the son regiment. However, frequent and mul- first burn. The objective on all sites is to tiple growing season burns will be re- consume the dry top litter layer while the quired to reduce the hardwood rootstocks wet lower duff layer will protect the roots (Boyer 1990), thereby providing condi- and root collar. Space burns as closely to- tions favorable to understory grasses and gether as fuel to carry the fire will allow. forbs. Be cautious in your prescriptions because a Potential Negative Impacts. As with all burn- patchy burn is preferable to a more coming, there is the potential for negative im- plete hot burn that could result in exces- pacts. The most obvious damage is direct sive tree mortality. The objective is to grad- tree mortality that can result from exces- ually reduce the duff layer at the base of sively hot burns that are too intense and kill trees over a cycle of four or five fires and tree crowns, including the buds. Tree mor- keep tree mortality at an acceptable level. tality can also occur with low-intensity but Once the excess duff layer is removed, ap- high-severity fires that slowly consume ac- ply a growing season burn, again as soon cumulated forest floor duff, and because of as there is sufficient fuel to carry a good their long residence time heat root and stem fire. 10. Restoring the Ground Layer 329

The most important factor in accomplish- USDA Forest Service, Southern Forest Exper- ing the goal of successful restoration with iment Station, Research Paper SO-256, New fire while minimizing the negative conse- Orleans, LA. quences is experience. Only through train- Glitzenstein, J.S., Platt, W.J., and Streng, D.R. ing and practice can you become proficient 1995. Effects of fire regime and habitat on tree at selecting proper conditions of tempera- dynamics in north Florida longleaf pine savannas. Ecol Monogr 65(4):441–476. ture, humidity, wind, fuel moisture, and fir- McPherson, G.R., Wade, D.D., and Phillips, C.B. ing techniques keyed to existing fuel types 1990. Glossary of Wildland Fire Management and loads that are likely to produce the Terms Used in the United States. Washington, DC: desired outcomes. This means obtaining a Society of American Foresters. contract burner from a consulting forestry Miller, S.R., and Bossuot, W.R. 2000. Flatwoods business or an experienced crew from a restoration on the St. Johns River Water Man- nonprofit or government agency until you agement District, Florida: A prescription to gain knowledge and experience needed to cut and burn. In Proceedings of the 21st Tall be a certified burner. A source of informa- Timbers Fire Ecology Conference, pp. 212–215. tion is the U. S. Department of Agriculture, Tall Timbers Research Station, Tallahassee, Forest Service, Southern Research Station FL. Wade, D.D., and Lunsford, J.D. 1989. A (http://www.srs.fs.usda.gov), which has a guide for prescribed fire in southern forests. number of relevant publications. Consider- USDA Forest Service, Southern Region, able advice and guidance is also available Technical Publication R8-TP 11, Atlanta, from state forestry agencies, forestry units GA. of southern universities, and local extension Waldrop, T.A., Van Lear, D.H., Lloyd, F.T., and agents. Harms, W.R. 1987. Long-term studies of prescribed burning in loblolly pine forests of the References Southeastern Coastal Plain. USDA Forest Ser- vice, Southeastern Forest Experiment Station, Boyer, W.D. 1990. Growing-season burns for General Technical Report SE-45, Asheville, control of hardwoods in longleaf pine stands. NC. 330 IV. Restoration

BOX 10.2 early 1990s, a formal decision was made to link the red-cockaded woodpecker recovery Restoring the Savanna to plan with restoration of the native savanna the Savannah River Site communities (Gaines et al. 1995). Don Imm and John Blake Savannah River Site, USDA Forest Service, Savanna Restoration New Ellenton, South Carolina 29809 Research

Savanna restoration research was targeted Background at addressing important questions tied to natural resource management decisions. The Savannah River Site (SRS) is a 80,128- Does historical evidence and current vegeta- ha Department of Energy facility that lies tion indicate the presettlement fire savanna within the upper Coastal Plain of South composition and distribution? How has land Carolina and is adjacent to the Savannah use since European settlement impacted River. The area is characterized as a warm those communities? Can we identify the temperate climate with moderately well to native savanna species, communities, and excessively drained sandy to sandy loam habitat relationships? What management surface soils underlain by sandy loam to practices are effective for restoring and sus- clay loam subsoils at varying depths. When taining these communities? Parallel efforts the SRS was established in 1951, the land were made to expand field surveys, mon- was classified as 40% agriculture and 60% itor sensitive species, map remnant com- forestland. However, a large portion of the munities, and organize existing ecological forestland was cut over (∼20,000 ha) and databases. had an average stocking of only one-fifth Research studies using a combination of of similar lands in the vicinity. The oldest historical plats, diaries, and field surveys pine stands dated from about 1890 follow- established that fire savanna communities ing Civil War land abandonment. dominated about 78% of the upland land- In 1951 the USDA Forest Service was en- scape (Frost 1997). These fire-influenced gaged to reforest the agriculture lands and communities ranged from canebrake wet- cut over forest areas. This project was largely lands and open woodland, to scrub-oak complete by the 1970s. During this same stands, and they contained an array of grass period an ecological baseline was estab- and forb species characteristic of the upper lished through various research institutions Coastal Plain (Peet and Allard 1995; Duncan (University of Georgia, University of South and Peet 1996). Cultivation, subsistence Carolina, and the Philadelphia Academy of hunting, dam construction, and selective Sciences) to document the status of the tree cutting coupled with fragmentation aquatic and terrestrial flora and fauna. In of the original landscape and fire protec- 1974, the Savannah River Site was estab- tion had major ecological impacts on native lished as the first National Environmental flora and fauna (White and Gaines 2000). Research Park. In the late 1970s, coinci- The absence of periodic burning and ear- dent with implementation of the first recov- lier crop agriculture reduced the abundance ery plan for the endangered red-cockaded of native grasses and forbs, as well as di- woodpecker, the prescribed burning was in- agnostic species such as wiregrass (Aristida creased to improve habitat conditions for beyrichiana) to a few locations (W. T. Batson that species, and reduce fuel loading. In the personal communication), and in doing so 10. Restoring the Ground Layer 331

reduced the amount of available seed for fu- of founder populations along an environ- ture establishment beneath newly planted mental gradient (Foster and Imm unpub- forests. Today, a large percentage of the up- lished). For category “2” areas, the SRS land forest is dominated by old field grasses established a large experiment on restora- and forbs (Smith 2000; Imm and McLeod tion and habitat relationships in the wet- 2005). Due to the exclusion of periodic lands and upland margins of periodically burning through most of the twentieth cen- burned and nonburned Carolina bays. For tury, early and midsuccession hardwood category “3” areas, studies include an eval- species co-dominate with southern pines. uation of thinning and herbicide use on Ecological studies of intact communities existing understory communities (Harring- (Duncan and Peet 1996) and old fields ton and Edwards 1999); an assessment of (Smith 2000) demonstrated that composi- midstory removal (mechanical, chemical) tion and diversity of the grass and herba- on existing populations of savanna species ceous community follow gradients that are (Glitzenstein, unpublished); and a test of strongly influenced by soils, topography, the influence of establishment conditions drainage, and geographical location. While on seed and seedling survival for different providing clues to ecological controls, they species (Primack and Walker 2003). were not precise enough to predict “presettlement” composition for specific parcels. Routine monitoring and surveys for sensitive species, and studies of remnant savanna Landscape Restoration fragments (Frost unpublished) are further Strategies enhancing our knowledge of the original landscape distribution. Numerous studies of Given the land use impacts and uncertainty habitat relationships for vertebrate species about original species composition, as well have been conducted. For vertebrate com- as SRS national defense missions, the stew- munities (birds, herpetofauna, and small ardship goal is to restore and sustain the mammals), early succession stands were native savanna species characteristic of the found to support a wide array of species region. Operational experience at SRS has characteristic of savannas (Grant et al. 1994; demonstrated that prescribed burning alone Dunning et al. 1995; Yates et al. 1997; Kilgo is rarely sufficient to redevelop the native et al. 2000). species composition or structure. Periodic Latent savanna communities at SRS can burning of large areas in South Carolina is be grouped into three general categories: (1) constrained by smoke management regula- old field pine with limited savanna species, tions. When combined with fuel moisture (2) wet pine–hardwood forests along wet- and weather restrictions, it is extremely dif- land margins and swamps, and (3) lon- ficult to effectively apply periodic burning to gleaf pine or pine–hardwood fragments large acreage. In addition, the total absence (<1 acre to several acres) with significant or low numbers of native flora and fauna remnant populations of savanna species (e.g., gopher tortoise) severely limits nat- in the understory. For category “1” areas, ural recolonization, and woody vegetation vegetation research was conducted on or- competition further suppresses plant de- dination factors of old-field communities velopment. The operational strategies tak- relative to remnant stands (Smith 2000); ing shape at the landscape scale combine overstory conditions and soil tillage effects prescribed burning, mapping, monitoring, on seed and seedling establishment for wire- reintroductions of native species, and vari- grass (Outcalt et al. 1999); and survival, ous silvicultural technologies to achieve the development, dispersal, and establishment stewardship goals. 332 IV. Restoration

The current recovery plan for the red- dependent vegetation of the Fall-line Sand- cockaded woodpecker was realigned to fa- hills, southeastern United States. SRI 96-23R, cilitate prescribed burning to reduce compe- US Forest Service-Savannah River, New Ellen- tition and litter and duff thickness (Edwards ton, SC. et al. 2003). The primary recovery area Dunning, J.B., Jr., Borgella, R., Jr., Clements, K., is easier to burn, and the soils, topogra- and Meffe, G.K. 1995. Patch isolation, corridor effects, and colonization by a resident spar- phy, and vegetation relationships are con- row in a managed pine woodland. Conserv Biol sistent with areas historically dominated by 9:542–550. savanna communities (Duncan and Peet Edwards, J.W., Smathers, W.M., Jr., LeMaster, 1996; Frost 1997; Imm and McLeod 2004). E.T., and Jarvis, W.L. 2003. Savannah River Early succession or regeneration areas are Site red-cockaded woodpecker management being maintained by reducing planting den- plan. US Forest Service-Savannah River, New sities and by initiating burning within a Ellenton, SC. year or two of planting. These areas act Frost, C.C. 1997. Presettlement vegetation and as surrogates for savanna habitat because natural fire regimes of the Savannah River they provide analogous structure required Site. SRI 97-10-R, US Forest Service-Savannah (Johannsen 1998; Krementz and Christie River, New Ellenton, SC. Gaines, G.D., Franzreb, K.E., Allen, D.H., Laves, 1999). The current reintroduction of the go- K.S., and Jarvis, W.L. 1995. Red-cockaded pher tortoise at SRS is focused on these ar- woodpecker management on the Savannah eas. River Site: A management/research success Within the landscape, detailed mapping story. In Red-Cockaded Woodpecker: Recovery, Ecol- of sensitive species and remnant fragments ogy, and Management, eds. D.L. Kulhavy, R.G. is being used to define appropriate silvi- Hooper, and R. Costa, pp. 81–88. Center for cultural and burning techniques. For rem- Applied Studies, College of Forestry, Stephen nant forest fragments with residual savanna F. Austin State University, Nacogdoches, species, silvicultural efforts include (1) in- TX. creased burning frequencies, (2) reductions Grant, B.W., Brown, K.L., Ferguson, G.W., and in tree densities through thinning, (3) me- Gibbons, J.W. 1994. Changes in amphibian biodiversity associated with 25 years of pine chanical and chemical removal of midstory forest regeneration: Implications for biodiver- shrubs, saplings, and overstory hardwoods, sity management. In Biological Diversity: Prob- and (4) the conversion of off-site species lems and Challenges, eds. S.K. Majumbar, F.J. to longleaf pine. For old-field sites, dom- Brenner, J.E. Lovich, J.F. Schalles, and E.W. inated by weedy species a cost-effective Miller, pp. 354–367. Easton: The Pennsylvania strategy for plants may be establishment Academy of Science. of founder populations of selected species Harrington, T.B., and Edwards, W.B. 1999. Un- in heavily thinned, or clear-cut gaps that derstory vegetation resource availability, and are periodically burned. With the excep- litterfall responses to pine thinning and woody tion of species such as wiregrass, seed in- vegetation control in longleaf plantations. Can troductions have not been very successful J For Res 29:1055–1064. Imm, D.W., and McLeod, K.W. 2005. Vegetation (Primack and Walker 2003). Establishment types. In Ecology and Management of a Forested of native plants by planting small container Landscape: Fifty Years of Natural Resource Stew- stock has been promising. ardship on the Savannah River Site, eds. J.C. Kilgo and J.I. Blake, pp. 85–131. (In Review). References Johannsen, K.L. 1998. Effects of thinning and herbicide application on vertebrate commu- Duncan, R.P., and Peet, R.K. 1996. A template nities in young longleaf pine plantations. MS for the reconstruction of the natural fire- thesis, University of Georgia, Athens. 10. Restoring the Ground Layer 333

Kilgo, J.C., Miller, K.V., and Moore W.F. 2000. bution of rare plant species at the Savannah Coordinating short-term projects into an ef- River Site and the Carolina Sandhills Na- fective research program: Effects of site prepa- tional Wildlife Refuge. SRI 03-29-R, US For- ration methods on bird communities in pine est Service-Savannah River, New Ellenton, plantations. Stud Avian Biol 21:144–147. SC. Krementz, D.G., and Christie, J.S. 1999. Scrub- Smith, G.P. 2000. Structure and composition of successional bird community dynamics in vegetation on longleaf pine plantation sites young and mature longleaf pine–wiregrass sa- compared to natural stands occurring along an vannas. J Wildl Manage 63:803–814. environmental gradient at the Savannah River Outcalt, K.W., Williams, M.E., and Onokpise, O. Site. MS thesis, Clemson University, Clemson, 1999. Restoring Aristida stricta to Pinus palustris SC. ecosystems on the Atlantic Coastal Plain, USA. White, D.L., and Gaines, K.F. 2000. The Sa- Restor Ecol 7(3):262–270. vannah River Site: Site description, land-use Peet, R.K., and Allard, D.A. 1995. Longleaf pine and management history. Stud Avian Biol 21: vegetation of the South Atlantic and Eastern 8–17. Gulf Coast regions: A preliminary classiﬁca- Yates, M.D., Loeb, S.C., and Guynn, D.C., Jr. tion. Proc Tall Timbers Fire Ecol Conf 18:45–81. 1997. The effect of habitat patch size on small Primack, R.J.L., and Walker, J. 2003. Dispersal mammal populations. Proc Annu Conf Southeast and disturbance as factors limiting the distri- Assoc Fish Wildl Agenc 51:501–510. Chapter 11 Reintroduction of Fauna to Longleaf Pine Ecosystems Opportunities and Challenges

Ralph Costa and Roy S. DeLotelle

Introduction in our fourth section, we provide a model that may be useful for planning future reintroduc- In this chapter, we discuss reintroduction, via tions of fauna in longleaf pine forests. translocation, of native fauna into longleaf pine forests. We focus on rare species, including those considered “sensitive,” “of special concern,” or “candidates” for listing by con- Reintroduction and servation groups, or state or federal agencies, Translocation: A and on species federally listed as either “threatened” or “endangered” under the Endangered Conservation Strategy Species Act of 1973, as amended. We cover Review our topic in four main sections. First, we provide a brief background on the broad con- The International Union for Conservation of cept of reintroductions and translocations as a Nature and Natural Resources/Species Sur- species-conservation tool. Second, we use the vival Commission, Re-introduction Special- red-cockaded woodpecker (Picoides borealis)as ist Group (IUCN/SSC RSG) provides general a case study to outline how this regional pro- guidelines for reintroductions considered ap- gram has succeeded. This section provides a propriate internationally (see IUCN/SSC RSG literature review of all relevant issues regard- 1995). Their goal is to “. . . help ensure that ing translocation of red-cockaded woodpeck- the re-introductions [hereafter referred to as ers. Our goal in using the red-cockaded wood- reintroduction(s)] achieve their intended con- pecker as a case study is to provide an empirical servation benefit, and do not cause adverse framework so the reader can apply the basics side-effects of greater impact.” The guidelines’ of what we have learned about this species’ target audiences are managers and scientists reintroduction and translocation potential to who are directly responsible for planning, ap- other species. In our third section we provide proving, and carrying out reintroductions. The brief summaries about other species that may guidelines contain six important sections: (1) benefit from a reintroduction program. Finally, definition of terms, (2) aims and objectives

r Ralph Costa U.S. Fish and Wildlife Service, Clemson Field Ofﬁce, Department of Forestry and Natural Resources, r Clemson University, Clemson, South Carolina 29634 Roy S. DeLotelle DeLotelle and Guthrie, Inc., 1220 SW 96th Street, Gainesville, Florida 32607.

335 336 IV. Restoration of reintroduction, (3) multidisciplinary ap- et al. (1989) concluded that “. . . without high proach, (4) preproject activities (biological, quality habitat . . . ” translocations would have and socioeconomic and legal requirements), poor success regardless of how many animals (5) planning, preparation, and release stages, are released. and (6) postrelease activities. Wolf et al. (1996) conducted a follow-up sur- The IUCN/SSC RSG defines “reintroduction” vey to Griffith et al. (1989) and examined 421 as “an attempt to establish a species in an area bird and mammal translocation programs in which was once a part of its historical range, Australia, New Zealand, and North America. but from which it has been extirpated or be- They documented increases from the results come extinct.” They define “translocation” as of Griffith et al. (1989) in three potentially im- a “deliberate and mediated movement of wild portant parameters related to successful pro- individuals or populations from one part of grams. The median number of animals translo- their range to another.” In this chapter, “aug- cated per program (31.5 to 50.5), the median mentation” refers to the translocation of red- duration of releases (2 to 3 years), and the cockaded woodpeckers from a donor popula- proportion of projects releasing greater than tion to another population to increase its size; 30 individuals (46% to 68%), all increased. “reintroduction” refers to the translocation of Like Griffith et al. (1989), Wolf et al. (1996) individuals from a donor population to un- also found that high-quality, i.e., “good to ex- occupied habitat that was probably occupied cellent,” habitat, releases into the core of the historically. We will use “translocation” as a species’ historical range, and large numbers generic term for most discussions related to of translocated individuals positively affected augmentations and/or reintroductions. In this translocation success. In contrast to Griffith et chapter, we will primarily discuss augmenta- al. (1989), Wolf et al. (1996) found no signif- tions, because only one reintroduction of red- icant relationship between translocation suc- cockaded woodpeckers has been accomplished cess and either first age of reproduction (“early as of 2003 (Hagan and Costa 2001). In the case or late breeder” in Griffith et al. 1989) or num- of red-cockaded woodpeckers, for all practical ber of offspring (“size of clutches” in Griffith purposes there is little difference from either et al. 1989). Both studies found translocation a planning or implementation perspective be- programs for mammals were more successful tween augmentations and reintroductions. than those for birds. Griffith et al. (1989) provided the first com- Gordon (1994) developed a dichotomous prehensive analysis of the status of translo- key to provide natural resource managers a cation as a conservation tool. They summa- tool to help assess the biological and genetic rized information on 93 species of birds and needs and impacts of translocating species. The mammals (90% game species, 86% success; key is useful for both plants and animals and 7% threatened, endangered, or sensitive, 46% considers the following factors: (1) degree of success) translocated from 1973 to 1986 in threat to species being considered, (2) dispersal Australia, Canada, Hawaii, New Zealand, and from release site, (3) genetic risks, (4) cause of the United States. Almost 700 translocations threat, (5) donor source, (6) competitive inter- were conducted annually, using both captive actions, (7) consumptive interactions, (8) con- and wild stock. They found success improved tamination risks, and (9) release site manage- with increased habitat quality, larger num- ment. Gordon (1994) suggests that prior to any bers of released animals, translocations into translocations, decisions should be well docu- the core of a species’ historical distribution ver- mented (the key provides the means to do so) sus on the periphery or outside the range, and and sufficient postrelease monitoring planned translocations into areas without competitors and implemented. of similar life form or a congeneric potential Scott and Carpenter (1987) argued that competitor. Due to limited survey question re- a lack of reliable published information ex- sponses they were unable to evaluate effects isted to adequately evaluate success of endan- of age, sex, and genetics on success. Griffith gered bird translocation programs. Without 11. Reintroduction of Fauna 337 such accounts, they concluded that it was dif- maintaining forest structure in fire-climax flat- ficult to accurately assess the species’ status or woods and pine savannas important to many evaluate success of specific procedures being protected wildlife species (Stout and Marion employed. Furthermore, they concluded that 1993). These pine forests have been exten- because of the high cost of endangered bird sively altered by cattle grazing, ditching, fire translocation programs, we must ensure high suppression, conversion to agricultural lands, survival probabilities of individual birds and logging, and naval stores extraction. Today, few meaningful contributions to the gene pool. To pine forests exist that are representative of pre- address these issues and concerns they pro- settlement conditions. Remaining pine forests posed the following guidelines (this partial differ from those of the presettlement era in list only includes those factors related to wild having longer fire return intervals, a more stock): (1) document bird preparation and re- even-aged structure, and a denser understory lease methodologies, e.g., capture techniques, with greater shrub cover and less herb cover. holding cages, and transportation methods, (2) These alterations have led to habitat fragmen- record release conditions (see Ellis et al. 1978), tation, reduced faunal diversity, and extirpa- and (3) frequently monitor movements and tion of many vertebrate species in most lon- activities of released birds through the first gleaf pine landscapes. breeding season. Scott and Carpenter (1987) The community’s dominance in the land- summarized by suggesting that improved doc- scape was facilitated by a suite of pyrogenic umentation of translocation programs applies traits that favored longleaf pine’s survival over not only to endangered birds but also to hardwoods or other species of pine (Stout and nonendangered birds and numerous other tax- Marion 1993). The physiognomy of classic pine onomic groups. forest is characterized by an emergent tree Lessons learned from this review of the layer of pines and a ground cover that, al- translocation literature include: (1) preplan- though appearing to be species-poor, is one ning with documentation is critical, (2) high- of the most species-rich on the planet (Walker quality habitat at the recipient site is manda- and Peet 1983; Platt et al. 1988b; Peet this vol- tory, (3) programs should be restricted to ume). Fire is important in maintaining diver- historic range, (4) large numbers of individ- sity of ground cover and the structural aspects uals should be moved, (5) a review of liter- necessary for many wildlife species (Engstrom ature relevant to one’s species or taxa must et al. 1984; Robbins and Myers 1992). In addi- be conducted, and (6) monitoring postrelease tion to fire, seasonal variation in water avail- success so that programs can be improved is ability, low and flat topography, and depression crucial. Readers are encouraged to review the areas also influence structure of pine forests summary literature discussed above and other and community types. current, relevant work related to particular Several types of depressional wetlands (e.g., species or taxa before engaging in transloca- cypress domes, woodland ponds) are found tion programs (e.g., see Jones, S. R., ed. 1990. within the matrix of the heterogeneous land- Endangered Species Update, Volume 8, No. 1). scape, as are isolated fragments of scrub, and larger areas of scrubby flatwoods or oak ridges. Size of these embedded communities varies Reintroduction of Fauna in greatly; however, collectively they can provide considerable coverage and diversity. Veg- Longleaf Pine Ecosystems etation of these embedded habitats within Longleaf pine (Pinus palustris) occurred in the pine forest matrix varies with edaphic monospecific stands over a larger area than conditions, hydroperiod, and basin topogra- any other tree species in the presettlement phy (Abrahamson and Hartnett 1990). A rel- landscape of the United States (Platt et atively rich fauna is associated with these al. 1988a; Abrahamson and Hartnett 1990; forests (Harris and Vickers 1984; Means this Frost this volume). Longleaf pine is key to volume). All amphibians and many reptiles 338 IV. Restoration

TABLE 1. List of potential fauna for reintroduction into longleaf pine ecosystems. Common name Scientific name Statusa Reintroduction potentialb Amphibians Mississippi gopher frog Rana capito sevosa E No Recovery plan (more info required) Flatwoods salamander Ambystoma cingulatum T No Recovery plan (more info required) Reptiles Eastern indigo snake Drymarchon corais couperi T Recovery plan (more info required) Gopher tortoise Gopherus polyphemus T Recovery plan (yes; disease problems) Blue-tailed mole skink Eumeces egregius lividus T Recovery plan (no; more info required) Sand skink Neoseps reynoldsi T Recovery plan (no; more info required) Louisiana pine snake Pituophis ruthveni C More info required Birds Mississippi sandhill crane Grus canadensis pulla E Recovery plan (yes; captive bred) Red-cockaded woodpecker Picoides borealis E Recovery plan (yes; wild stock, artificial cavity installation) Bald eagle Haliaeetus leucocephalus T Recovery plan (yes; wild stock, nest site construction) Southeastern american kestrel Falco sparverius paulus N3 Yes; nest box installation Bachman’s sparrow Aimophila aestivalis N3 More info required Mammals Red wolf Canus rufus E Recovery plan (yes; captive bred) Florida panther Felis concolor coryi E Recovery plan (yes; captive bred) Florida black bear Ursus americanus floridanus N2 More info required Lousiana black bear Ursus americanus luteolus T Recovery plan (no; more info required) Sherman’s fox squirrel Sciurus niger shermani N2 More info required a Status is based on Endangered Species Act listing (E = endangered, T = threatened, C = candidate) or if not federally listed, The Nature Conservancy and the Natural Heritage Network rankings (N = national status; 2 = imperiled, 3 = vulnerable to extirpation or extinction). b Potential is based on information in an approved recovery plan (“yes” or “no” indicates whether reintroduction/augmentation is discussed as a recovery activity) or the authors’ evaluation that more information (info) is required for unlisted species and for listed species with and without recovery plans, even if species with plans do not specifically list reintroductions/augmentations as a recovery activity. of the longleaf pine ecosystem are obligatorily additional listed, rare, or sensitive species linked with water (e.g., breeding and feeding may be potential candidates for similar ini- sites) or upland ridges and, therefore, main- tiatives (Table 1). Significantly, these poten- taining a healthy link between uplands and tial candidates cover all terrestrial vertebrate wetlands with varying hydroperiods will result taxa, including reptiles, amphibians, birds, and in higher abundance and diversity of species mammals. The list of potential candidates is within these taxonomic groups (Vickers et al. substantial, covers many life forms over a 1985). broad geographic area, includes multiple phys- As a conservation strategy, reintroducing ex- iographic provinces, ecoregions, and longleaf tirpated and augmenting existing small popu- pine habitats, and represents the analyses and lations of fauna in longleaf pine forests is in opinions of numerous federal and state agen- its infancy. Red-cockaded woodpeckers are the cies and nongovernmental organizations. only resident species with a significant his- Reintroducing or augmenting native fauna tory of translocations, while several species in longleaf pine ecosystems should be consid- have been supported with limited reintro- ered for the following reasons: (1) recovery ductions. These include red wolf (Canus ru- of federally or state listed threatened or en- fus), Mississippi sandhill crane (Grus canaden- dangered species, (2) conservation of federal sis pulla), and southeastern American kestrel and state candidate and sensitive species, and (Falco sparverius paulus) programs. Numerous (3) conservation of rare or keystone species. 11. Reintroduction of Fauna 339

In a broad sense, these specific reasons address ited the longleaf pine forests from Virginia larger conservation goals, including enhancing south to Florida, and west to Texas (Costa long-term survival of a species, maintaining or 2001). In 2003, 5800 occupied clusters of cav- restoring biodiversity, and increasing conser- ity trees were known throughout the species vation awareness (IUCN/SSC RSG 1995). Ad- range, an estimated 99% decline (Costa and ditionally, specific reasons address legal man- Jordan 2003). Initial dramatic population de- dates, e.g., achieving recovery, political, and clines were a result of large-scale, extensive administrative issues, e.g., precluding the need habitat loss, primarily from logging, fire sup- to list a species, cultural and social issues, e.g., pression, and the naval stores industry (see informing and educating the public about the U.S. Fish and Wildlife Service 2003 for details value of saving rare species, and finally, ecolog- and additional references). ical issues, e.g., reintroducing keystone species, Once historic populations were fragmented which in turn benefits biodiversity by posi- into smaller, disjunct populations, other local tively affecting a host of other species. factors adversely impacted persistence of many local populations. Factors including population isolation, fragmentation, and size, hard- Translocation Case Study: wood midstory encroachment, lack of suitable cavity trees, and demographic and environ- Red-cockaded Woodpecker mental stochasticity caused further declines (Costa and Escano 1989; Conner and Rudolph This section covers seven major topics: (1) 1991; Crowder et al. 1998; Letcher et al. 1998; need for translocations, (2) threats to small Walters et al. 2002). In combination, these lo- populations, (3) why translocations work, (4) cal factors contributed to the sometimes rapid translocation success, (5) translocation con- and steady declines, and in numerous cases ex- cerns, (6) translocation strategies, and (7) the tirpations, of remnant populations. future. The red-cockaded woodpecker is found Walters (1991) provided a further interpre- in several primary pine forest types within tation of red-cockaded woodpecker population its range, including longleaf pine, shortleaf dynamics and its implications for dealing with pine (Pinus echinata), slash pine (Pinus elliottii), population decline, i.e., preventing extirpa- and loblolly pine (Pinus taeda). Although forest tions. His findings support the principle that types within a particular property can be essen- combating limiting habitat factors, particu- tially “pure” on a stand-by-stand basis, mixed larly cavity availability, is key to saving and stands, with one pine species comprising the restoring small populations of red-cockaded majority of the overstory, are more typical woodpeckers. In addition, well-planned and throughout the species range. Translocations -executed translocations, adequate quantities of red-cockaded woodpeckers have taken place of good-quality nesting and foraging habitat, in all forest types. To date, no studies have in- and an aggressive prescribed burning program dicated that the forest type affects the outcome will all be necessary to save very small popu- (see Edwards and Costa 2004). Therefore, we lations from extirpation (e.g., see Brown and include literature and data from studies in all Simpkins 2004; Drumm et al. 2004; Hedman forest types and assume that our discussion et al. 2004; Stober and Jack 2004). of red-cockaded woodpecker translocations is relevant to all appropriate southern pine forest types. Achieve Recovery Recovery of the federally listed red-cockaded Need for Translocations woodpecker (listed as “endangered” in 1970, Prevent Extirpations 35 Federal Register 16047) depends on establishing sufficient numbers of varying size Historically, as many as 920,000 groups of populations, well-distributed throughout their red-cockaded woodpeckers may have inhab- historic landscape, to enable the species to 340 IV. Restoration counteract threats inherent to survival of small species, translocation of other fauna into lon- populations. Specifically, recovery, or “delist- gleaf pine ecosystems could have various ing,” of the red-cockaded woodpecker involves objectives. These include, to: enhance the establishing at least 29 populations in 11 recov- long-term survival of a species, reestablish a ery units (i.e., ecoregions) with the following keystone species, maintain and restore natural sizes (size = potential breeding groups): 1 with biodiversity, promote conservation awareness, 1000, 10 with 350, 9 with 250, 3 with 100, or a combination of these (IUCN/SSC RSG and 6 with 40 (see U.S. Fish and Wildlife Ser- 1995). Prior to any fauna translocation pro- vice 2003 for details). The 29 “populations” are gram in longleaf pine forests, a feasibility study, comprised of 65 individual properties on pri- including a species status review and thorough vate, state, and federal lands. Individual prop- review of available literature on the species, erties may or may not be contiguous to other must be completed. In addition to identifying properties comprising a population. the need for the program, this background re- Any population or noncontiguous property search will help identify and, thereby, focus harboring fewer than 30 potential breeding the stated objectives of the translocation. Once groups is eligible to receive translocated birds. identified, objectives must be clearly stated, Of the 29 recovery populations, 15 (52%) con- understood by all involved parties, and capable tain fewer than 30 potential breeding groups. of being evaluated on a regular basis to ensure Of the 65 properties, 38 (58%) harbor fewer they are being accomplished. than 30 potential breeding groups; 32 (49%) are noncontiguous. In addition to these pop- Threats to Small Populations ulations directly involved in delisting, man- Threats to small populations include de- agement and conservation of numerous other mographic and environmental stochasticity, populations, i.e., “significant” support popula- genetic uncertainty (both inbreeding and ge- tions (see U.S. Fish and Wildlife Service 2003), netic drift), and natural catastrophic events is critical to help achieve recovery. Many of (Shaffer 1981, 1987). Based on recent re- these support populations also qualify for, and search, population sizes and spatial configu- are participating in, translocation programs. rations of red-cockaded woodpecker groups Factors necessitating a translocation pro- within populations necessary to withstand gram include the fragmented nature of the potential extirpation risks associated with de- Southeast’s longleaf pine forests, wide and iso- mographic and environmental stochasticity lated distribution of populations, and inad- are better understood (Crowder et al. 1998; equate dispersal to increase and sustain de- Letcher et al. 1998; Walters et al. 2002). mographic health of many small populations. Small, isolated populations are subject to mul- Demographic Constraints tiple threats and the only way to combat these threats is by increasing population size. Research has demonstrated that the probabil- Translocation is a powerful management tool ity of maintaining red-cockaded woodpecker capable of relatively rapidly increasing pop- populations is dramatically improved as population size to a more self-sustaining level. ulation size increases and territories are maxi- Once populations harbor 30 or more potential mally aggregated (e.g., 1 group per 200 acres) breeding groups, continued population growth (Crowder et al. 1998; Letcher et al. 1998). can be achieved without reliance on inter- Since most territories are large, higher popula- population translocations (see U.S. Fish and tion densities can be difficult to attain (Hooper Wildlife Service 2003). However, occasional et al. 1982; DeLotelle et al. 1987). Addition- interpopulation translocations may be neces- ally, Crowder et al. (1998) suggested that prox- sary to maintain or improve genetic viability imity of adjacent populations might increase (Haig et al. 1993). the persistence of small populations (also see Besides preventing extirpation of local pop- DeLotelle et al. 2004). Importantly, results of ulations and achieving recovery for listed both Crowder et al. (1998) and Letcher et al. 11. Reintroduction of Fauna 341

(1998) support the concept of maintaining ex- tion dynamics.” That is, when environmental isting and establishing new populations, even stochasticity results in annual variation in mor- if available habitat limits the ultimate popula- tality or reproduction, the effect is essentially tion size to 10–25 groups. It is vital that these absorbed by the helper class and, therefore, the small populations be dense and maximally ag- overall number of occupied territories remains gregated within suitable habitat (Walters et al. unaffected. However, ability of the helper class 2002). to function as an effective buffer against threats Neither the Crowder et al. (1998) nor the of environmental stochasticity is a function of Letcher et al. (1998) models allowed for im- not only the population’s size, but also criti- migration. Letcher et al. (1998) acknowledged cally, its density and aggregation (Walters et that immigration might help support popula- al. 2002). Significantly, our ability to strate- tion growth and stability in small populations. gically, via translocations, locate new groups However, they noted that the effect of the where needed to maximize density and con- population’s spatial distribution on immigra- nectivity within populations is and will remain tion remains untested. In the absence of immi- a key management objective as we build small gration, via dispersal, translocation provides a populations to and beyond the threshold sizes management tool to help maintain small pop- necessary to withstand natural demographic ulation viability. (100 potential breeding groups) and environmental (250 potential breeding groups) Environmental Stochasticity threats. Environmental stochasticity can affect red- Genetic Considerations cockaded woodpecker populations in several ways. For example, an exceptionally severe Both the short- and long-term survival of winter could result in high adult mortality, or small- to moderate-size red-cockaded wood- a prolonged drought during the nesting sea- pecker populations will require genetic man- son could result in poor nestling survival and, agement. Haig and Nordstrom (1991) provided therefore, lower recruitment. These types of an excellent primer and guide on genetic man- annual population-level effects have the po- agement of small populations. Ultimately, in tential, at least over the short term, to affect small populations, loss of genetic diversity re- population parameters such as group size and sults in a species’ inability to adapt to changing recruitment of helpers. In small populations conditions (Selender 1983). Haig and Nord- these effects can be even more pronounced strom (1991) pointed out that “. . . loss of ge- leading to population declines (DeLotelle et al. netic diversity is manifested by more severe ef- 1995). fects resulting from . . . ” random genetic drift Walters et al. (2002) conducted a population and an increase in levels of inbreeding. While viability analysis for red-cockaded woodpeck- genetic drift can threaten even large popu- ers by incorporating environmental stochas- lations, inbreeding depression threatens only ticity parameters, e.g., probabilities of produc- small populations. The effect of genetic drift, ing different numbers of fledglings each year, i.e., rate of loss of genetic variation, increases into the demographic model developed by with shrinking population size and mutation Letcher et al. (1998). Their study results pro- rate. Inbreeding depression is one effect of ge- duced two important conclusions. First, den- netic drift (Lacy 1987). sity and distribution of groups have as signif- Inbreeding affects fecundity, fertility, de- icant an effect on maintaining viable popula- velopment rate, age of sexual maturity, and tions through time (that is population fitness) number of offspring, i.e., factors that con- as population size itself. Second, small, highly trol survival and productivity of individu- aggregated populations were relatively stable. als (Ralls et al. 1988). Inbreeding depres- Walters et al. (2002) attribute their findings to sion has been demonstrated in red-cockaded the “. . . buffering effect of helpers on popula- woodpecker populations (Daniels and Walters 342 IV. Restoration

2000a; Daniels et al. 2000). Daniels et al. Haig et al. (1996), sampling small popula- (2000) found that populations of at least 40, tions in south Florida, found that birds were but perhaps as many as 100, potential breed- no more differentiated or isolated than other ing groups are necessary to avoid inbreed- populations throughout their range and that ing depression and concluded that viability of there were no genetic differences among birds small isolated and declining populations is se- in longleaf and slash pine habitats. They con- riously threatened by inbreeding depression. cluded, “. . . habitat type does not constrain However, they also calculated that an immi- dispersal . . . among geographically proximate gration rate of two or more migrants per year populations in south Florida,” but recom- would likely protect against inbreeding depres- mended translocating red-cockaded wood- sion. Where immigration rates are insufficient peckers from nearby populations and using (less than two per year), Daniels et al. (2000) donors with similar habitat types. However, recommended translocation as a possible tool Edwards and Costa (2004) suggested that, to reduce inbreeding. when nearby donor populations of the same Stangel et al. (1992), examining genetic habitat type are limited to other small popu- variation and population structure, found that lations, such as in south Florida, translocating most small populations exhibited normal lev- birds from larger, nearby populations of differ- els of heterozygosity. Based on their find- ent habitat types is acceptable and not likely to ings, they suggested that small populations: be deleterious. (1) should not be considered “lost causes,” (2) Generally, genetic diversity increases while are valuable reservoirs of unique genetic com- threats associated with genetic drift decrease binations, and (3) serve as “steppingstones” with population size. Therefore, population for gene flow among larger populations. They growth strategies must be integral components recommended management and conservation of small population conservation and man- of both small and large populations to main- agement programs. Translocations provide one tain the species’ historic genetic population important tool to grow larger populations and, structure. Because they found evidence that thereby, increase genetic diversity. Similarly, “. . . genetic distance increases with geographic deleterious effects of genetic drift, including in- distance . . . ” they recommended that donor breeding depression, can be minimized by nat- populations be as near as possible to recipi- ural immigration into small populations and ent populations to minimize “disrupting locally “managed” by translocation of individuals into adapted populations.” populations. All genetic research and analyses Haig et al. (1994) found no population- to date indicate that small populations of red- specific, diagnostic genetic markers among the cockaded woodpeckers are worth saving and 101 red-cockaded woodpeckers sampled from likely can be genetically managed via translo- 14 populations rangewide. Significantly, both cations. Similarly, we believe the “small popu- Stangel et al. (1992) and Haig et al. (1994) lations are valuable” paradigm may hold true concluded that population-specific alleles were for many longleaf pine ecosystem fauna, al- not useful for differentiating populations. This though this theory remains untested for most suggests that although the current distribu- species. tion of red-cockaded woodpeckers consists of many, relatively small, fragmented popula- Catastrophic Events tions, these populations have not been isolated long enough to result in the loss or gain Two types of natural catastrophes potentially of specific genes within a population (Haig et threaten red-cockaded woodpecker popula- al. 1994). Based on their overall findings, and tions: (1) severe winds (hurricanes, down- in agreement with Stangel et al. (1992), Haig bursts, and tornados) and (2) southern pine et al. (1994) concluded that nearby popula- beetle epidemics. While impacts from torna- tions, as opposed to distant ones, should serve dos and downbursts will typically be local- as donors. ized, hurricanes have the potential to affect 11. Reintroduction of Fauna 343 entire populations. Effects of Hurricane Hugo stochasticity are overcome. If species-specific on the red-cockaded woodpecker population data on population sizes and associated re- and its habitat on the Francis Marion Na- silience to threats are not available, informa- tional Forest, South Carolina, were devastat- tion for similar taxa should be researched. It ing (Hooper et al. 1990). Because most pop- is imperative that research identifies and un- ulations are located in the Atlantic and Gulf derstands the potential for population loss and Coastal Plains they face a substantial risk from reduction from natural catastrophes. Addition- hurricanes (Hooper and McAdie 1995). Reduc- ally, it is critical to design translocation pro- ing the threat of hurricanes at the species level grams to withstand the possible effects of catas- can be accomplished by distributing numerous trophic events to the maximum extent practi- populations throughout the species range, in- cable. Proceeding with translocation programs cluding insular populations. with population goals (sizes, locations, and Southern pine beetles are not a population- configurations) that are set without knowl- level threat to red-cockaded woodpeckers in edge and consideration of threats to small pop- longleaf pine forests. Although longleaf pine ulations is inefficient and will jeopardize the has high natural resistance to southern pine program’s success. beetles, loblolly and shortleaf pine are not nearly as resistant, especially under stress (for example, growing on poor soils, at high stock- Why Translocations Work ing, severe drought conditions). The rapid Red-cockaded Woodpecker Demographics and dramatic loss of over 40,000 acres of and Sociobiology shortleaf, pitch (Pinus rigida), and Virginia pine (Pinus virginiana) in the Daniel Boone The red-cockaded woodpecker is a nonmigra- National Forest, Kentucky, and the subse- tory, territorial, cooperative breeder with all quent human-induced “extirpation” of the group members participating in nesting season red-cockaded woodpecker population illus- activities and territorial defense (Lennartz et al. trates the catastrophic potential of southern 1987). Groups can consist of a solitary bird, pine beetles (Mills et al. 2004). Proper forest usually a male, or a potential breeding pair management, including maintaining properly with zero to four helpers. Helpers are usually stocked stands and restoring longleaf pine to male offspring from previous breeding seasons; appropriate sites, can improve a forest’s resis- however, female helpers are common in some tance to catastrophic beetle epidemics. Fortu- populations (DeLotelle and Epting 1992). Wal- nately, at a smaller scale, the loss of individual ters et al. (1988a) studied the demographics cavity trees can be mitigated with artificial cav- and sociobiology of a large population in the ity replacement (Copeyon 1990; Allen 1991). North Carolina sandhills. Their model is appropriate to North Carolina and perhaps other populations, but some populations in different Summary of Threats to Small parts of the species’ range exhibit different lev- Populations of Longleaf Pine els of these demographic components. The fol- Ecosystem Fauna lowing percentages are from the Walters et al. (1988a) North Carolina study; percentages are Managers and biologists considering reintro- the annual change from one breeding season ductions or augmentations of fauna in lon- to the next. gleaf pine ecosystems must thoroughly un- Male fledglings may remain (27%) on their derstand the threats to small populations of natal territory as helpers for several years wait- the target species. This is critical because small ing for the opportunity to inherit it, or oc- populations will remain at risk for a rela- cupy an adjacent territory, and subsequently tively long period of time. Therefore, it is vi- acquire breeding status upon the death of the tal to know at what population sizes the risks breeding male. Approximately 13% of male of demographic, environmental, and genetic fledglings disperse and 39% of those become 344 IV. Restoration breeders on another territory. Another 31% Not only has our knowledge about numbers of the dispersing males find a vacant territory of dispersing birds and distances moved been or establish a new one and remain there as a considerably expanded in recent years, but also solitary bird. These birds have trouble attract- our understanding of how dispersal poten- ing mates and those that do are usually un- tially affects demographic and genetic health successful as breeders the first year (Daniels of populations substantially increased. Lay and Walters 2000a; Leonard et al. 2004). Dis- and Swepston (1973), studying red-cockaded persing males that do not establish or oc- woodpeckers in east Texas, recorded the first cupy a territory, but continue searching for long-distance movement of the species, 42 km one, are called “floaters”; they represent about by an adult male. It would be 15 years before 25% of the annual fledgling pool. Fledgling Walters et al. (1988b) reported the next long- male annual mortality is high, about 57%. distance dispersal of a bird, 86 km by a breeding Female fledgling mortality is also high, esti- female in the North Carolina Sandhills. Subse- mated at 68%. Of the female fledglings that quent long-distance dispersals ranging in dis- disperse from their natal territory and survive tance from 27 to 275 km have been recorded (about 31%), 92% become breeders their first in South Carolina (Jackson 1990), from Ok- year. Once established as breeders, females re- lahoma to Arkansas (Montague and Buken- main in that status (56%), disappear (31%), hofer 1994), from North Carolina to South or move to another territory (12%) the fol- Carolina (Ferral et al. 1997), in Texas (Con- lowing year. Young males and females typ- ner et al. 1997), and in Florida (Lowery and ically disperse more frequently in the early Perkins 2002). fall (October) and the late spring (March) Schiegg et al. (2002) examined effects of than other times of year (J. Walters, Vir- fragmented populations on dispersal. Their ginia Polytechnic Institute, personal commu- results suggested that habitat fragmentation nication, and R. DeLotelle unpublished data). could affect dispersal success, thereby ad- Dispersal patterns of both female and male versely affecting population dynamics in small subadults from their natal territories and their (25 groups) populations. They suggested that ability to breed their first year are behaviors the cooperative breeding system of the red- that apparently make the species conducive to cockaded woodpecker “. . . may stabilize small translocation. aggregated populations, which may even act as populations sources.” Schiegg et al. (2002) Red-cockaded Woodpecker Dispersals and concluded that small populations likely have Metapopulation Theory conservation value. Similarly, Haig et al. (1993:296) suggested that as her small study With the exception of a few large data sets population (Savannah River Site, SC) infor the North Carolina Sandhills red-cockaded creased in size, it “. . . could serve as a source woodpecker population (see Walters et al. of immigrants to numerous other local popu- 1988a; Daniels and Walters 2000b), rangewide lations.” Indeed, at least five birds have dis- knowledge regarding the frequency of both persed from this small population to distant short- and long-distance dispersals of this populations: two to Fort Gordon, GA, and species was limited (or unpublished) prior to three to Fort Jackson, SC; four of five became the 1990s. With the increased emphasis on breeders postdispersal. Daniels et al. (2000), saving small populations and the associated investigating inbreeding, also concluded that need to band large segments of donor popu- small populations are important and suggested lations and all birds in recipient populations enhancing dispersal via retention and man- for the translocation program, the number of agement of as many small populations as birds annually banded since the early 1990s possible within a region. Additionally, they has steadily increased. This existing and grow- recommended “. . . linking disjunct inhabited ing inventory of banded birds has proven in- areas . . . ” (Haig et al. 1993:147) using recruit- valuable to conservation efforts for the species. ment clusters. 11. Reintroduction of Fauna 345

TABLE 2. Long-distance dispersal patterns for 110 marked red-cockaded woodpeckers. Age Breedera Population sizeb Mean distance Small & Total Subadults Adults Yes No dispersed medium Large Female 64 (58.2%) 38 (52.7%) 21 (65.6%) 41 (60.3%) 9 (52.9%) 31.4 (n = 59) 28 (66.7%) 31 (53.5%) Male 46 (41.8%) 34 (47.3%) 11 (34.4%) 27 (39.7%) 8 (47.1%) 27.8 (n = 44) 14 (33.3%) 28 (46.5%) Total 110 72 32 68 17 29.6 (n = 103) 42 59 a Indicates breeding status for 85 of the 110 birds where status was documented. b Size of population bird dispersed to: small (40 potential breeding groups; PBG), medium (41–249 PBG), large (>249 PBG).

DeLotelle et al. (2004) summarized 110 dis- R. Woodroffe, and J. Gittleman, 2000; chap- persals (see Table 2) ranging in distance from ter entitled “The Metapopulation Paradigm: A 7 to 325 km; 60% exceeded 17 km. More Fragmented View of Conservation Biology”) dispersers were females (58.2%) than males states that “[t]he ever-growing diversity of em- and subadults (69.2%) than adults. Adult fe- pirical and theoretical studies that demonstrate males dispersed more frequently (65.6%) than the importance of spatial structure in deter- adult males. The majority (80%) of birds dis- ring ecological and evolutionary trajectories persing became breeders. Overall, females and also indicates that long-term conservation pro- males dispersed about the same average dis- grams need to focus on regional rather than tance (about 30 km); however, subadults dis- local within-population dynamics.” This is the persed farther (30 km) than adults (22 km). approach we advocate. Finally, 41.6% of dispersals were within or to In the context of metapopulation theory, all small and medium populations from a larger of the above findings and recommendations population, e.g., large to medium or medium represent important conservation issues. With to small. DeLotelle et al. (2004), based on few exceptions, all researchers and managers these data, years of research in central Florida, who have investigated red-cockaded wood- and a broader (see Harrison 1994; Gutierrez pecker dispersal as it relates to genetic, and and Harrison 1996) interpretation of “classic” in some cases demographic, maintenance of metapopulation theory (see Levins 1969), sug- small populations concluded that saving such gested that red-cockaded woodpeckers may populations has conservation value. Transloca- currently function as metapopulations in cer- tions provide the opportunity, both short and tain parts of their range. long term, to save, expand, and maintain iso- The metapopulation definition that we lated and insular small populations. Addition- choose is less stringent than the conserva- ally, translocations, along with improvements tive definition of “demographic rescue” (Han- in habitat linkages via “steppingstones” (Stan- ski and Simberloff 1997; DeLotelle et al. gel et al. 1992) or “island corridors” (Costa 2004). We believe that present-day aggrega- and Edwards 1997), can be potentially used tions of small red-cockaded woodpecker pop- to restore or maintain structure and function ulations in relatively close proximity to one an- of metapopulations. other more appropriately fit the “demographic The relatively high and improving success recovery” definition. In this sense, demo- rate of red-cockaded woodpecker transloca- graphic recovery involves movement among tions is not surprising when the bird’s so- red-cockaded woodpecker populations suffi- ciobiology, likely historic and documented cient to assist one another in demographic sta- present dispersal patterns, genetic and demo- bility (see DeLotelle et al. 2004). Further, a re- graphic metapopulation structure, and pop- cent book (Genetics, Demography, and Viability of ulation demographics are understood (Wal- Fragmented Populations, edited by A.G. Young, ters et al. 1988a; Haig et al. 1993, 1996; G.M. Clarke, M.L. Gosling, G. Cowlishaw, Letcher et al. 1998; DeLotelle et al. 2004). 346 IV. Restoration

That is, translocations replicate probable his- is crucial to understanding successes and toric short- and long-range dispersal patterns failures of any translocated animals. Without that would have occurred in large contigu- such information adaptive management will ous areas of forest, and in naturally frag- not be possible. mented landscapes where birds likely func- In this section, we primarily summarize tioned as a metapopulation, e.g., central and overall success of translocations at the popula- south Florida. That red-cockaded woodpeck- tion scale. That is, we measure success by eval- ers disperse within and among today’s highly uating changes of basic population parameters, fragmented populations and landscapes is well including number of occupied clusters, birds, documented (Walters et al. 1988a; Daniels and and potential breeding groups. All of our study Walters 2000a,b; DeLotelle et al. 2004). Dis- populations harbored fewer than 23 potential persal is a primary behavior of the species, di- breeding groups when they began a transloca- rectly related to an individual’s fitness, i.e., the tion program. Therefore, by definition, all were ability of most females and some males to ob- considered “small,” i.e., fewer than 30 poten- tain breeding status. Translocations simulate tial breeding groups. Indeed, 27 (90%) of the subadults’ natural behavior to disperse. Given 30 populations harbored fewer than 14 poten- the very high, annual natural mortality rates tial breeding groups; 17 (57%) had fewer than (57% male, 68% female) of subadults (Wal- 5. Populations exceeding 30 potential breeding ters et al. 1988a), and high retention rates (per- groups are not eligible to be translocation recent of birds remaining in their recipient pop- cipients (U.S. Fish and Wildlife Service 2003). ulation through at least one breeding season) Such populations are expected to expand on of translocated subadults (65%, Hedman et al. their own via use of recruitment clusters, pre- 2004; 72%, Hagan et al. 2004a; 82%, Stober scribed burning, and, as needed, intrapopula- and Jack 2004), the probability of survival and tion translocation. breeding for individual subadult birds, on av- The growth of the Marine Corps Base Camp erage, is actually increased if they are translo- Lejeune red-cockaded woodpecker population cated from their natal territory to a new pop- (34 potential breeding groups in 1992 to 67 ulation. groups in 2002) is an excellent example of how populations larger than 30 potential breeding groups can be significantly increased in a rel- Translocation Success atively short time, without translocation, by Definitions of Success using appropriate habitat improvement and management techniques (Walters 2004). Sim- Translocation success of individual red- ilarly, recipient populations must also have cockaded woodpeckers has been defined in rigorous habitat programs in place to receive various ways (Allen et al. 1993; Costa and birds, including four suitable cavities per re- Kennedy 1994; Hess and Costa 1995; Carrie cruitment cluster, two recruitment clusters et al. 1999; Franzreb 1999; Edwards and available for each pair of birds received, and an Costa 2004). Therefore, meaningful inter- aggressive prescribed burning program (U.S. pretations and comparisons of existing data Fish and Wildlife Service 2003). Given that sets remain difficult. Translocation success quality habitat conditions were required and at the population scale is relatively easy all recipient populations were essentially con- to quantify, being measured in number of sistent in this regard, we believe the rapid potential breeding groups. Instituted in 1998, growth of our study populations was directly the U.S. Fish and Wildlife Service’s Annual related to translocation. Furthermore, we have Red-cockaded Woodpecker Population Data no examples of populations similar in size than Report now provides a comprehensive and our study populations that have substantially systematic reporting procedure to help track increased in size without translocations. and document success of all translocated birds. We strongly recommend identification of Intensive monitoring at recipient populations success criteria for all longleaf ecosystem fauna 11. Reintroduction of Fauna 347 translocations prior to program implemen- minor donor populations and secondary recip- tation. However, more than one definition ient populations involved fewer than 10 birds may be necessary depending on the programs’ during the entire period they are not included objectives. Additionally, it is imperative that in the analyses below. development and implementation of com- Overall, numbers of occupied clusters, to- prehensive postrelease monitoring protocols tal birds, and potential breeding groups more for all translocation programs be established. than doubled in the 30 primary recipient pop- Without clearly defined and understood mea- ulations from initiation of translocation to the sures of success and comprehensive monitor- nesting season of 2002 or 2003; the latest years ing to confirm translocation results, program available for the corresponding data (Table 3). improvements will be questioned and support On average, by population, occupied clusters for the initiative will be at risk. increased from 7.5 to 15.8, number of birds from 17.7 to 43.6, and potential breeding Translocation Results groups from 6.0 to 13.9. Populations received birds for 2 to 14 years, average of 6.4. The num- Here we offer a preliminary summary of ber of birds received annually ranged from 1.4 the overall population level successes of red- to 10.8 and averaged 5.0. From initiation of cockaded woodpecker translocations. From translocation to 2002 or 2003, each population 1989 to 2002, 1014 birds were translocated: received, in total, an average of 29 birds; range 941 came from 11 major donor populations of 3–130. The majority (70%) of populations and 73 came from 19 minor donor popula- received more than 20 birds and 43% received tions. Birds were translocated to 30 primary 30 or more. These results demonstrate that, on recipient populations and 19 secondary re- average, small populations receiving five birds cipient populations. During the 14 years, the per year doubled in size in approximately 6 mean number of birds translocated from the years. 19 minor donors was 4, ranging from 1 to 8. In the 11 major donor populations, num- Most of these birds went to primary recipi- ber of pre-translocation occupied clusters and ent populations. Of the 19 secondary recipi- potential breeding groups increased 8.7% ent populations, 17 (89%) received fewer than and 17.3%, respectively, during this period 10 birds (range 1–14). These birds came from (Table 4). Over the 14-year period, 941 birds both major and minor donor populations. Be- were translocated from the 11 donor popula- cause the majority of translocations involving tions, averaging 85.5 (range 12–253) birds per

TABLE 3. Pretranslocation program initiation and 2002 or 2003 population demography (number of active clusters, adult birds, and potential breeding groups) for 30 primary recipients of translocated red-cockaded woodpeckers, and total number of birds translocated to the properties from 1989 to 2002. Pretranslocation 2002 or 2003 # birds # birds # populations # clustersa # birdsb # PBGc translocatedd # yearse per yearf # clustersg # birdsh # PBGi Total 30 225 532 180 866 148.5 475 1307 416 Mean 7.5 17.7 6.0 28.9 6.4 4.95 15.8 43.6 13.9 a Number of active clusters prior to translocation. b Number of adult birds prior to translocation. c Number of potential breeding groups prior to translocation. d Total number of birds translocated to the populations from different populations (i.e., intrapopulation moves are not included). e Mean number of years populations have received birds; most populations received some birds each year. f Mean number of birds received annually. g Number of active clusters pre-nesting season in 2002 or 2003. h Number of adult birds pre-nesting season in 2002 or 2003. i Number of potential breeding groups pre-nesting season in 2002 or 2003. 348 IV. Restoration

TABLE 4. Pretranslocation program initiation and 2002 or 2003 population demography (number of active clusters and potential breeding groups) for the 11 primary donors of translocated red-cockaded woodpeckers and total number of birds translocated from the properties from 1989 to 2002. Pretranslocation 2002 or 2003 # birds # birds # populations # clustersa # PBGb translocatedc # yearsd per yeare # clustersf # PBGg Total 11 1958 1626 941 97.8 2145 1968 Mean 178.0 148.0 67 birds per year 8.3 8.9 195.0 179.0 a Number of active clusters prior to translocation. b Number of potential breeding groups prior to translocation. c Total number of birds translocated from all populations from 1989 to 2002. d Mean number of years populations donated birds. e Number of birds per year donated. f Number of active clusters pre-nesting season in 2002 or 2003. g Number of potential breeding groups pre-nesting season in 2002 or 2003. population and 67 birds per year. On average, peckers have been translocated to 49 different populations donated birds for 8.3 (range 2–14) populations since 1986, we have a good un- years and translocated 8.9 (range 3–18.1) birds derstanding of what affects success at both the annually. Although, in total, donor popula- individual bird and population level for this tions increased in size, 6 of the 11 populations species. Notably, most studies prior to those only remained stable during their donor years. published in Red-cockaded Woodpecker: Road to Population increases in the other 5 populations Recovery (Costa and Daniels 2004) focused on were directly attributable to habitat manage- factors related to success of individual bird ment programs, including prescribed burning translocations, e.g., weather, forest type, re- and artificial cavity installation in occupied and cruitment cluster conditions, age, status, and recruitment clusters. Without such population sex of bird, distance moved, time of year, and habitat management and evaluation pro- and method of release. However, numerous cedures, managers may falsely conclude that small population translocation studies pub- removal of birds from the population may lished in Costa and Daniels (2004) focused contribute to their inability to increase their on factors that, in the authors’ opinions, af- population. Conversely, a declining population fected overall success of their population-level with an aggressive habitat management and growth, e.g., kleptoparasite control, numbers translocation program in place may indicate an of birds translocated, and condition of forest adverse effect from translocation. However, no structure. Using all of these “success” studies, such examples exist, and furthermore U.S. Fish we review factors important for red-cockaded and Wildlife Service donor population translo- woodpecker translocations and suggest how cation guidelines (see U.S. Fish and Wildlife these and perhaps other factors would affect Service 2003) would very likely prevent this reintroduction of other fauna into longleaf scenario from occurring. pine forests. Factors affecting translocation success of red-cockaded woodpeckers and other fauna Factors Potentially Related to Success may be broadly categorized as follows: envi- Various factors, working independently or ronmental, e.g., weather, forest type, phys- cumulatively, affect the short- and long- iographic province (equivalent to ecoregions term success of individual bird translocations. and recovery units), habitat conditions (e.g., Similarly, these factors in combination with cavity suitability and forest structure), and population-level variables affect the long-term kleptoparasites (predators and competitors for translocation success at the population scale. other species); demographic, e.g., age, status Because more than 1000 red-cockaded wood- (breeder, helper, fledgling), and sex; logistics, 11. Reintroduction of Fauna 349 e.g., distance translocated; temporal, e.g., sea- consider effects of weather, forest type, or son of year and time of day moved; transloca- physiographic province on translocation suc- tion type, e.g., single bird to an established soli- cess of any animal species. Effects of these vari- tary bird or multiple, unrelated subadult pairs; ables on red-cockaded woodpecker transloca- release method, e.g., from cavity, aviary, or tions were not significant. Likely, in part, the free-release, and program methodology, e.g., nonsignificance of weather was directly re- duration of translocations and numbers of an- lated to temperature and precipitation guide- imals translocated. Although no red-cockaded lines established for red-cockaded woodpecker woodpecker research team has incorporated, translocations to avoid potential adverse ef- or controlled for all of these variables in their fects (see U.S. Fish and Wildlife Service 2003). translocation analyses, one or more authors We suggest that similar guidelines be devel- have examined these factors for their effect at oped for any animal species being consid- the individual bird or population level. ered for translocation. Additionally, we recommend that for other species being considered for reintroduction into longleaf pine forests, Environmental Factors caution be exercised in choosing forest type Weather, Forest Type, Physiographic and physiographic province or ecoregion of Province the donor population; near is better than distant. We base this recommendation on the Only one study has examined effects of need to minimize disruption of locally adapted weather conditions, forest type, and physio- populations. graphic province at both capture and release sites on translocation success of individual Habitat Conditions birds. Edwards and Costa (2004), by establishing the conservative success criteria of “an in- Cavity Suitability and Quantity. Hess and Costa dividual remaining at the release cluster, fol- (1995) attributed several failed translocations lowed by pairing and breeding,” were able to in 1989/1990 to poor condition and quality of make inferences regarding effects of selected release cavities. Fortunately, all of these situ- environmental variables on translocation suc- ations could be remedied post-1990/1991 due cess of about 158 individual birds from 1989 to to development of artificial cavity technology 1995. Overall, 27% of translocations were suc- (see Copeyon 1990; Allen 1991). They also rec- cessful. Success was not related to weather, for- ommended a minimum of three suitable cav- est type, or physiographic province. Haig et al. ities be available when a female is being re- (1996) suggested forest habitat type does not leased in a solitary male cluster. Similarly, nu- limit dispersal of red-cockaded woodpeckers. merous managers have recommended that al- This may, in part, explain why Edwards and ternative roost cavities, for example, at least Costa (2004) found no significant differences four per cluster, should be available at translo- in success of translocated birds from one forest cation recipient clusters (Hagan et al. 2004a; type to another. Although Edwards and Costa Lohr 2004; Stober and Jack 2004). Others rec- (2004) found no differences in success related ommended providing extra cavities to mini- to physiographic province, in consideration of mize impacts of kleptoparasites and ensure all minimizing the potential for genetic disrup- potential fledglings had a roost cavity (Marston tion of “locally adapted populations,” Stangel and Morrow 2004). et al. (1992) and Haig et al. (1994, 1996) sug- After a decade of artificial cavity use for gested using local or geographically proximate translocations, several researchers have populations as donors. Typically, these donors presented data suggesting cavity age and would be in the same physiographic province cavity maintenance may affect translocation as the recipient. success of individual birds. Saenz et al. (2004) Published reviews of translocations (e.g., suggested that the newest artificial (inserts) Griffith et al. 1989; Wolf et al. 1996) do not cavities, those less than 1 year old, positively 350 IV. Restoration influenced translocation success when small populations is that their population- compared to translocations to older (greater level translocation successes are largely a re- than 1 year old) artificial cavities. However, sult of “best possible condition” of cluster due to confounding variables in their data habitat (Hess and Costa 1995), “high-quality” set they acknowledged, “. . . it is possible that habitat (Hagan et al. 2004a), “high quality the number of successful translocations is of the habitat’s structure” (Stober and Jack independent of the condition [age] of the 2004), and plentiful habitat in “excellent con- inserts . . . .” However, Marston and Marrow dition” (Hedman et al. 2004). Maintaining the (2004) also suggested that their population open, midstory-free forest structure required increase was in part attributed to maintaining for participation in red-cockaded woodpecker cavity inserts in a “newer-like condition.” Neal translocation programs is best accomplished and Montague (2000), strong proponents of with prescribed fire (see U.S. Fish and Wildlife prolonging insert suitability via regular inte- Service 2003). rior cleaning and exterior maintenance, have Habitat condition at recipient sites for all an- substantially increased their red-cockaded imal reintroductions is of paramount impor- woodpecker population using artificial cavi- tance (Griffith et al. 1989, 1990; Kleiman 1990; ties and translocations. Many of their cavities IUCN/SSC RSG 1995; Wolf et al. 1996). It is have been in use for numerous (5–10+) critical that all limiting habitat factors be identi- years. fied, addressed, and mitigated. All of the target species’ required habitat components, includ- Forest Structure. Most managers and researchers ing suitable area for translocations and natu- agree that habitat quality of the release clus- ral population growth, and appropriate struc- ter and surrounding territory is important to ture and composition, must be in place prior to translocation success; however, its impact has program initiation. Adequate amounts of high- never been measured. It is, and will remain for quality habitat, including microhabitat compo- the near future, difficult to impossible to eval- nents, like nest structures, must be the foun- uate effects of some factors thought to affect dation of any longleaf pine ecosystem fauna translocation success. Primarily, this is because translocation program. investment in individual birds is high and managers cannot afford (ecologically, politically, Kleptoparasites socially, or economically) to use translocated birds in experiments to test, for example, their Relationships between kleptoparasites(Kappes response to less-than-optimal habitat. Addi- 1997) and translocated red-cockaded wood- tionally, sample sizes would always be small, peckers are difficult to study due to the i.e., fewer than 30 potential breeding groups, significant value of individual donor birds for both control and treatment groups, leav- and sample size (see above). For example, ing correlative or experimental results to be managers would not intentionally want to questioned. expose their translocated birds to predation Red-cockaded woodpeckers evolved in an or cavity kleptoparasitism by not controlling open, parklike ecosystem and have subse- target predators or kleptoparasites if they quently been extirpated from forests where believed control was necessary. Although these conditions were not present, e.g., forests some associations between red-cockaded with a dense hardwood midstory (see Costa woodpeckers and red-bellied woodpeckers and Escano 1989). We believe translocation have been investigated, our overall knowledge success is improved when forest structure of kleptoparasite impacts remains very limited, replicates historic conditions to the maxi- particularly for small populations (Kappes and mum extent possible (see Platt et al. 1988a). Harris 1995). For example, little to no research Although untested, and, therefore, specula- exists on relationships between other avian tive, professional opinions among managers cavity usurpers, e.g., red-headed woodpeckers and researchers responsible for expanding (Melanerpes erythrocephalus) and northern 11. Reintroduction of Fauna 351

flickers (Colaptes auratus), and red-cockaded number of fledglings in groups where south- woodpeckers. ern flying squirrels were removed. They sug- Similarly, only one study examined effects gested removal of southern flying squirrels is of southern flying squirrels on red-cockaded unnecessary in “healthy” populations of red- woodpecker nesting success in a small popula- cockaded woodpeckers. tion (Conner et al. 1996). However, relation- Given the current absence of experimental ships between translocated birds and south- data regarding interactions between southern ern flying squirrels were not investigated in flying squirrels and red-cockaded woodpeck- that study. Additionally, the study was cor- ers in small populations involved with translo- relative, not experimental, and was based on cations, we rely, for now, on professional observational evidence of a small sample size. opinion whether controlling southern flying Conner et al. (1996) found no negative effect squirrels increases population-level translo- of southern flying squirrels on red-cockaded cation success. Numerous managers and re- woodpeckers and concluded that any effect on searchers involved in growing very small pop- a “healthy” red-cockaded woodpecker popu- ulations to larger sizes, e.g., from fewer than 3 lation is likely minimal; we believe impacts to 8 to 10 occupied clusters, within a relatively may not be minimal if set within an exist- short period of time (4 to 5 years), suggested ing metapopulation structure. By definition, that controlling southern flying squirrels was a small populations receiving birds via translo- contributing factor in their translocation suc- cations cannot be considered healthy. They cess. In some cases, large numbers of squir- are constantly at risk from potential stochas- rels were removed, e.g., 26–108 per month for tic demographic and environmental threats. 5 years (Allen et al. 1993) and 2304 over 9 Indeed, Conner et al. (1996) acknowledged years (Franzreb 1997b), while red-cockaded that when control is considered beneficial, woodpecker populations increased substan- it should only occur while the red-cockaded tially from 1 to 6 pairs and 1 to 19 pairs, re- woodpecker population is “small and vulner- spectively. able to extirpation.” Small and vulnerable de- Managers of many other small population scribes translocation recipient populations. expansion programs, dependent on translo- A few additional studies, both controlled ex- cations, have agreed that southern flying periments, researched southern flying squir- squirrel control is an important management rel and red-cockaded woodpecker relation- activity. Marston and Marrow (2004) sug- ships (Laves and Loeb 1999; Mitchell et al. gested that southern flying squirrel removal 1999). However, both of these studies were contributed to: (1) increased nesting and conducted in large (greater than 100 poten- fledgling success, (2) occupation of recruit- tial breeding groups) populations and neither ment clusters by dispersing fledglings, and involved translocated birds. Laves and Loeb (3) quality and longevity of cavities. All of (1999) found that southern flying squirrels’ these observations related directly or indi- use of red-cockaded woodpecker cavities in rectly to translocation success. Other stud- the breeding season had a significant negative ies have been more direct about benefits of effect on red-cockaded woodpecker reproduc- southern flying squirrel removal on translo- tion. Southern flying squirrel removal resulted cation success. Stober and Jack (2004) stated, in significantly more fledglings produced in “flying squirrel removal is an essential man- treatment (removal) clusters versus control agement action when establishing new red- (no removal) clusters. They concluded that in cockaded woodpecker populations.” Hedman very small or declining red-cockaded wood- et al. (2004) believed that southern flying pecker populations (e.g., translocation recip- squirrels were “. . . exerting considerable pres- ient populations), where any nest loss could sure . . . ” on translocated birds (presumably result in extirpation, southern flying squir- for cavities, although this was not stated), rel control might be necessary. In contrast, and concluded that controlling southern fly- Mitchell et al. (1999) found no increase in the ing squirrels had a “major role” in the growth 352 IV. Restoration of their red-cockaded woodpecker population. competitors, their presence and abundance, Similarly, Poirier et al. (2004) believed that and need for and method of control at the re- the 14-fold red-cockaded woodpecker popula- cipient location. Control, if required, should be tion increase would not have occurred without carefully monitored and designed in conjunc- a southern flying squirrel removal program. tion with the increasing size of the translocated Lohr (2004), studying a very small popula- population to be a relatively short-term pro- tion, suggested southern flying squirrels posed gram. If, based on preliminary analyses, it is a threat to efforts to expand that population. determined control will be necessary for years In the first red-cockaded woodpecker reintro- (10+) even as the translocated population induction (see Hagan and Costa 2001), Hagan et creases in size, the overall reintroduction pro- al. (2004a) believed southern flying squirrels gram should be reevaluated. were potentially “limiting population expansion”; therefore, they were controlled for the first several years of the reintroduction pro- Demographic Factors gram. Age, Status, and Sex In conclusion, many managers and researchers agree that controlling southern fly- Early (1986 to 1995) research and population ing squirrels is/may be necessary in small, rescue translocations of red-cockaded wood- at-risk red-cockaded woodpecker populations. peckers used birds of various ages and sta- We do not advocate control as a long-term tus, e.g., breeder, helper, or subadult, of both population maintenance activity in medium sexes (Allen et al. 1993; Hess and Costa 1995). to large populations, but as a potential short- Documented successes of these translocations term population establishment or small pop- were somewhat inconsistent, in part due to ulation stabilization strategy. Kappes (2004) differences in success definitions, small sam- concluded that kleptoparasites might affect ple sizes, and inadequate monitoring (Rudolph red-cockaded woodpecker population size et al. 1992; Costa and Kennedy 1994). In (number of groups) through their impacts on spite of these inconsistent results, several find- reproduction and exacerbating cavity limita- ings were noteworthy. Franzreb (1999) found tion (see Poirier et al. 2004). For these reasons no significant differences in success between and others discussed above, it may/will be nec- translocation of younger (5–7 months old) essary, on a case-by-case basis, to control klep- and older (8–12 months old) subadult females. toparasites, particularly southern flying squir- Similarly, success was not related to subadult rels, in translocation recipient populations. age (4 to 8 months) for either males or females Translocation success has been positively as- in another study (Edwards and Costa 2004). sociated with introducing animals into areas Most studies did not recommend translocat- without congeneric competitors, morpholog- ing helper or adult males due to their hom- ically similar species, or species with similar ing tendency or poor success rates (Rudolph life history traits (Griffith et al. 1989, 1990). et al. 1992; Carrie et al. 1996; Franzreb 1999; Results discussed above, regarding southern see Table 5). Conversely, most studies recom- flying squirrels, support the premise that mended translocating either subadult or adult competitors (kleptoparasites) with at least one females; success ranged from moderate to high similar, but critical, life history trait, i.e., cav- (Allen et al. 1993; Costa and Kennedy 1994; ity use, can potentially affect success of a Hess and Costa 1995; Franzreb 1999; Edwards translocation program. This effect will likely and Costa 2004; see Table 5). be particularly severe if the shared trait in- Based on past research and translocation volves the primary limiting factor for the intro- results, we have an excellent understanding duced species, for example roosting and nest- of which age class and sex ratios to translo- ing cavities for red-cockaded woodpeckers. cate to maximize translocation success for Reintroduction programs for other longleaf red-cockaded woodpecker populations. It is pine ecosystem fauna should identify potential imperative for other translocation candidates 11. Reintroduction of Fauna 353

TABLE 5. Red-cockaded woodpecker translocation successa reported from 1994 to 2004.b Translocation type Study Female to male Male to female Potential pair Success criteria Allen et al. (1993) 40% (4 of 10) 25% 33% (2 of 6) Remained in the population and successfully bred Costa and Kennedy (1994) 62% (48 of 77) 38% 33% (18 of 54) Ranged from “interacted well” to “fledged young” Hess and Costa (1995) 61% (11 of 18) — — Remaining at release cluster through subsequent breeding season Carrie et al. (1999) — — 65% (11 of 17)c Remained in the population and successfully bred Franzreb (1999) 82% (18 of 22) 33% 40% (4 of 10) Remained in the vicinity of release cluster for ≥30 days Edwards and Costa (2004) 42% (36 of 86) 15% 13% (10 of 79) Remaining at the release cluster, followed by pairing and nesting a Success was measured in various ways. Success applies to individual or multiple translocation “events” over time involving individual birds, pairs of birds, or multiple pairs of birds. Success is not, in this table, referring to the success of growing target populations. b Source (with slight modifications): Edwards and Costa (2004). c Included translocation of 5 potential pairs to inactive clusters and later 6 additional single birds (3M, 2F) to solitary individuals, and 1 solitary male was released to determine whether he would remain at the release site. that appropriate age class, breeding status, to their capture clusters (Allen et al. 1993). and sex ratios are determined. This is best Others have also documented birds, particu- accomplished basing decisions on species’ life larly males, returning to their capture clusters history, ecology gained from studies of wild when moved short to moderate distances, e.g., populations, and by following adaptive man- 3 km (Rudolph et al. 1992), 35 km (Reinman agement principles based on field trials with 1984) and 16, 26, and 29 km (same bird moved rigorous monitoring. We recommend follow- three times) (Carrie et al. 1996). Following the ing these procedures with all translocation U.S. Fish and Wildlife Service’s translocation programs. bird allocation guidelines and given the paucity of donor populations and dispersion of recipi- Logistic Factors ent populations throughout the Southeast, few Distance Translocated red-cockaded woodpeckers are now translocated less than 32 km, except for intrapopu- Several studies have evaluated effects of dis- lation moves. tance moved on red-cockaded woodpecker We found no “distance moved” related sum- translocation success. Franzreb (1999) found maries for other species’ translocations. For that less than 8% of birds moved 18 to 464 longleaf pine ecosystem fauna translocations, km returned home, while almost 42% of birds we recommend that unless there is some com- moved less than 6 km returned home. Edwards pelling reason to move animals long distances, and Costa (2004), summarizing 159 transloca- for example, their ability to home, they be tions, found no significant differences among moved as short a distance as practical. This distances moved. Their mean distance for suc- strategy may minimize costs, stress related cessful moves was 350 km (range of 2 to to time in captivity, and potential for dis- 630) and for unsuccessful moves was 405 km ease transmission. Additionally, doing so may (range of 8 to 493). In another study, four help avoid disruption of local gene complexes adult males moved less than 19 km returned and ensure that probable local morphological 354 IV. Restoration adaptations are not compromised in a different suggest the time of year animals are translo- environment. cated should be based on their dispersal patterns or response to breeding seasons. Addi- tionally, weather, ability to obtain food, and Temporal Factors ability to avoid, or at least minimize, predation Season of Year and competition must be considered. Finally, translocations should attempt to minimize so- Edwards and Costa (2004) found that success cial disruption of, and maximize the probabil- rates for birds translocated in fall (September– ity of integration into, the extant population. December) and spring (January–April) were Regarding time of day, and with the exception not different. Since birds are not moved un- of the possible need for daylight to acclimate til the fall, it must be that maturity of young to specific habitat needs for some species, we birds has been accounted for by the time they believe it to be a minor or unimportant fac- reach 5–6 months of age. These translocation tor for most species’ translocations. If these or periods are consistent with dispersal patterns other temporal factors are thought to affect the for both young males and females in natural potential for translocation success, they should situations. be addressed prior to initiation of the program. Time of Day Translocation Type Edwards and Costa (2004) found no significant differences in success whether birds were Single Bird to Established Solitary Bird moved during the day or night. Day moves in- Success of translocating single birds to es- volved trapping the bird at dawn, transporting tablished solitary birds has varied depending and feeding it all day, and placing it in its re- on age and sex of both the bird moved and lease cavity at dusk to be released the following the recipient bird (see Table 5). Allen et al. morning. Night moves involved trapping the (1993) found that success of translocating fe- bird at dusk, transporting it to its recipient clus- males to solitary males had mixed results, pri- ter, typically by midnight, placing it in its re- marily depending on the male’s age. Females lease cavity, and releasing it at dawn. Although moved to adult males all bred, whereas no fe- “time in captivity” data were not collected, typ- males moved to subadult males bred. Costa ically day moves involved considerably more and Kennedy (1994) documented that 62% time in captivity than night moves. Addition- of females translocated to solitary males were ally, day moves required handling birds six successful. In contrast, only 25% of males to eight times for feeding (at 45-minute in- moved to solitary females were successful. tervals), a potentially stressful situation. How- Hess and Costa (1995) found that 61% of fe- ever, on average, a bird moved in the day spent males translocated to solitary males were suc- 4 to 6 more hours in its release cavity, for a to- cessful. Franzreb (1999) recorded 82% success tal of about 12 hours, than a bird moved in the for females moved to resident males. Finally, night. This not only provided more potential Edwards and Costa (2004) documented suc- rest compared to birds moved at night, which cess rates of 42% for females moved to solitary only spent 6 to 8 total hours in their release males. cavity, but also allowed for additional cavity and site acclimation time. However, the con- Multiple, Unrelated Subadult Pairs to tribution, if any, of these factors to success is Recruitment Clusters questionable, given the lack of significant differences between day and night moves. Although early translocations (1986 to 1995) Based on our red-cockaded woodpecker of unrelated, subadult pairs to establish poten- data set we offer the following recommenda- tial breeding groups were minimally successful tions for other fauna being considered for rein- (see Table 5), most managers and researchers troduction in longleaf pine forests. First, we continued to believe the technique had 11. Reintroduction of Fauna 355 promise (Costa and Kennedy 1994; Franzreb We recommend the structure and compo- 1999; Edwards and Costa 2004). Importantly, sition of fauna translocation cohorts that are none of these early attempts involved multi- reintroduced into an area be based on the ple pairs of birds, a technique that would not species sociobiology. The number of animals be instituted as a management practice until to introduce with each translocation attempt, the late 1990s (see Rudolph et al. 1992). by age class, sex, breeding or social status, and In 1994/1995, Carrie et al. (1999) further other demographic factors, should be initially investigated the pair translocation technique, determined based on knowledge of wild pop- but used multiple pairs, although not simulta- ulations’ demographics and sociobiology. If a neously. They achieved a 65% success rate, de- thorough understanding of species sociobiol- fined as birds remained in the population and ogy is lacking, then research should be con- bred. Based on this success, and the relatively ducted to answer critical questions. poor success of single-pair translocations, the U.S. Fish and Wildlife Service adopted a policy of not, except in extenuating circumstances, Release Method translocating single pairs. Instead, pair translo- From Cavity cations after about 1997 involved multiple With very few exceptions, all red-cockaded pairs, typically a minimum of two to three, and woodpecker translocations have involved the not uncommonly four or five, and were con- release method described by DeFazio et al. ducted on the same day. (1987). This usually involves trapping the bird It was decided to thoroughly evaluate the at dusk with a small mist net or mosquito multiple-pair, simultaneous release concept bag net attached to a telescoping pole (Jack- while concurrently attempting for the first son 1977), placing it in a wooden/wire box, time to reintroduce a red-cockaded wood- transporting it to the release cluster, placing it pecker population de novo, i.e., in the absence in the appropriate cavity, placing wire screen of a founder population. The Turner Endan- over the cavity entrance with a string attached gered Species Fund and the U.S. Fish and that reaches the ground, and pulling the screen Wildlife Service embarked on a 5-year research off at dawn. When a bird is released in an es- project (Hagan and Costa 2001). The study tablished solitary bird group the translocated proposal was fully implemented in 1998 and bird is not released until the resident bird ex- results were remarkably successful. By 2002, its its cavity. When unrelated subadults are re- the population harbored 15 occupied clusters, leased in a recruitment cluster, they are re- 15 potential breeding groups, and 58 birds (Ha- leased simultaneously after each bird appears gan et al. 2004a). The success of simultane- at its screen. Although this release method ously releasing multiple pairs of birds is directly has proven acceptable for red-cockaded wood- related to the principle that success increases peckers, different methods had been used for with increasing numbers of animals released. other bird species (see Ellis et al. 1978; Barclay All small population growth success sto- and Cade 1983). ries from the mid to late 1990s through the early 2000s employed the multiple-pair (typi- From Aviary cally two to five pairs), same-day release strategy (Hagan and Costa 2001) rigorously im- Scott and Carpenter (1987) suggested that one plemented by Hagan et al. (2004a) (Brown technique for translocating birds is to do a and Simpkins 2004; Hedman et al. 2004; Lohr “soft” release, i.e., holding them in captivity 2004; Marston and Marrow 2004; Morris and for some designated time at the release site. Werner 2004; Stober and Jack 2004). This They hypothesized this process may help birds model is the new and successful paradigm for become acclimated, and perhaps imprinted, expanding small, at-risk red-cockaded wood- to their new environment. Franzreb (1997a) pecker populations, and reintroducing new designed and tested for durability a mobile, populations. metal-framed aviary for the eventual holding 356 IV. Restoration and release of red-cockaded woodpeckers. Al- Free-release involved capturing birds at dawn, though proving durable, Edwards et al. (1999) transporting them to their release cluster, found this design’s takedown and reassembly and immediately releasing them. Of the free- to be complex and, in combination with its release birds, 36% remained at their release weight, essentially nonmobile. Therefore, Ed- cluster for at least 30 days, while 14% of cavity wards et al. (1999) modified the design and released birds did so. Although retention at the constructed an aviary using PVC plastic pipe release clusters was relatively low, the major- for the frame. Their final design met their ob- ity (71%) of both the free-released and cavity jectives; it was lightweight and easy and fast released (76%) birds stayed within the study (less than a day) to assemble. The efficacy of area (see Walters et al. 2004). In addition, this design was used in two different experi- six free-releases in central Florida were suc- ments. cessful while two failed (R. DeLotelle unpub- Franzreb (2004) confined 22 red-cockaded lished data). Two of the six successful translo- woodpeckers in aviaries. Thirteen (59%) birds cations included males to solitary females. If completed the experiment, remaining in the further studies verify free-release success rates aviary for 9 to 14 days. The other 9 birds ei- are at least equivalent to cavity release rates, it ther died in the aviary (n = 2) or were released may be a preferred method for short-distance (n = 7) for various reasons prior to 9 days. Of moves. All of the free-releases of Walters et al. the 13 birds remaining in the aviary, 61% were (2004) were intrapopulation and, therefore, successful upon release; success was defined birds were at their release cluster in several as remaining at, or in the vicinity of, the re- hours or less, providing adequate time for lo- lease cluster for 30 days or more. Therefore, cating cavities and acclimating to the site prior overall success rate was 36% (59% × 61%). to dark. Free-releases made later in the day The 61% of the birds successfully completing would minimize or eliminate these potential the experiment were not statistically different benefits and may decrease success rates. from “hard” release (see DeFazio et al. 1987) We believe that for the majority of potential results (63%) reported by Franzreb (1999) longleaf pine ecosystem reintroduction can- for translocations at the same location. How- didates, some variation of cavity (e.g., nest ever, the 36% overall success is considerably box, burrow) or free-release method will prove poorer than hard release results. Edwards et practicable and successful. With some excep- al. (2004), using the same success criteria as tions, e.g., red wolf, captive propagation and Franzreb (2004), conducted six aviary translo- holding wild animals in captivity prior to re- cations (birds held 10 to 14 days), paired with lease is not likely to become a major program six concurrent hard releases. Only one soft- in translocations. These methods can be pro- release bird (17%) remained for greater than hibitively expensive and logistically challeng- 30 days (256+ days); another stayed 5 days ing and, therefore, with the exception of an- and the other four dispersed the day of re- imals at high risk (e.g., whooping cranes and lease. Based on results of these two studies, Florida panthers), not preferred as long as wild, which had worse or no better success rates donor stock is available. than hard releases, managers are understand- ably unwilling to use aviaries for red-cockaded woodpecker translocations, given equipment Translocation Methodology and construction costs when compared to hard Number and Duration of Translocations releases, bird-maintenance requirements, and and Numbers of Animals Translocated logistics. Griffith et al. (1990) found that success rates of Free-Release threatened, endangered, and sensitive species translocations improved with increasing num- Walters et al. (2004) reported success rates bers of animals released and years of releases. for free-released and cavity released birds. Wolf et al. (1996) also found numbers of 11. Reintroduction of Fauna 357 animals released to be associated with suc- ment program, prescribed burning, and well- cess; however, number and duration of re- planned and implemented silvicultural prac- leases were not important. Although our red- tices. Except for Marine Corps Base Camp Leje- cockaded woodpecker data did not provide une, all of these sites are translocation donor comparisons between populations for num- populations. bers released and duration and number of re- Several studies have investigated effects of leases, we did find that, on average, small red- removing subadult birds from donor popula- cockaded woodpecker populations could be tions. Walters et al. (2004) documented a 4% doubled in size in approximately 6 years by increase in their “donor plot” (a subset of the translocating about 30 birds at a rate of 5 per entire population) as five to eight birds were year. Based on our results, and those of oth- removed annually. While unable to specifically ers, we recommend that to the maximum ex- identify why removal had no effect, they sug- tent practicable translocation of other fauna gested the location of donor clusters in the cen- into longleaf pine forests involve relatively ter of the population and rigorous adherence large numbers of individuals released annually to U.S. Fish and Wildlife Service guidelines over multiple years. We believe such a strategy specifically designed to minimize group- and would increase both the survival probability of population-level impacts, as probable reasons. individual animals during the short term and However, Walters et al. (2004) did report a re- the probability of successfully augmenting an duction of group size in their donor plot. Hagan existing, or reintroducing a new, population et al. (2004b) found no significant differences over the long term. in number of fledglings, mean group size, and percent change in occupied clusters between donor and nondonor groups. Acknowledging Translocation Concerns an experimental design problem, they still con- Effects of Translocation on Donor cluded that overall, translocations did not ap- Populations pear to harm the donor population. In addition to these studies, it is encouraging A fundamental premise of red-cockaded that several primary donor populations con- woodpecker translocation programs is that tinued to significantly increase in size, while donor populations must remain stable, if they birds were removed (see above). These and have attained their population goal, or a min- other examples provide critical documentation imum size or be increasing if they have not supporting the premise that large populations (U.S. Fish and Wildlife Service 2003). Expand- can function as a valuable source for translo- ing the size of larger populations has proven cated birds while simultaneously increasing relatively straightforward when proper habitat their own size. Smaller populations can also and population management techniques are serve as donors as long as rigorous group and applied. Numerous examples of the population population monitoring is conducted, translo- growth potential (5–10% annually, on aver- cation guidelines are followed, and birds are age) of medium to large populations exist, in- removed during years of surplus (DeLotelle et cluding populations at Fort Bragg, NC (Britcher al. 1995; U.S. Fish and Wildlife Service 2003). and Patten 2004), Fort Stewart, GA (Carlile Because of the paucity of literature on ef- et al. 2004), Fort Benning, GA (Doresky et fects of removing animals from wild donor al. 2004), Elgin Air Force Base, FL (Petrick populations, we strongly recommend that wild and Hagedorn 2004), Marine Corps Base Camp populations proposed as donors in longleaf Lejeune, NC (Walters 2004), and Carolina pine fauna translocation programs be moni- Sandhills National Wildlife Refuge, SC (U.S. tored annually. Monitoring must be adequate Fish and Wildlife Service unpublished data). to ensure that potential adverse impacts from The common attributes of their recovery pro- removals can be detected and distinguished grams included aggressive recruitment clus- from normal demographic variation. Monitor- ter establishment, an artificial cavity manage- ing is also required to determine availability, 358 IV. Restoration i.e., numbers and frequency, of animals for gious pathogens and, therefore, have no ac- translocation. Understanding demography, required immunity to them if they arrive with productive potential, current size, and overall translocated animals (IUCN/SSC RSG 1995). trend of donor populations is critical not only For migratory and wide-ranging species and to determine availability of animals but also to those species prone to dispersals, e.g., red- evaluate potential impacts of removal. cockaded woodpeckers, between geographically isolated populations, the potential for Diseases and Parasites transmission of diseases, parasites, and other pathogens may be minimal. However, translo- Prior to the 1990s, only one study had been cating large numbers of animals over rela- conducted on blood parasites of red-cockaded tively short time periods may be a concern. woodpeckers (Love et al. 1953). In recent stud- At relatively low rates of natural immigra- ies several blood parasites have been found tion, recipient populations may have time to and some have been identified (Luttrell et al. adapt and develop normal levels of resistance 1995; Pung et al. 2000). Based on their find- and immunity to diseases, parasites, and other ings, Pung et al. (2000) suggested caution if pathogens. Under high rates of immigration, subadult, unexposed birds are translocated to as mimicked by an intensive translocation pro- populations where risk of infection is elevated gram, adaptation may be difficult to impossi- by other woodpeckers or higher numbers of ble and result in deleterious effects to recipient dipteran vectors. However, examination of in- populations. Caution must also be exercised to dividual birds for blood parasites is not practical minimize risk of exposing translocated animals given the numbers and timing of birds translo- to vectors of disease agents that may be present cated. Importantly, no instances of disease- at release sites, but not donor sites, for which related mortality have been documented from they have no acquired immunity (IUCN/SSC the many translocations conducted to date. RSG 1995). Nor do we believe the risk is significant, pri- The IUCN/SSC RSG (1995), Kleiman (1990), marily based on the species’ natural dispersal and many others have cautioned against behavior. the possibility of transmitting diseases via a Pung et al. (2000) found a “rich commu- translocation program. We share these con- nity” of arthropods in cavities, a cosmopoli- cerns. However, given the frequency of long- tan mite, Androlaelaps casalis, being the most distance dispersals (DeLotelle et al. 2004) and common blood-feeding arthropod. This mite research results to date, we do not believe that was found in 76% of the cavities at an av- disease or transmissions of other pathogens erage density of 51 (range 1–233) per cavity. via translocations is a serious concern for red- However, they concluded that A. casalis had cockaded woodpeckers. Given the level of cur- no effect on red-cockaded woodpecker fitness. rent knowledge about high disease level in go- Overall, they found occurrence and density of pher tortoises (Gopherus polyphemus) and lack blood-feeding insects low, even in cavities that of information on other potential longleaf pine had been in constant use for 20 years. ecosystem translocation candidates, we are not There is potential to transmit diseases, endo- as confident suggesting that transmissions of and ectoparasites, and infectious or contagious disease and other pathogenic agents are not pathogens via translocation programs for lon- a serious consideration for other species. We gleaf pine forest fauna. For species that would recommend a thorough review of available lit- otherwise not emigrate from one population to erature on each species, and field and labora- another, e.g., reptiles, amphibians, and small tory analyses for diseases, parasites, bacteria, mammals, potential for adverse effects at both and other potential pathogens on a sample of the recipient population and for translocated both donor and recipient populations. These animals may be elevated. That is, the recip- measures and precautions will help guide final ient population may never have been ex- decisions on whether or not to proceed with posed to certain diseases, parasites, or conta- translocations. Additionally, based on findings 11. Reintroduction of Fauna 359 of literature reviews and field and labora- tive (WRTC). With some modifications, this tory studies, if potential pathogens are discov- model has been replicated for remaining por- ered but a decision is made to proceed with tions of the birds’ range. In 2003, the South- a translocation program, specific guidelines, ern Range Translocation Cooperative (SRTC) procedures, and protocols must be developed (Alabama, Florida, Georgia, and Mississippi) and implemented to avoid or minimize impacts had 7 donor populations and approximately on recipient populations and translocated ani- 30 recipient populations, while the North- mals. ern Range Translocation Cooperative (NRTC) (Kentucky [formerly], North Carolina, South Geographic Variation Carolina, and Virginia) had 3 donor populations and 3 recipient populations. Within co- Mengel and Jackson (1977) measured multi- operatives, birds are primarily allocated to el- ple morphological characteristics of 430 red- igible populations based on population size, cockaded woodpecker museum specimens availability of suitable habitat with artificial from throughout their range. They found lat- cavities installed, and success of past transloca- itudinal differences in wing and tail length, tions. With release of the red-cockaded wood- both being longer farther north. Additionally, pecker recovery plan revision, it is expected culmens were relatively shorter in interior and that a population’s role in recovery, e.g., a “pri- northern populations, suggesting morphologi- mary core” versus a “significant support” pop- cal differences can also occur along habitat gra- ulation, will be increasingly important in its dients from Coastal Plain to Interior forest pop- recipient status in the bird allocation process ulations. (U.S. Fish and Wildlife Service 2003). Although small and inconclusive from a The allocation process is essentially one of management perspective, the body of knowl- informed consent, whereby participants at an- edge regarding diseases, parasites, bacteria, nual translocation cooperative meetings de- and geographic variation suggests that for termine recipient population priorities (see all longleaf pine ecosystem fauna transloca- Saenz et al. 2002). Importantly, respective state tions, the distance translocated should be lim- wildlife agencies and the U.S. Fish and Wildlife ited as much as possible. Limiting distance Service have necessary veto power if they be- translocated may help minimize transmission lieve a particular recipient or donor is not el- of pathogens and parasites because closer pop- igible to receive or provide birds, respectively. ulations may be locally adapted to potential For example, a recipient may be determined threats. Similarly, minimizing latitudinal dis- ineligible if release cluster habitat is not in ex- tance moved may improve an animal’s proba- cellent condition. A donor population would bility of survival given presumed adaptive sig- likely be denied interpopulation translocation nificant of known, or assumed, morphological authority if its population trend were not sta- differences seen geographically or by habitat ble or increasing. gradients. With few donors and many recipients in the WRTC and SRTC, the annual supply of Translocation Strategies available birds never meets the demand. This Regional Scale fact, along with the informed consent process, leaves managers unsure, on a year-to- In 1995, a multistate (Arkansas, Louisiana, Ok- year basis, whether they will be receiving lahoma, and Texas) partnership was formed birds. Saenz et al. (2002) recognized this is- to allocate annually the limited number of sue and designed and tested various potential red-cockaded woodpeckers available from the regional translocation models. Overall, they only 2 donor populations to the approxi- concluded that a partitioning strategy that an- mately 20 recipient populations (Saenz et nually allocated all available birds to the six al. 2002). This partnership is known as largest recipient populations, called the “elitist” the Western Range Translocation Coopera- model, was best at meeting conservation and 360 IV. Restoration translocation goals, defined as management red-cockaded woodpecker populations using objectives. Importantly, the “random” model, translocations. representing the informed consent system, The basic strategy for any small red- performed well, achieving rates of popula- cockaded woodpecker population translocation growth only slightly lower, and with tion program involves the following: (1) aug- fewer population extirpations, than the elitist ment solitary bird groups, typically male, (2) model. build from edges of the population’s core, Although Saenz et al. (2002) suggested us- (3) connect demographic subpopulations by ing a single strategy to minimize costs and sim- building from their proximate borders, (4) ac- plify the program, they also recognized flexi- complish 2 and 3 above by simultaneously bility is important for other reasons, including translocating multiple (at least three), unre- ensuring genetic diversity, addressing fluctu- lated subadult pairs annually or as frequently ating budgets, meeting previous commitments as possible until the population goal is reached, from prior years, and satisfying diverse goals of (5) provide one additional recruitment cluster managers. For the foreseeable future, the infor each pair released, and (6) arrange recruit- formed consent system will likely continue as ment clusters in clumped arrays as much as the basic red-cockaded woodpecker allocation possible. Following these guidelines will help process in the WRTC and SRTC. In the NRTC, ensure that population-specific translocation with only three recipients and three donors, strategy goals of: maximizing retention and long-term agreements have been established survival of individual birds, minimizing cost, between donor and recipient pairs, eliminat- and expanding the population to its potential ing the need for annual allocation meetings. size of 10–30 potential breeding groups, will be However, numerous additional populations in achieved as rapidly as possible (U.S. Fish and North and South Carolina harbor fewer than Wildlife Service 2003). 30 potential breeding groups and could be par- Overall, regional and population-level ticipating in translocation programs if they de- strategies for reintroducing fauna in lon- sired, thereby likely creating a need for annual gleaf pine ecosystems will vary by species. meetings. Types and scales of strategies employed will be dependent upon number and types of Population Scale cooperators involved, number of animals available, resources available, logistical con- At the population scale, red-cockaded wood- siderations, number of populations at risk of pecker (and most other longleaf pine for- extirpation, calculated extirpation risk to each est species) translocations have four poten- population, and species life history and ecol- tial applications: (1) prevent extirpations, (2) ogy. Developing and testing species-specific reduce isolation of groups or subpopula- strategies provides a foundation for adaptive tions, (3) genetic management, and (4) rein- management and increasing success. troduction (U.S. Fish and Wildlife Service 2003). Red-cockaded woodpecker transloca- The Future tions have been used since the late 1980s and early 1990s to achieve goals 1 and 2, and The red-cockaded woodpecker translocation since the mid-1990s to explore goal 4. Genetic program has been remarkably successful. management is likely occurring as other goals However, various challenges remain for main- are being achieved, but only one translocation taining and improving the program. Contin- program specifically designed to achieve ge- ued success hinges on filling key information netic management has been established (Haig gaps and increasing the number of available et al. 1993). After more than a decade (1989– birds to meet population needs. 2002) of pursuing these goals, we under- Additional research on rates of natural dis- stand how to stabilize, expand, and establish persal and new research on factors promoting 11. Reintroduction of Fauna 361 juvenile dispersal between populations are the NRTC, the three recipient populations needed (U.S. Fish and Wildlife Service 2003; all pay their respective donors approximately DeLotelle et al. 2004). Investigating these is- $1000 for the opportunity to receive each bird. sues will help us evaluate genetic threats and Donors use these funds to maintain the level of determine whether translocations will be nec- monitoring necessary to translocate subadult essary for genetic management of small- to birds. medium-size populations over the long term. It may be possible to increase the number of The paucity of information on diseases, para- birds available for translocation in the short- sites, and other potential pathogens, e.g., West term by allowing smaller populations to be Nile virus, within translocation donor and re- donors. This strategy may be particularly ap- cipient populations suggests additional studies plicable when the only “official” donor pop- on this aspect of red-cockaded woodpecker bi- ulations (see U.S. Fish and Wildlife Service ology should be increased. 2003) are far from the recipient population Although translocation success increased and economic, research, and logistical oppor- substantially from the initial efforts in the tunities for intensive monitoring are in place late 1980s/early 1990s to the late 1990s/early at these small populations. The South/Central 2000s, additional research is needed to im- Florida Recovery Unit may provide such an prove methods. Specifically, more information opportunity. It consists of many (about 16), on time of year of release, release method, small (3–50 occupied clusters) populations. number of pairs simultaneously released, ef- Most are intensively monitored, i.e., totally fects of southern flying squirrels, and dif- marked, and several have long-term research ferent regional translocation strategies would programs established. This level of knowledge be valuable. Additionally, more research is may make it ecologically possible and logisti- needed on effects of subadult removal on cally practical to move carefully selected birds donor populations. The limited research to between these populations and among poten- date on this issue has suffered from small sam- tial metapopulations. Assessing the potential ple size and inadequate study design. Data on donor population level impact would be a crit- this subject have been gathered from moni- ical and required component of any program toring programs rather than well-designed re- involving removing birds from small popula- search projects. tions. Several approaches may increase the an- There are numerous, small red-cockaded nual supply of birds available for transloca- woodpecker populations across the Southeast tion. However, fully implementing these op- well below their potential population size and tions is challenging, given budget constraints, yet not participating in translocation programs. logistical challenges, and unknown potential Reasons for nonparticipation vary but include impacts to donor populations. As an exam- inadequate funding to monitor their popula- ple, increasing output of birds from donor tion and prepare habitat, unqualified staff, or populations would likely require establish- lack of commitment. Overcoming these barri- ment of recipient-driven compensation sys- ers is challenging and ultimately requires lead- tems for donors, or sharing resources to min- ers willing to secure funding, hire and train imize economic burden on donors. The argu- staff, and identify red-cockaded woodpecker ment could be made that resources used by recovery as a management priority. It is hoped donors to help recipients could be used to in- that as more donor populations become avail- crease the rate of growth of donors’ own pop- able, current recipient populations leave the ulation via habitat improvement initiatives. translocation program (by reaching the thresh- Ironically, as donor populations increase in old of 30 potential breeding groups), and pub- size they could provide additional subadult lic awareness increases regarding longleaf pine birds for translocation if managers could in- ecosystem restoration, more small populations crease their nesting season monitoring. In will participate in translocations. 362 IV. Restoration Other Potential Species for are listed as threatened, and five have critically small populations requiring the growth of ex- Reintroduction and isting populations or the establishment of addi- Translocation in Longleaf tional populations. Disease appears to be a major problem for several species; therefore, any Pine Forests risk of disease via translocation efforts must be minimized (Cunningham 1996; Trenham and Will Translocations Work for Marsh 2002; L. LaClaire, U.S. Fish and Wildlife Other Longleaf Pine Ecosystem Service, personal communication 2003). Mon- Fauna? itoring efforts must be conducted to evaluate success and potential problems with transloca- Understanding, and then designing the tions. translocation program for red-cockaded woodpeckers to capitalize on the life history Species Accounts and ecological behaviors of the species has Below we provide species accounts for 10 been largely responsible for the documented vertebrates listed in Table 6 that have, at successes. Before considering translocation some level, been involved in translocation pro- programs for other longleaf pine fauna, biol- grams. Although the remaining six species ogists must ask several important questions. (not including the red-cockaded woodpecker) Does the species exhibit any life history traits may have potential for translocation programs, that make it “naturally” conducive to translo- available literature on translocation for these cations, like dispersal behavior? Conversely, species is inadequate to provide meaningful does its ecology preclude likely success if species accounts. translocated, e.g., a species with a strong homing instinct? Are there age- or status- Amphibians specific (e.g., reproductive or social structure) characteristics that enhance or detract from Two listed amphibian species are dependent the probable success of translocating specific on longleaf or slash pine forests that surround individuals? What is the species’ annual re- isolated, ephemeral wetlands. The Mississippi productive potential, that is, are few offspring gopher frog (Rana capito sevosa) is listed as or multiple broods common? Answers to these endangered in Mississippi (66 Federal Register and other questions directly related to the 62993, 2001) and the flatwoods salamander species’ ecology, as it relates to being receptive (Ambystoma cingulatum) is listed as threatened to translocations, are critical. These questions in Florida, Georgia, and South Carolina. Sev- must be addressed and evaluated to not only eral populations of flatwoods salamander re- increase success rates for appropriate species, cently have been documented in the above but also to identify those fauna that are poor states (64 Federal Register 15691, 1999); how- candidates for translocation or are candidates ever, only one population of Mississippi go- for other kinds of enhancement efforts that pher frogs is known to exist. Primary habi- do not involve translocation. tat for adults of both species includes upland Numerous other rare vertebrate fauna, be- sandy areas of longleaf or slash pine forest sides red-cockaded woodpeckers, are poten- surrounding ephemeral wetlands. Small adult tial candidates for reintroduction into longleaf population size, vulnerability to environmen- pine forests (Table 1). Most have ongoing re- tal changes or conditions, and loss of breeding covery efforts including some form of rein- sites threaten the flatwoods salamander and troduction method (Table 6). Table 6 includes Mississippi gopher frog. five species each of mammals and reptiles, Both species breed in flooded, ephemeral four bird species (not including red-cockaded grassy ponds or cypress domes, and other iso- woodpeckers), and two amphibian species. lated wetlands that lack predatory fish popula- Five species are listed as endangered, seven tions (Mount 1975). The grassy nature of the 11. Reintroduction of Fauna 363

TABLE 6. Population trend and reintroduction methods for listed and rare vertebrates found in longleaf pine forests. Common name Population Trend Method of increase Amphibians Mississippi gopher frog 1 population in southern Stable, vulnerable Translocation MS Flatwoods salamander 51 populations with 71% in Stable, vulnerable Translocation? Florida; 122 breeding sites. Reptiles Eastern Indigo snake Mostly FL and south GA; Declining, particularly Translocation possible in AL, MS, and on private lands questionable SC Gopher tortoise Southwestern AL, Probably declining Translocation northeastern panhandle of LA, and southeastern MS Blue-tailed mole skink Limited to southern portion Probably declining Unknown of Lake Wales Ridge in FL Sand skink Ocala National Forest and Probably declining Translocation? southern portion of Lake Wales Ridge in FL Louisiana pine snake Small population in western Unknown Some releases, LA and eastern TX program on hold Birds Mississippi sandhill crane 100 adults Stable Captive release Red-cockaded woodpecker 5800 groups Increasing throughout Translocation/ their range artificial cavities Southern bald eagle ∼1000 territories in FL Increasing throughout Translocation/ their range artificial nest sites Southeastern American kestrel Southern portion of LA, Increasing on Nest boxes MS, AL, GA, and SC; protected lands throughout FL highlands Bachman’s sparrow Common in appropriate Declining due to No information habitat habitat loss Mammals Red wolf 1 small population in Stable, vulnerable Captive breeding eastern NC and release Florida panther 1 population with 2 Declining/stable, Captive breeding subpopulations; 30–50 vulnerable and release, adults possibly Florida black bear 4 primary populations Declining/stable Unknown Louisiana black bear 40 to 60 adults Declining Unknown Sherman’s fox squirrel Widely distributed in FL Declining, habitat loss Translocation? see and degradation Delmarva fox squirrel literature entire pond or surrounding grassy transitions Mississippi gopher frogs is on a topographic of wetlands are particularly important to sur- high within a small drainage basin. Nearby vival of tadpoles or larvae. Loss of this micro- wells are used to maintain appropriate water habitat has occurred through detrimental tim- levels during years of low rainfall (L. LaClaire, ber practices, other development activities, and U.S. Fish and Wildlife Service, personal com- anthropogenic changes to drainage patterns. munication 2003). Ponds used by flatwoods The single remaining breeding pond used by salamanders are usually down slope from 364 IV. Restoration surrounding uplands and may be seasonally alteration, predation by humans, road mortal- connected to other wetlands. Hollow stumps ity, and isolation of mature animals from one and occupied and abandoned gopher tortoise another (Lohoefener and Lohmeier 1981; U.S. burrows are underground refugia important to Fish and Wildlife Service 1990a). In Florida, adult Mississippi gopher frogs, which histori- translocation of gopher tortoises is often con- cally occurred in areas with gopher tortoises. ducted as part of development mitigation. The remaining Mississippi gopher frog popula- However, there are significant problems with tion occurs in an area where gopher tortoises this process related to the social system of the have been extirpated, although efforts are un- animal and the potential to spread upper res- derway to reestablish a resident gopher tortoise piratory tract disease, as well as other diseases, population (L. LaClaire, U.S. Fish and Wildlife to uninfected populations (Cox et al. 1994, Service, personal communication 2003). J. Barish, Florida Fish and Wildlife Conser- Translocations of these species have not been vation Commission, personal communication conducted and using translocations as a means 2003). Currently, all donor populations must of enhancing or reestablishing amphibians is be screened for disease prior to any movement controversial (Marsh and Trenham 2001, Sei- of individuals. gal and Dodd 2002). Dodd and Seigal (1991) The sand skink (Neoseps reynoldsi) is listed as conducted a through review of the benefits threatened by the U.S. Fish and Wildlife Ser- and costs of amphibian and reptile relocations, vice and inhabits sandy ridges with underly- repatriations, and translocations. The possi- ing moist soil (Christman 1992). This species bility of disease transmission appears to be may be particularly abundant in ecotonal ar- the most serious potential negative aspect of eas between pine savannas and oak scrub. It translocation (Cunningham 1996; Daszak et al. also occurs in rosemary scrub, turkey oak bar- 1999; L. LaClaire, U.S. Fish and Wildlife Ser- rens, and sandy areas of high pine. Within their vice, personal communication 2003). For ex- known distribution all suitable habitats were ample, Seigal and Dodd (2002) point out that probably inhabited (P. Molar, Florida Fish and one potential site for translocation of Missis- Wildlife Conservation Commission, personal sippi gopher frog tadpoles or eggs contained communication 2003). The preferred habitat an undescribed fungus that decimated all ranid within its range is extremely limited. Areas tadpoles within a matter of weeks. Addition- free of abundant roots, with scattered shrubby ally, there appear to be several diseases to vegetation and patches of bare sand seem to which frogs are susceptible, particularly un- be preferred habitats; however, surveys have der stressful conditions (L. LaClaire, U.S. Fish found them in a few areas with an extensive and Wildlife Service, personal communication tree canopy. Habitat loss, both historic and cur- 2003). rent, is the major threat to this species. Much of its habitat along Florida’s Lake Wales Ridge Reptiles has been converted to citrus groves and human development. Within apparently contin- The gopher tortoise is a resident of high sandy uous habitat, small drainages likely result in areas and dry microhabitats within wet pine genetic segregation and this genetic structur- flatwoods distributed throughout the Lower ing has implications for conservation of all Coastal Plain and Sandhills. The species is scrub-associated lizards (Branch et al. 2003, listed as federally threatened in the western- P. Molar, Florida Fish and Wildlife Conser- most portion of its range, including south- vation Commission, personal communication western Alabama, the northeastern panhan- 2003). Any strategy for population enhance- dle of Louisiana, and southeastern Mississippi ment by means of translocation or relocation (U.S. Fish and Wildlife Service 1990a). The should consider genetic implications and re- listed population occurs within the historic quire a comprehensive plan to conserve ge- range of longleaf pine. Threats include habitat netic diversity (Branch et al. 2003). 11. Reintroduction of Fauna 365

To date, the only translocation effort for Birds sand skinks involved complete removal and transportation of the top 6 inches of soil from There are three nonmigratory species of sand- the project area along with the resident sand hill cranes in the Southeast. The Mississippi skink population to an abandoned citrus grove. population (Grus canadensis pulla) is federally The recipient site was believed to be suitable listed as endangered, occurring only on the and occupied by sand skinks prior to its con- Mississippi Sandhill Crane National Wildlife version to citrus. The skinks have persisted, Refuge in southeastern Mississippi (U.S. Fish but apparently are not breeding (P. Molar, and Wildlife Service 1991). This population Florida Fish and Wildlife Conservation Com- was greatly reduced by hunting prior to the mission, personal communication 2003). If ad- twentieth century. Currently the population ditional relocations are attempted, recipient is limited by suitable habitat, primarily lon- sites must include appropriate edaphic condi- gleaf pine savannahs. Lower levels of genetic tions and be restored to native conditions. Dis- heterozygosity may result in poor hatching eases do not appear to be a problem for this success and some debility in captive chicks species. (U.S. Fish and Wildlife Service 1991). Over the The eastern indigo snake (Drymarchon corais) past several years the population has remained occurs in southeast Georgia and peninsular relatively stable at approximately 100 adults, Florida with disjunct populations in the Florida with slightly more than 20 nests during 2003. Panhandle and perhaps Alabama (Conant Although apparently stable, predation of nests 1975). The species inhabits a variety of habitats from a variety of mammals is problematic (L. in southern Florida; however, in the northern LaClaire, U.S. Fish and Wildlife Service, per- parts of its range it is more restricted to sand sonal communication 2003). Coyotes, foxes, ridge habitats of pine interspersed with wet- and bobcats are the most signiﬁcant predators lands (Speake et al. 1978). Fire exclusion, habi- and are controlled through trapping programs. tat loss, and extensive (historical) trapping for Coupled with habitat restoration, an ongoing the pet trade are responsible for their decline. program of releasing captive-bred cranes helps In northern parts of their range, loss of gopher stabilize the single extant population. tortoises and their burrows also are limiting The southeastern American kestrel (Falco factors (Speake and Mount 1973). Males are sparverius paulus) is resident in the lower south- territorial and cannibalistic and may cover up eastern Coastal Plain and primarily uses open to 160 ha (P. Molar, Florida Fish and Wildlife habitats such as pastures, longleaf pine/turkey Conservation Commission, personal commu- oak sandhills, and grasslands (Johnsgard 1990; nication 2003). To date, the only translocation Stys 1993). The largest population occurs in efforts included releases of snakes by Speake, Florida and is associated with upland ridges in from a variety of sources, on St. Marys and the northern two-thirds of the peninsula (Stys St. Vincent Islands in Georgia and southern 1993). Loss of habitat from logging or ﬁre sup- Alabama (P. Molar, Florida Fish and Wildlife pression appears to be the primary cause of Conservation Commission, personal commu- the species’ decline. Additionally, the loss of nication 2003). Feral hogs occurred in all re- red-cockaded woodpeckers and their cavities lease sites and apparently killed most snakes may have contributed to declines in southeast- during high water conditions. Regular moni- ern American kestrels. Nest boxes installed in toring was apparently not conducted. If rein- formerly occupied habitat, where adjacent ex- troduction programs are implemented, release tant populations still exist, have been success- sites should contain 4000 to 4800 ha of good ful in reestablishing breeding populations. For habitat and be free of feral and domestic preda- example, at Fort Jackson in South Carolina in- tors (P.Molar, Florida Fish and Wildlife Conser- stallation of nest boxes over the past 8 years vation Commission, personal communication has resulted in a large breeding population (T. 2003). Marston, U.S. Army, personal communication 366 IV. Restoration

2003). Nest boxes are repaired as needed, nesting attempts in a southern Ontario popu- breeding is monitored, and nestlings are lation (Hunter et al. 1997). Because of the DDT banded. Over the past 6 years an average of ban and, to a lesser extent, reintroduction pro- 17 nests produced an average clutch size of 4.7 grams, populations are growing in most areas eggs with about 34% of eggs failing. On aver- of the Southeast and the species has been pro- age, these nests produced about 2.9 nestlings posed for delisting by the U.S. Fish and Wildlife at time of nestling banding (T. Marston, U.S. Service. Army, personal communication 2003). Simi- lar success has occurred in other areas of the Mammals species range, for example, Fort Gordon, GA. The southern bald eagle (Haliaeetus leuco- Two large, predatory, federally endangered cephalus leucocephalus) is federally listed as mammals with similar ecological and social endangered (Curnutt 1996). Historically dis- constraints to their reintroduction are the tributed throughout North America, pesti- Florida panther (Felis concolor coryi) and red cide poisoning and habitat loss eliminated the wolf (Canis rufus). Historically, these mam- species from most of its former range. Banning mals occurred over much of the southeast- DDT in 1972 resulted in a dramatic increase in ern United States and ranged over a variety Florida populations while the 10 other south- of habitats including longleaf pine forests (An- eastern states only had 187 breeding pairs in derson 1983; U.S. Fish and Wildlife Service 1991 (U.S. Fish and Wildlife Service 1987; 1990b). Translocation of red wolves and the Jenkins and Sherrod 1993). Bald eagles occupy Florida panthers’ close relative, mountain li- a variety of habitats, although proximity to ons (Felis concolor stanleyana) have resulted in large bodies of open water with large trees (es- individuals surviving and breeding in the wild pecially pine) for nest sites, are important re- (U.S. Fish and Wildlife Service 1990b; Belden quirements. Nests are usually in live, tall trees and McCown 1996). within several miles of a large lake or river. For red wolves, interbreeding with coyotes Florida currently has one of the largest ea- was a major problem in the last remaining gle populations in the lower 48 states and it population and potentially a problem for any continues to expand. Eggs from Florida pop- reintroduction efforts. During the 1980s, the ulations have been used to reintroduce eagles remaining red wolves persisted in low num- into several other states (Wood 1982). Typi- bers in a small area of southeastern Texas and cally, eggs were collected after 2 weeks of natu- southwestern Louisiana. From this area, over ral incubation, transported to the Sutton Avian 400 canids were captured and screened to ob- Research Center in Oklahoma, incubated by tain 43 red wolf-types for the captive breeding Bantam hens (Gallus sp.) or artificially (Nes- program. Of these, only 15 adults survived and bitt et al. 1998), and young released at 11–12 produced offspring that fit the red wolf-type for weeks of age at recipient sites (Sherrod et al. reintroduction. 1987; Nesbitt et al. 1998). Florida populations Reintroduction of red wolves has been con- remained stable or increased during the period sidered or attempted in the Land between of egg removal. the Lakes (LBL) region in western Tennessee An alternate strategy for population en- and Kentucky and Cades Cove (Great Smoky hancement or establishment includes the con- Mountains National Park) in eastern Ten- struction of artificial nesting platforms to en- nessee. The LBL project failed prior to the re- courage the return of eagles to former breed- lease of the first animal because of public oppo- ing territories (Hunter et al. 1997). The nest sition (U.S. Fish and Wildlife Service 1990b). platforms complete with nest material are con- Captive-bred wolves were successfully intro- structed in tops of well-formed trees simi- duced at Cades Cove and survived and bred in lar in structure to those normally used by the wild. However, because of poor pup sur- eagles. From 1994 to 1998, three successful vival, inability of adults to establish territories, artificial nests accounted for 15.4% of the and low availability of prey, the wolves were 11. Reintroduction of Fauna 367 removed from the Park. Since these initial the captive-raised animals (particularly the attempts, red wolves have been successfully males) were involved in a number of interac- reintroduced to the remote Alligator River Na- tions with humans and resulted in bad public- tional Wildlife Refuge in eastern North Car- ity (Belden and McCown 1996). Based on this olina (Parker 1987; Phillips 1990). Releases study, Belden and Hagedorn (1993) concluded usually included family units of adults and off- that reintroductions of a panther population spring. Mortality of adult females was high; in northern Florida or elsewhere were bio- however, the population has survived and is logically feasible. However, issues of funding, producing young. gaining public support, and identifying suitable In the late 1990s, it was estimated that 30 landscape-scale areas need to be resolved be- to 50 adult Florida panthers remained in the fore such reintroductions can be attempted. Big Cypress National Preserve and Everglades National Park physiographic regions in south Florida (Belden and McCown 1996). In 2003, Planning for Future 87 adult Florida panthers were known to exist. For recovery it is proposed that this popu- Reintroduction of Longleaf lation be managed and that panthers be rein- Pine Ecosystem Fauna troduced to three additional areas (Belden and Hagedorn 1993). To that end, and as a feasi- Based on our literature review, case study bility study, 19 mountain lions were released of red-cockaded woodpeckers, and what we as Florida panther surrogates into northern have learned about reintroduction of other Florida. These releases included 11 females longleaf pine ecosystem fauna, we have devel- and 8 vasectomized males. These lions, includ- oped a translocation success ranked response ing captive-raised and wild-caught animals, model (see Table 7). Prior to discussing its ap- established 15 home ranges. Unfortunately, plication, one observation about the model

TABLE 7. Translocation success ranked response model for longleaf pine ecosystem potential reintroduction fauna candidates. Probability of success Factors potentially influencing translocation success High Medium Low Ownership of recipient property Federal State Private Public support Substantial Likely Questionable Availability of funding Good Fair Poor Cost per animal translocated <$1500 $1500–$5000 $5000+ Logistical challenges Few Manageable Numerous Source of donor stock Wild (hard release) Wild (soft release) Captive bred Number of animals available Many Some Few annually Number of years required to 5 5 to 10 10+ reach population goal Spatial scale (number of acres <100 100 to 1000 1000 to 10,000+ required per territory) Species status Federally listed State listed/federal candidate Rare/sensitive/special concern Species ecological classification Generalist Adaptable Specialist Number of species unique Few (1 or 2) Manageable (3 to 4) Numerous (5 or more) habitat requirements Number of significant, Few (1) Manageable (2 to 3) Numerous (4 or more) non-habitat-related, threats Target population size (number Small (20 to 50) Medium (50 to 200) Large (200+) of breeding adults) Species reproductive potential High Moderate Low Species dietary habit Omnivorous Herbivorous Carnivorous 368 IV. Restoration is noteworthy. The ranking factors fall in cussed and evaluated in the context of “prob- various categories, including administrative, ability of program implementation.” For ex- economic, social/cultural (i.e., human dimen- ample, implementation will be more difficult sions), ecological, and management-oriented. without public support and adequate amounts This is significant because it highlights the of funding and donor stock. Even if the above need to approach each potential transloca- factors are adequately addressed, implementation program with a multidisciplinary coalition will still be challenging if many unique tion of partners. The partners, individually habitat requirements must be satisfied and and collectively, must not only gain knowl- numerous threats, e.g., predators and com- edge about each factor, but also develop an petitors, must be managed. Additionally, if a adequate understanding of the interrelation- species has low reproductive potential, specific ships among factors to assess, design, and ul- dietary requirements, or requires a vast terri- timately implement a translocation program. tory, implementation will also be very chal- Typically, partners may include governmental lenging. Therefore, from an implementation natural resource agencies (federal and state), perspective, the model can be used to help nongovernmental conservation organizations, quantify and qualify both the annual and long- private industry, private landowners, and uni- term probability that the program will be suc- versities. cessful. With this knowledge, administrators The model has multiple applications. First, and managers can project and plan for fu- it can be used to highlight areas of program ture needs, while simultaneously calculating vulnerability by identifying those factors for a expected program benchmarks given start-up particular species with a “low” ranking. While and probable short-term needs for each appro- nothing can be done for some factors with a priate factor. low ranking, for example species is an ecolog- In addition to the above applications, the ical specialist or has low reproductive poten- model has also been designed to assist man- tial, other factors can often be addressed. For agers, biologists, and administrators predict the example, if the reintroduced species will be ex- overall probability of success of translocation posed to numerous threats from competitors, programs for vertebrate, terrestrial fauna of resources and management emphasis can be longleaf pine ecosystems. The model contains directed toward mitigating this factor, thereby 16 factors that we consider critical for assess- minimizing its impact and reducing the overall ing probable success of translocations. Initially, vulnerability of the project’s success. In other ranking each factor as high, medium, or low words, the model can be used to focus limited could assess the probability of a successful resources on management of key limiting fac- translocation program. Determination of suc- tors. cess probability after this initial analysis may Second, the model can also direct biolo- be possible if a clear majority (e.g., 75%; 12 or gists and researchers to those factors for which more) of factors fall within one rank. However, more information is required about their tar- using this system and lacking a clear major- get species to make informed decisions. By do- ity, it would be necessary to “score” or “inter- ing so, research, monitoring, and management pret” the results to assign an overall probability programs, capable of filling in the blanks in rank to the program. For example, if six factors the model, can be designed and implemented. were assigned “high,”, six factors “medium,” Again, this helps focus limited resources on and four factors “low,” one could logically con- those factors that may affect program success. clude at least a medium probability of success Probable success of any fauna translocation for the project. A scoring system provides chal- program will ultimately be directly related to lenges when no clear pattern develops, there- the ability of managers and biologists to im- fore, a factor weighting system may be neces- plement actions necessary to reach the target sary to help interpret results. Once weighted, population size. Therefore, many factors listed some factors for any number of reasons, e.g., in the ranked response model could also be dis- economic, administrative, social, or ecological, 11. Reintroduction of Fauna 369 become more important to program success monitoring, ability, via technology and habitat than other factors. Under this scenario, a mi- management programs, to restore and main- nority of weighted factors in one rank (e.g., tain the forest’s structure and composition, and four in “low”) may “outweigh” the influence public support. Many of these issues are inter- of the ranks with more variables listed, e.g., related and must be addressed in a compre- five in “medium” and seven in “high.” Ulti- hensive fashion. Additionally, some programs mately, the inability to satisfy an individual, such as prescribed burning, while essential for but critical, factor, e.g., availability of funding long-term management of most longleaf pine or number of animals available annually, could ecosystem species, are not necessarily con- make program implementation impossible. trolled or administered by the fauna reintro- Although we considered weighting each fac- duction team. Therefore, reintroduction efforts tor, at this time we do not believe adequate also require close coordination and coopera- information exists to do so with reasonable active partnerships with habitat managers and curacy for the entire suite of potential longleaf others to ensure success. pine fauna translocation candidates (see Table We believe significant potential exists to 1). That is, the multiple and complex rela- further the cause of longleaf pine ecosystem tionships that exist between and among these restoration via establishing new and augment- various factors, in the context of all longleaf ing existing small, at-risk populations of listed pine fauna as one group, are not well known and rare vertebrate fauna. We encourage re- or understood. However, at the single species sponsible officials, knowledgeable scientists, scale, knowledgeable individuals may be able and interested conservationists to explore the to weight factors for which adequate informa- possibilities, identify opportunities, and imple- tion exists. The red-cockaded woodpecker may ment well-designed programs. Each success provide such an opportunity. will be an important contribution to our legacy We encourage managers to use and build on of restoring the biodiversity of the endangered our model but we caution that it has limita- longleaf pine ecosystem. tions related to its generic coverage. Simulta- neously, we encourage biologists who develop species-specific models to consider a weight- Acknowledgments ing system. In conclusion, we recommend that this model be used, in conjunction with other We appreciate the editorial reviews received appropriate decision-making tools, and likely from Dr. George Tanner, University of Florida, more species-specific variables, to guide offi- and an anonymous reviewer; their comments cials responsible for listed and rare species con- and suggestions improved our chapter consid- servation. erably. A special thank you to Dr. Jeff Walters, Virginia Polytechnic Institute, for his detailed review of an earlier draft of our manuscript. Conclusion His insights and knowledge regarding red- cockaded woodpeckers focused his review and Conservation of biodiversity in longleaf pine comments on issues most relevant to a thor- forests via fauna translocations will be chal- ough understanding of our topic. Address- lenging. Once life history and ecology of each ing his thoughtful comments significantly im- species are well understood and it has been de- proved our chapter. We also thank Ms. Nickie termined to be a viable candidate for translo- Nichols, U.S. Fish and Wildlife Service stu- cation, complex operational issues must still be dent intern and Clemson University graduate considered. Program implementation-related student, working on red-cockaded woodpeck- issues include: availability of donor stock, ers, for formatting and constructing our ta- landscapes large enough to accommodate vi- bles. Thanks to Ms. Nicole Edwards, U.S. Fish able populations, adequate funding to initiate and Wildlife Service computer assistant, Clem- and sustain the program, including postrelease son University, for assembling our literature 370 IV. Restoration cited section. Finally, we thank our editors, red-cockaded woodpecker population at Fort Dr. Shibu Jose, Dr. Eric Jokela, and Dr. Debbie Stewart, Georgia. In Red-cockaded Woodpecker: Road Miller, for inviting us to prepare this chapter to Recovery, eds. R. Costa and S. J. Daniels. Blaine, and, thereby, contribute to this much-needed WA: Hancock House Publishers. In Press. text on the longleaf pine ecosystem. Carrie, N.R., Moore, K.R., Stephens, S.A., and Keith, E.L. 1996. Long-distance homing of a translocated red-cockaded woodpecker. Wildl Soc References Bull 24:607–609. Carrie, N.R., Conner, R.N., Rudolph, D.C., and Car- Abrahamson, W.G., and Hartnett, D.C. 1990. Pine rie, D.K. 1999. Reintroduction and postrelease flatwoods and dry prairies. In Ecosystems of Florida, movements of red-cockaded woodpecker groups eds. R.L. Myers and J. J. Ewel, pp. 103–149. Or- in eastern Texas. J Wildl Manage 63:824–832. lando: University of Central Florida Press. Christman, S.P. 1992. Threatened: sand skink, Allen, D.H. 1991. An insert technique for con- Neoseps reynoldsi (Stejneger). In Rare and Endan- structing artificial red-cockaded woodpecker cav- gered Biota of Florida, ed. P.E. Moler. Gainesville: ities. U.S. Forest Service General Technical Report University Press of Florida. SE-73. Conant, R. 1975. A Field Guide to Reptiles and Amphib- Allen, D.H., Franzreb, K.E., and Escano, R.F. ians of Eastern and Central North America. Boston: 1993. Efficacy of translocation strategies for red- Houghton Mifflin Co. cockaded woodpeckers. Wildl Soc Bull 21:155– Conner, R.N., and Rudolph, D.C. 1991. Forest habi- 159. tat loss, fragmentation, and red-cockaded wood- Anderson, A.E. 1983. A critical review of litera- pecker populations. Wilson Bull 103:446–457. ture on puma (Felis concolor). Colorado Division Conner, R.N., Rudolph, D.C., Saenz, D., and Schae- of Wildlife Special Report 54. fer, R.R. 1996. Red-cockaded woodpecker nest- Barclay, J.H., and Cade, T.J. 1983. Restoration of ing success, forest structure, and southern flying the Peregrine Falcon in the eastern United States. squirrels in Texas. Wilson Bull 108:697–711. In Bird Conservation, ed. S.A. Temple, pp. 3–40. Conner, R.N., Rudolph, D.C., Schaefer, R.R., and Madison: University of Wisconsin Press. Saenz, D. 1997. Long-distance dispersal of red- Belden, R. C., and Hagedorn, B. W. 1993. Feasibility cockaded woodpeckers in Texas. Wilson Bull of translocating panthers into northern Florida. J 109:157–160. Wildl Manage 57:388–397. Copeyon, C.K. 1990. A technique for constructing Belden, R.C., and McCown, J.W. 1996. Florida cavities for the red-cockaded woodpecker. Wildl panther reintroduction feasibility study. Florida Soc Bull 18:303–311. Game and Fresh Water Fish Commission, Bureau Costa, R. 2001. Red-cockaded woodpecker. In of Wildlife Resources. Study No. 7507, Final Re- Wildlife of Southern Forests: Habitat and Management, port. Tallahassee, FL. ed. J.G. Dickson, pp. 309–321. Blaine, WA: Han- Branch, L.C., Clark, A.M., Moler, P.E., and Bowen, cock House Publishers. B.W. 2003. Fragmented landscapes, habitat speci- Costa, R., and Daniels, S.J., eds. 2004. Red-cockaded ficity, and conservation genetics of the three Woodpecker: Road to Recovery. Blaine, WA: Hancock lizards in Florida scrub. Conserv Genet 4:199–212. House Publishers. In Press. Britcher, J.J., and Patten, J.M. 2004. Red-cockaded Costa, R., and Edwards, J.W.1997. Cooperative con- woodpecker management on Fort Bragg: Then servation agreements for managing red-cockaded and now. In Red-cockaded Woodpecker: Road to Re- woodpeckers on industrial forest lands: What are covery, eds. R. Costa and S. J. Daniels. Blaine, WA: the motivations? In Proceedings of the Symposium Hancock House Publishers. In press. on the Economics of Wildlife Resources on Private Brown, C., and Simpkins, S. 2004. The Chicka- Lands, ed. R. Johnson, pp. 111–124. Auburn, AL: sawhay Ranger District story: Saving a small pop- Auburn University. ulation from extirpation. In Red-cockaded Wood- Costa, R., and Escano, R. 1989. Red-cockaded pecker: Road to Recovery, eds. R. Costa and S. J. woodpecker: Status and management in the Daniels. Blaine, WA: Hancock House Publishers. southern region in 1986. U.S. Forest Service Tech- In press. nical Publication R8-TP12. Carlile, L.D., Ten Brink, C., Mitchell, C.L.R., Costa, R., and Jordan, N. 2003. Biological opinion Puder, S.E., Spadgenske, E. W., and Beaty, T. on all section 10(a)(1)(A) management, moni- A. 2004. An intensively managed and increasing toring and research permits issued to all private, 11. Reintroduction of Fauna 371

state and federal agencies and individuals in- DeLotelle, R.S., Epting, R.J., and Demuth, G. 1995. volved with management, conservation and re- A 12-year study of red-cockaded woodpeckers covery of the red-cockaded woodpecker through- in central Florida. In Red-cockaded Woodpecker: Re- out the range of the species. U.S. Fish and Wildlife covery, Ecology and Management, eds. D.L. Kul- Service, Clemson, SC. havy, R.G. Hooper, and R. Costa, pp. 259–269. Costa, R., and Kennedy, E. 1994. Red-cockaded Center for Applied Studies in Forestry. College woodpecker translocations 1989–1994: State-of- of Forestry, Stephen F. Austin State University, our-knowledge. Annual Proceedings of the American Nacogdoches, TX. Zoo and Aquarium Association, pp. 74–81. Atlanta, DeLotelle, R.S., Epting, R.J., Leonard, D.L., Jr., and GA: Zoo Atlanta. Costa, R. 2004. Management strategies for recov- Cox, J., Kautz, R., MacLaughlin, M., and Gilbert, T. ery of red-cockaded woodpeckers populations: A 1994. Closing the Gaps in Florida’s Wildlife Habi- metapopulation proposal. In Red-cockaded Wood- tat Conservation System. Florida Game and Fresh pecker: Road to Recovery, eds. R. Costa and S.J. Water Fish Commission. Tallahassee, FL. Daniels. Blaine, WA: Hancock House Publishers. Crowder, L.B., Priddy, J.A., and Walters, J.R. 1998. In Press. Demographic isolation of red-cockaded wood- Dodd, C.K., Jr., and Seigal, R.A. 1991. Relocation, pecker groups: A model analysis. Project Final Re- repatriation, and translocation of Amphibians: port, prepared for U.S. Fish and Wildlife Service. Proven Management Method or Experimental Cunningham, A.D. 1996. Disease risks of wildlife Technique? 16:552–554. translocation. Conserv Biol 10:349–353. Doresky, J., Barron, M., and Swiderek, P. 2004. Curnutt, J.L. 1996. Southern bald eagle. In Rare and Landscape scale restoration and red-cockaded Endangered Biota of Florida, Volume IV. Birds, eds. woodpecker recovery? In Red-cockaded Wood- J.A. Rodgers, Jr., H.W. Kale II, and A. T. Smith, pecker: Road to Recovery, eds. R. Costa and S.J. pp. 179–187. Orlando: University Press of Florida. Daniels. Blaine, WA: Hancock House Publishers. Daniels, S.J., and Walters, J.R. 2000a. Inbreeding In press. depression and its effects on natal dispersal in red- Drumm, R.L., Boyd, K., and Camp, S.N. 2004. cockaded woodpeckers. Condor 102:482–491. Red-cockaded woodpecker management on Fort Daniels, S.J., and Walters, J.R. 2000b. Between- Gordon, Georgia. In Red-cockaded Woodpecker: year breeding dispersal in red-cockaded wood- Road to Recovery, eds. R. Costa and S.J. Daniels. peckers: Multiple causes and estimated cost. Ecol- Blaine, WA: Hancock House Publishers. In ogy 81:2473–2484. Press. Daniels, S.J., Priddy, J.A, and Walters, J.R. 2000. Edwards, J.W., and Costa, R. 2004. Rangewide suc- Inbreeding in small populations of red-cockaded cess of red-cockaded woodpecker translocations. woodpeckers: Insights from a spatially explicit In Red-cockaded Woodpecker: Road to Recovery, eds. individual-based model. In Genetics, Demography R. Costa and S.J. Daniels. Blaine, WA: Hancock and Viability of Fragmented Populations, eds. A.G. House Publishers. In Press. Young and G.M. Clarke, pp. 129–147. London: Edwards, J.W., Dachelet, C.A., and Smathers, W.M. Cambridge University Press. 1999. A mobile aviary to enhance translocation Daszak, P., Berger, L., Cunningham, A.A., Hyatt, success of red-cockaded woodpeckers. In Fish and A.D., Green, D.E., and Speare, R. 1999. Emerg- Wildlife Research and Management: Applying Emerg- ing Infectious diseases and amphibian population ing Technologies, pp. 48–53. Proceedings of the declines. Emerg Infect Dis 5:735–748. 37th Annual Meeting of the Canadian Society DeFazio, J.T., Hunnicutt, M.A., Lennartz, M.R., of Environmental Biologists. Edmonton, Alberta, Chapman, G.L., and Jackson, J.A. 1987. Red- Canada. cockaded woodpecker translocation experiments Edwards, J.W., Mari, Y., and Smathers, W. 2004. in South Carolina. Proc Southeast Assoc Fish Wildl Evaluation of a mobile aviary to enhance translo- Agenc 41:311–317. cation success of red-cockaded woodpeckers. In DeLotelle, R.S., and Epting, R.J. 1992. Reproduc- Red-cockaded Woodpecker: Road to Recovery, eds. R. tion of the red-cockaded woodpecker in central Costa and S.J. Daniels. Blaine, WA: Hancock Florida. Wilson Bull 104:285–294. House Publishers. In Press. DeLotelle, R.S., Epting, R.J., and Newman, J.R. Ellis, D.H., Dobrott, S.J., and Goodwin, J.L., Jr. 1987. Habitat use and territory characteristics 1978. Reintroduction techniques for masked bob- of red-cockaded woodpeckers in central Florida. white. In Endangered Birds: Management Techniques Wilson Bull 99:202–217. for Preserving Threatened Species, ed. S.A. Temple, 372 IV. Restoration

pp. 345–354. Madison: University of Wisconsin Daniels. Blaine, WA: Hancock House Publishers. Press. In press. Engstrom, T., Crawford, R.I., and Baker, W.W. 1984. Haig, S.M., and Nordstrom, L.H. 1991. Genetic man- Breeding bird populations in relation to chang- agement of small populations. In Westview Special ing forest structure following fire exclusion: A 15- Studies in Natural Resources and Energy Management: year study. Wilson Bull 96:437–450. Challenges in the Conservation of Biological Resources: Ferral, P., Edwards, J.W., and Armstrong, A.E. 1997. A Practitioner’s Guide, pp. 119–138. Boulder, CO: Long-distance dispersal in red-cockaded wood- Westview Press. peckers. Wilson Bull 109:277–284. Haig, S.M., Belthoff, J.R., and Allen, D.H. 1993. Pop- Franzreb, K.E. 1997a. A mobile aviary design to al- ulation viability analysis for a small population of low the soft release of cavity nesting birds. U.S. red-cockaded woodpeckers and an evaluation of Forest Service Research Note SRS-5. enhancement stragegies. Conserv Biol 7:289–301. Franzreb, K.E. 1997b. Success of intensive manage- Haig, S.M., Walters, J.R., and Plissner, J.H. 1994. ment of a critically imperiled population of red- Genetic evidence for monogamy in the red- cockaded woodpeckers in South Carolina. J Field cockaded woodpecker, a cooperative breeder. Be- Ornithol 68:458–470. hav Ecol Sociobiol 34:295–303. Franzreb, K.E. 1999. Factors that influence translo- Haig, S.M., Bowman, R., and Mullins, T.D. 1996. cation success in the red-cockaded woodpecker. Population structure of red-cockaded woodpeck- Wilson Bull 111:38–45. ers in south Florida: RAPDs revisited. Mol Ecol Franzreb, K.E. 2004. Effects of a mobile aviary in 5:725–734. translocation success of red-cockaded woodpeck- Hanski, I., and Simberloff, D. 1997. The metapop- ers. In Red-cockaded Woodpecker: Road to Recovery, ulation approach, its history, conceptual domain, eds. R. Costa and S.J. Daniels. Blaine, WA: Han- and application. In Metapopulation Biology, Ecology, cock House Publishers. In press. Genetics, and Evolution, eds. I. Hanski and M.E. Gordon, D.R. 1994. Translocation of species into Gilpin, pp. 5–26. San Diego, CA: Academic Press. conservation areas: A key for natural resource Harris, L.D., and Vickers, C.R. 1984. Some faunal managers. Nat Areas J 14:31–37. community characteristics of cypress ponds and Griffith, B., Scott, J.M., Carpenter, J.W., and Reed, the changes induced by perturbation. In Cypress C. 1989. Translocation as a species conservation Swamps, eds. K.C. Ewel and H.T. Odum, pp. 171– tool: Status and strategy. Science 245:477–480. 185. Gainesville: University Press of Florida. Griffith, B.J., Scott, J.M., Carpenter, J.W., and Reed, Harrison, S. 1994. Metapopulations and conserva- C. 1990. Translocations of captive-reared terres- tion. In Large-scale Ecology and Conservation Biology, trial vertebrates, 1973–1986. Endangered Species eds. P.J. Edwards, N.R. Webb, and R.M. May, pp. Update 8(1):10–13. 111–128. Oxford: Blackwell. Gutierrez, R.J., and Harrison, S. 1996. Applications Hedman, C.W., Poirier, J.R., Durfield, P.E., and Reg- of metapopulation theory to spotted owl man- ister, M.A. 2004. International Paper’s habitat agement: A history and critique. In Metapopu- conservation plan for the red-cockaded wood- lations and Wildlife Conservation Management, ed. pecker: Implementation and early success. In Red- D. McCullough, pp. 167–185. Covelo, CA: Island cockaded Woodpecker: Road to Recovery, eds. R. Costa Press. and S. J. Daniels. Blaine, WA: Hancock House Hagan, G., and Costa, R. 2001. Rare woodpeckers Publishers. In press. reintroduced to North Florida. Endangered Species Hess, C., and Costa, R. 1995. Augmentation from Bulletin 26:30–31. the Apalachicola National Forest: The develop- Hagan, G., Costa, R. and Phillips, M.K. 2004a. The ment of a new management technique. In Red- first reintroduction of red-cockaded woodpeck- cockaded Woodpecker: Recovery, Ecology and Manage- ers into unoccupied habitat: A private land and ment, eds. D.L. Kulhavy, R.G. Hooper, and R. conservation success story. In Red-cockaded Wood- Costa, pp. 385–388. Center for Applied Studies pecker: Road to Recovery, eds. R. Costa and S.J. in Forestry, College of Forestry, Stephen F. Austin Daniels. Blaine, WA: Hancock House Publishers. State University, Nacogdoches, TX. In press. Hooper, R.G., and McAdie, C.J. 1995. Hurri- Hagan, G. Engstrom, R.T., Cox, J., and Spivey, P. canes and the long-term management of the 2004b. Effects of translocation on a large red- red-cockaded woodpecker. In Red-cockaded Wood- cockaded woodpecker population. In Red-cockaded pecker: Recovery, Ecology and Management, eds. D.L. Woodpecker: Road to Recovery, eds. R. Costa and S.J. Kulhavy, R.G. Hooper, and R. Costa, pp. 148–166. 11. Reintroduction of Fauna 373

Center for Applied Studies in Forestry, College Kleiman, D.G. 1990. Decision-making about a rein- of Forestry, Stephen F. Austin State University, troduction: Do appropriate conditions exist? En- Nacogdoches, TX. dangered Species Update 8(1):18–19. Hooper, R.G., Niles, L.J., Harlow, R.F., and Wood, Lacy, R.C. 1987. Loss of genetic diversity from man- G.W. 1982. Home ranges of red-cockaded wood- aged populations: Interacting effects of drift, mu- peckers in coastal South Carolina. Auk 99:675– tation, immigration, selection, and population 682. subdivision. Conserv Biol 1:143–158. Hooper, R.G., Watson, J.C., and Escano, R.E.F. 1990. Laves, K.S., and Loeb, S.C. 1999. Effects of southern Hurricane Hugo’s initial effects on red-cockaded flying squirrels Glaucomys volans on red-cockaded woodpeckers in the Francis Marion National For- Picoides borealis reproductive success. Anim Conserv est. Trans North Am Wildl Nat Resourc Conf 55:220– 2:295–303. 224. Lay, D.W., and Swepston, D.A. 1973. Red-cockaded Hunter, P., Mahony, N.A., Ewins, P.J., and Field, M. woodpecker study. Job Completion Report, 1997. Artificial nesting platforms for bald eagles in Project W-80-R-16, Job 10. Texas Parks and southern Ontario, Canada. J Raptor Res 31(4):321– Wildlife Department. 326. Lennartz, M.R., Hooper, R.G., and Harlow, R.F. International Union for Conservation of Na- 1987. Sociality and cooperative breeding of red- ture and Natural Resources/Species Survival cockaded woodpeckers (Picoides borealis). Behav Commission, Re-introduction Specialist Group Ecol Sociobiol 20:77–88. (IUCN/SSC RSG). 1995. IUCN/SSG RSG guide- Leonard, D.L., Jr., DeLotelle, R.S., and Epting, lines for re-introductions. African Wildlife Foun- R.J. 2004. Factors contributing to variation dation, Nairobi, Kenya, Africa. in fledgling production in central Florida red- Jackson, J.A. 1977. A device for capturing tree cav- cockaded woodpeckers. In Red-cockaded Wood- ity roosting birds. North Am Bird Bander 2:14–15. pecker: Road to Recovery, eds. R. Costa and S.J. Jackson, J.A. 1990. Intercolony movements of red- Daniels. Blaine, WA: Hancock House Publishers. cockaded woodpeckers in South Carolina. J Field In press. Ornithol 61:149–155. Letcher, B.H., Priddy, J.A., Walters, J.R., and Crow- Jenkins, M.A., and Sherrod, S.K. 1993. Recent bald der, L.B. 1998. An individual-based, spatially- eagle nest records in Oklahoma. Bull Okla Ornithol explicit simulation model of the population dy- Soc 26:25–28. namics of the endangered red-cockaded wood- Johnsgard, P.A. 1990. Hawks, Eagles, Falcons of North pecker, Picoides borealis. Biol Conserv 86:1–14. America. Washington, DC: Smithsonian Institu- Levins, R. 1969. Some demographic and genetic tion Press. consequences of environmental heterogeneity for Jones, S.R. ed., 1990. Captive propagation and rein- biological control. Bull Entomol Soc Am 15:237– troduction: A strategy for preserving endangered 240. species? Endangered Species Update 8(1):3–88. Lohefener, R., and Lohmeier, L. 1981. Comparison Kappes, J.J., Jr. 1997. Defining cavity-associated of gopher tortoise (Gopherus polyphemus) habitats interactions between red-cockaded woodpeck- in young slash pine and old longleaf pine areas of ers and other cavity-dependent species: Interspe- southern Mississippi. J Herpetol 15:239–242. cific competition or cavity kleptoparasitism? Auk Lohr, S.M. 2004. Red-cockaded woodpecker recov- 114:778–780. ery efforts in an isolated and small South Carolina Kappes, J.J., Jr. 2004. Community interactions as- Sandhills population. In Red-cockaded Woodpecker: sociated with red-cockaded woodpecker cavities. Road to Recovery, eds. R. Costa and S.J. Daniels. In Red-cockaded Woodpecker: Road to Recovery, eds. Blaine, WA: Hancock House Publishers. In press. R. Costa and S. J. Daniels. Blaine, WA: Hancock Love, G.J., Wilkin, S.A., and Goodwin, M.H. 1953. House Publishers. In press. Incidence of blood parasites in birds collected in Kappes, J.J., Jr., and Harris, L.D. 1995. Interspe- southwestern Georgia. J Parasitol 38:52–57. cific competition for red-cockaded woodpecker Lowery, L., and Perkins, J. 2002. Long dispersal of a cavities in the Apalachicola National Forest. In red-cockaded woodpecker in central Florida. Fla Red-cockaded Woodpecker: Recovery, Ecology and Man- Field Nat 30:42–43. agement, eds. D.L. Kulhavy, R.G. Hooper, and R. Luttrell, M.P.,Stangel, P.W., and Bartlett, C.M. 1995. Costa, pp. 389–393. Center for Applied Studies in A survey for blood parasites in red-cockaded Forestry, College of Forestry, Stephen F. Austin woodpeckers. In Red-cockaded Woodpecker: Recovery, State University, Nacogdoches, TX. Ecology and Management, eds. D.L. Kulhavy, R.G. 374 IV. Restoration

Hooper, and R. Costa, pp. 332–334. Center for Petrick, C.J., and Hagedorn, B.W. 2004. Population Applied Studies in Forestry, College of Forestry, status and trend of red-cockaded woodpeckers Stephen F. Austin State University, Nacogdoches, on Eglin Air Force Base, Florida. In Red-cockaded TX. Woodpecker: Road to Recovery, eds. R. Costa and S.J. Marsh, D.M., and Trenham, P.C. 2001. Metapopula- Daniels. Blaine, WA: Hancock House Publishers. tion dynamics and amphibian conservation. Con- In press. serv Biol 15:40–49. Phillips, M.K. 1990. Red wolf: Recovery of an Marston, T.G., and Morrow, D.M. 2004. Red- endangered species. Endangered Species Update cockaded woodpecker conservation on Fort 8(1):79–81. Jackson military installation: A small popula- Platt, W.J., Evans, G.W., and Rathbun, S.L. 1988a. tion’s response to intensive management in the The population dynamics of a long-lived conifer Sandhills of South Carolina. In Red-cockaded Wood- (Pinus palustris). Am Nat 131:491–525. pecker: Road to Recovery, eds. R. Costa and S.J. Platt, W.J., Evans, G.W., and Rathbun, S.L. 1988b. Daniels. Blaine, WA: Hancock House Publishers. Effects of fire season on flowering of forbs and In press. shrubs in longleaf pine forests. Oecologia 76:353– Mengel, R.M., and Jackson, J.A. 1977. Geographic 363. variation of the red-cockaded woodpecker. Condor Poirier, J., Hedman, C., and Register, M. 2004. 79:349–355. Southern flying squirrel use of artificial red- Mills, L.M., Feltner, K.J., and Reed, T.O. 2004. The cockaded woodpecker cavities at International rise and fall of the red-cockaded woodpecker Papers Southlands Forest: The past, present, and population in Kentucky: A chronology of events future. In Red-cockaded Woodpecker: Road to Recov- preceding extirpation. In Red-cockaded Woodpecker: ery, eds. R. Costa and S.J. Daniels. Blaine, WA: Road to Recovery, eds. R. Costa and S.J. Daniels. Hancock House Publishers. In press. Blaine, WA: Hancock House Publishers. In Pung, O.J., Carlile, L.D., Whitlock, J., Vives, S.P., press. Durden, L.A., and Spadgenske, E. 2000. Survey Mitchell, L.R., Carlile, L.D., and Chandler, C.R. and host fitness effects of red-cockaded wood- 1999. Effects of southern flying squirrels on nest pecker blood parasites and nest cavity arthropods. success of red-cockaded woodpeckers. J Wildl J Parasitol 86:506–510. Manage 63:538–545. Ralls, K., Ballou, J.D., and Templeton, A. 1988. Es- Montague, W.G., and Bukenhofer, G.A. 1994. Long timates of lethal equivalents and the cost of in- range dispersal of a red-cockaded woodpecker. breeding in mammals. Conserv Biol 2:185–193. Proc Arkansas Acad Sci 48:259–260. Reinman, J.P. 1984. Woodpeckers receive second Morris, V., and Werner, C. 2004. Withlacoochee chance. Fla Nat 57:17. State Forest Croom Tract red-cockaded wood- Robbins, L.E., and Meyers, R.L. 1992. Seasonal ef- pecker recovery effort. In Red-cockaded Woodpecker: fects of prescribed burning in Florida: A review. Road to Recovery, eds. R. Costa and S.J. Daniels. Miscellaneous Publication No. 8. Tall Timbers Re- Blaine, WA: Hancock House Publishers. In press. search Station, Tallahassee, FL. Mount, R. 1975. The Reptiles and Amphibians of Al- Rudolph, D.C., Conner, R.N., Carrie, D.K., and abama. Auburn, AL: Auburn Printing Company. Schaefer, R.R. 1992. Experimental reintroduc- Neal, J.C., and Montague, W.G. 2000. Serviceabil- tion of red-cockaded woodpeckers. Auk 109:914– ity of RCW cavities. U.S. Forest Service, Ouachita 916. National Forest, Waldron, AR. Saenz, D., Baum, K.A., Connor, R.N., Rudolph, Nesbitt, S.A., Jenkins, A., Sherrod, S.K., Wood, D.C., and Costa, R. 2002. Large-scale translo- D.A., Beske A., White, J.H., Schulz, P.A., and cation strategies for reintroduced red-cockaded Schwikert, S.T. 1998. Recent status of Florida’s woodpeckers. J Wildl Manage 66:212–221. bald eagle population and its role in eagle reestab- Saenz, D., Schaefer, R.R., Conner, R.N., Best, M.S., lishment efforts in the southeastern United States. Rudolph, D.C., and Carrie, D.K. 2004. Influ- Proc Annu Conf Southeast Assoc Fish Wildl Agenc ence of artificial cavity age on red-cockaded 52:377–3883. woodpecker translocation success. In Red-cockaded Parker, W.T. 1987. A plan for reestablishing the red Woodpecker: Road to Recovery, eds. R. Costa and S. J. wolf on Alligator River National Wildlife Refuge, Daniels. Blaine, WA: Hancock House Publishers. North Carolina. Red Wolf Management Series: In press. Technical Report No. 1. U.S. Fish and Wildlife Ser- Schiegg, K., Walters, J.R., and Priddy, J.A. 2002. vice, Atlanta, GA. The consequences of disrupted dispersal in 11. Reintroduction of Fauna 375

fragmented red-cockaded woodpecker Picoides bo- Coastal Plain. In Biodiversity of the Southeastern realis populations. J Anim Ecol 71:710–721. United States: Lowland Terrestrial Communities, eds. Scott, J.M., and Carpenter, J.W. 1987. Release of W. H. Martin, S. G. Boyce, and A. C. Echter- captive-reared or translocated endangered birds: nacht, pp. 373–446. New York: John Wiley and what do we need to know? Auk 104(3):544– Sons. 545. Stys, B. 1993. Ecology and habitat protection Seigal, R.A., and Dodd, C.K., Jr. 2002. Translo- needs of southeastern American kestrel (Falco cations of amphibians: Proven management sparverius paulus) on large-scale development method or experimental technique? Conserv Biol sites in Florida. Florida Game and Fresh Wa- 16(2):552–554. ter Fish Commission, Nongame Wildlife Pro- Selender, R.K. 1983. Evolutionary consequences of gram Technical Report No. 13. Tallahassee, inbreeding. In Genetics and Conservation: A Refer- FL. ence for Managing Wild Animal and Plant Popula- Trenham, P.C., and Marsh, D.M. 2002. Amphibian tions, eds. C.M. Shonewald-Cox, S.M. Chambers, translocation programs: Reply to Seigel and Dodd. B. MacBryde, and L. Thomas, pp. 201–215. Men- Conserv Biol 16(2):555–556. low Park, CA: Benjamin/Cummings. U.S. Fish and Wildlife Service. 1987. Habitat man- Shaffer, M.L. 1981. Minimum population sizes for agement guidelines for the bald eagle in the species conservation. Bioscience 31:131–134. southeast region. U.S. Fish & Wildlife Service, Shaffer, M.L. 1987. Minimum viable populations: Washington, DC. Coping with uncertainty. In Viable Populations for U.S. Fish and Wildlife Service. 1990a. Gopher tor- Conservation, ed. M.E. Soule, pp. 69–86. London: toise (Gopherus polyphemus) recovery plan. U.S. Cambridge University Press. Fish & Wildlife Service, Atlanta, GA. Sherrod, S.K. Jenkins, M.A., McKee, G., Tatum, S., U.S. Fish and Wildlife Service. 1990b. Red wolf re- and Wolfe, D. 1987. Using wild eggs for produc- covery plan. U.S. Fish & Wildlife Service, Atlanta, tion of bald eagles for reintroduction into south- GA. eastern United States. In Proceedings Third South- U.S. Fish and Wildlife Service. 1991. Mississippi east Nongame and Endangered Wildlife Symposium, sandhill crane recovery plan. U.S. Fish & Wildlife eds. R.R. Odum, K.A. Riddleberger, and J. C. Service, Atlanta, GA. Ozier, pp. 14–20. Georgia Department of Natural U.S. Fish and Wildlife Service. 2003. Recovery plan Resources. for the red-cockaded woodpecker (Picoides bore- Speake, D.W., and Mount, R.H. 1973. Some possi- alis): second revision. U.S. Fish & Wildlife Service, ble ecological effects of rattlesnake roundups in Atlanta, GA. the southeastern Coastal Plain. Proc 27th Annu Vickers, C.R., Harris, L.D., and Swindel, B.F. 1985. Conf Southeast Assoc Game Fish Comm 27:267– Changes in herpetofauna resulting from ditching 277. of cypress ponds in coastal plains ﬂatwoods. For Speake, D. W., McGlincy, J.A., and Colvin, T.A. Ecol Manage 11:17–29. 1978. Ecology and management of the eastern Walker, J.L., and Peet, R.K. 1983. Composition indigo snake in Georgia: A progress report. In Pro- and species diversity of pine–wiregrass savannahs ceedings of the Rare and Endangered Wildlife Sym- of the Green Swamp, North Carolina. Vegetatio posium, Technological Bulletin WL 4, eds. R.R. 55:163–179. Odom and L. Landers, pp. 64–73. Game and Walters, J.R. 1991. Application of ecological prin- Fish Division, Georgia Department of Natural Re- ciples to the management of endangered species: sources. The case of the red-cockaded woodpecker. Annu Stangel, P.W., Lennartz, M.R., and Smith, M.H. Rev Ecol Syst 22:505–523. 1992. Genetic variation and population structure Walters, J.R. 2004. Unusual dynamics in a rapidly of red-cockaded woodpeckers. Conserv Biol 6:283– increasing population of red-cockaded wood- 292. peckers at Camp Lejeune, North Carolina. In Red- Stober, J.M., and Jack, S.B. 2004. Down for the cockaded Woodpecker: Road to Recovery, eds. R. Costa count? Red-cockaded woodpecker restoration on and S. J. Daniels. Blaine, WA: Hancock House Ichauway. In Red-cockaded Woodpecker: Road to Re- Publishers. In press. covery, eds. R. Costa and S.J. Daniels. Blaine, WA: Walters, J.R., Doerr, P.D., and Carter, J.H., III. Hancock House Publishers. In press. 1988a. The cooperative breeding system of Stout, I.J., and Marion, W.R. 1993. Pine ﬂatwoods the red-cockaded woodpecker. Ethology 78:275– and xeric pine forests of the southern (Lower) 305. 376 IV. Restoration

Walters, J.R., Hansen, S.K., Carter, J.H., III, Manor, in promoting growth of the red-cockaded P.D.,and Blue, R.J. 1988b. Long distance dispersal woodpecker population on Eglin Air Force of an adult red-cockaded woodpecker. Wilson Bull Base, Florida. In Red-cockaded Woodpecker: Road 100:494–496. to Recovery, eds. R. Costa and S.J. Daniels. Walters, J.R., Crowder, L.B., and Priddy, J.A. 2002. Blaine, WA: Hancock House Publishers. In Population viability analysis for red-cockaded press. woodpeckers using an individual-based model. Wolf, C.D., Grifﬁth, B., Reed, C., and Temple, S.A. Ecol Appl 12:249–260. 1996. Avian and mammalian translocations: Up- Walters, J.R., Gault, K.E., Hagedorn, B., Pet- date reanalysis of 1987 survey data. Conserv Biol rick, C.J., Phillips, L.F., Jr., Tomcho, J., Jr., 10:1142–1154. and Butler, A. 2004. Effectiveness of recruit- Wood, D.A. 1982. Florida’s bald eagle. Fla Wildl ment clusters and intrapopulation translocation 35:42. Chapter 12 Spatial Ecology and Restoration of the Longleaf Pine Evosystem

Thomas S. Hoctor, Reed F. Noss, Larry D. Harris, and K. A. Whitney

Introduction between fire and hydrology were key ecological processes facilitated and mediated largely Trees of the genus Pinus have been dom- by the fire ecology of longleaf pine. These pro- inant species in the southeastern Coastal cesses combined with variation in climate, to- Plain (SECP) landscape since the Pleistocene pography, and soils resulted in spatial hetero- (c. 2 million years). When Spanish conquis- geneity across spatial scales ranging from lo- tadors arrived 500 years ago, longleaf pine cal sites and ecotones to regional landscapes. (Pinus palustris) forest covered millions of This heterogeneity supported a vast array of hectares, and was the dominant land cover flora and fauna, unsurpassed virtually any- type throughout the SECP from southeastern where in the temperate zone. Species that Virginia to eastern Texas. Longleaf-dominated evolved in this ecological context were fre- natural communities were also a significant quently dependent on vast longleaf pine land- component of some Piedmont and southern scapes maintained by spatially extensive dis- Appalachian landscapes up to 600 m above turbance regimes (Means and Grow 1985; mean sea level (Varner et al. 2003a). From the Noss 1988; Myers 1990; Means 1996; Harris wet flatwoods of Louisiana to the sandhills of et al. 1996a). southeastern Virginia, and from the ancient Not surprisingly, then, the wholesale reduc- sand dune ridges and flatwoods of south- tion of longleaf pine landscapes to small rem- central Florida to the Ridge and Valley, Cum- nant patches has led to widespread ecologi- berland Plateau, and Blue Ridge physiographic cal dysfunction and endangerment of many provinces of northern Georgia and Alabama, species (Noss et al. 1995). Remnant patches longleaf pine has played a leading role in the are incapable of supporting the functional eco- ecology and evolution of southeastern biota. logical processes needed to maintain ecologi- Over much of its former extent, longleaf pine cal integrity or the spatial extensiveness and communities served as the matrix surround- ecological juxtapositions necessary to maining many different wetland and smaller upland tain viable populations of many species that communities. In these landscapes, interactions evolved in longleaf pine landscapes (Fig. 1).

r Thomas S. Hoctor Department of Landscape Architecture, University of Florida, Gainesville, Florida 32611. r Reed F. Noss Department of Biology, University of Central Florida, Orlando, Florida 32816. Larry D. r Harris Department of Wildlife Ecology and Conservation, University of Florida, Gainesville, Florida 32611. K. A. r Whitney Department of Urban and Regional Planning, University of Florida, Gainesville, Florida 32611.

377 378 IV. Restoration

NORTH CENTRAL FLORIDA WOODPECKER SPECIES ABUNDANCE

1950

RED COCKADED .40 .30 .20 RED-HEADED RED-BELLIED COMMON FLICKER DOWNY .10 HAIRY ABUNDANCE YELLOW-BELLIEDPILEATED SAPSUCKER

.10 .20 .30 1980-85 RELATIVE .40 .50 .60

FIGURE 1. Dramatic changes in wildlife communities are one result of the demise of longleaf pine forests and ecological dysfunction. As longleaf pine communities have become vastly diminished in north-central Florida, the woodpecker guild has been altered dramatically.

The demise of longleaf pine began soon after ecological processes, ecological context, inter- European colonization, with commercial tim- actions between natural communities, suffi- bering and agricultural conversion decimat- cient aerial extent (both of individual patches ing some longleaf pine landscapes well before and networks of patches), and functional con- 1900. These changes have left an indelible nectivity. The restoration of longleaf pine landmark on our understanding of longleaf pine scapes is required to reestablish the dominant ecology, which may cloud future attempts at role of this species in the regional ecology of restoration. the southeastern United States. Given the distinctive ecology and current We use the term “evosystem” in this chap- condition of longleaf pine communities, an ter to emphasize the critical influence of a understanding of landscape ecology and re- regionally dominant ecosystem on the evolu- gional reserve design principles is crucial for tion of associated species (Hutchinson 1965; guiding restoration. A regional-scale strategy Harris et al. 1996a). As the predominant ma- that integrates the principles and techniques trix vegetation in much of the southeastern of landscape ecology, conservation biology, United States, longleaf pine provided the con- and restoration ecology is required to restore text within which a diverse flora and fauna the diversity of longleaf pine communities. evolved (Means 1996; Platt 1999). This con- Such a strategy should include restoration and text has been disrupted, and longleaf pine maintenance of all variants of longleaf pine no longer exerts the ecological or evolution- communities across their historic range. Con- ary influence it once did. The massive decline servation at this scale requires attention to (greater than 97% from the pre-European dis- between-stand diversity and function, spatial tribution, by virtually all estimates; e.g., Ware 12. Spatial Ecology and Restoration 379 et al. 1993) of a regional matrix vegetation these scales, occurring on extremely well- has ramifications on the persistence and future drained sandhills (e.g., “desert in the rain,” evolution of far more species—and on the op- Wells and Shunk 1931), in seasonally flooded eration of fundamental ecological processes— flatwoods throughout the SECP, on moun- than would the loss of unique natural com- tain slopes in Georgia and Alabama, and on munities that were never common (Noss et al. relict sand dune ridges in central and south- 1995). Yet, with its emphasis on rarity, conser- central Florida (Noss 1988; Myers 1990; Platt vation strategy has more often focused on the 1999). Across this ecological spectrum, lon- latter than the former. gleaf pine communities varied from open savannas with often stunted trees on drier, less productive sites to massive old-growth forests Landscape Ecology and with trees at least 450 years old on more productive clay soils or adjacent to riparian com- Historic Longleaf Pine munities (Noss 1988; Platt et al. 1988; Means Communities 1996; Harris 1999). Although there are com- monalities among these communities, associ- In its pristine condition with millions of trees mea- ated herbaceous species and associated natural suring a yard or more in basal diameter, the Pinus communities show a broad range of variation palustris consocies unquestionably presented one of (Platt 1999). In the SECP, even extremely sub- the most wonderful forests in the world . . . . The tle differences in elevation can have profound complete destruction of this forest constitutes one effects on species composition and vegetation of the major social crimes of American history. structure (Noss 1988; Noss and Harris 1990; (Wells and Shunk 1931) Platt 1999). For instance, a change in eleva- In presettlement times, longleaf pine forests tion of less than a few meters can result in covered approximately two-thirds (more than a gradient of communities ranging from ex- 37 million hectares) of the upland surface area tremely well-drained and desertlike sandhills in the SECP (Wahlenberg 1946; Noss 1988; to wet flatwoods and swamps. Along seepage Ware et al. 1993; Means 1996; Platt 1999; slopes and other ecotones between uplands Frost this volume). The concept of matrix com- and wetlands, pronounced changes in plant munities or dominant land cover is integral species composition and structure can occur in landscape ecology. Landscape ecology has over an elevation range of only a few cen- become increasingly relevant in efforts to un- timeters (Noss and Harris 1990; Means 1996; derstand, protect, restore, and manage bio- Platt 1999). As a fire-adapted, matrix land diversity because it focuses on the interac- cover, longleaf pine communities play a key- tions between spatial patterns and ecological stone role in maintaining between-habitat di- processes (Turner 1989; Forman 1995). These versity throughout the SECP (Means 1996). processes and patterns generate and control Both gradual and abrupt ecotones are promi- the distribution of biodiversity. The concept of nent in longleaf pine-dominated landscapes landscape, either implicitly or explicitly, incor- and play significant roles in the propagation porates the notion that two or more identifi- and influence of spatial ecological processes ably different ecosystems (or natural commu- such as fire. High plant diversity, primarily in nities), and their associated species, interact in the herbaceous layer, is a prominent charac- a spatially patterned mosaic (Forman and Go- teristic of most intact longleaf pine commu- dron 1986; Harris et al. 1996a). nities (Clewell 1981; Walker and Peet 1983; Landscape ecology increasingly has recog- Peet and Allard 1993), and many of these nized the significance of spatial heterogene- plant species achieve their greatest abundance ity over a range of scales from regional to lo- within ecotones (Noss 1988). Hence, tradi- cal (Forman 1995; Turner et al. 1995; Poiani tional plant community classifications, which et al. 2000). Longleaf pine communities ex- avoid ecotones and sample the “characteristic” hibited striking spatial heterogeneity across interior areas of communities (Noss 1987a), 380 IV. Restoration miss much of the diversity in landscapes such physiognomy. Dominance in the overstory is as those dominated by longleaf pine. Due to shared with such species as blackjack oak the complexity of environmental gradients at (Quercus marilandica) and sand hickory (Carya multiple spatial scales, several hundred plant pallida), while ice and snowstorms join the rel- species can occur in areas as small as 1 ha. Beta atively less frequent fire in maintaining lon- (between-habitat) and gamma (regional) di- gleaf pine (Varner et al. 2003a,b). On other versity are correspondingly high (Platt 1999). sites in the lower Piedmont and upper Coastal We consider landscapes to include two or Plain, tree co-dominants include southern red more, spatially and compositionally distinct oak (Quercus falcata), flowering dogwood (Cor- community types that interact through a va- nus florida), post oak (Quercus stellata), loblolly riety of spatial and temporal processes (Noss pine (Pinus taeda), shortleaf pine (Pinus echi- 1987a; Turner 1989; Forman 1995; Harris et al. nata), and sweetgum (Liquidambar styraciflua) 1996a; Poiani et al. 2000). Adjacency and/or (Beckett and Golden 1982; Peet and Allard connectivity are important aspects of land- 1993). scape pattern that facilitate interactions among landscape components through the flow of organisms, materials, water, and energy (For- Longleaf Pine Exploitation man 1995; Harris et al. 1996a). Such interactions and functions were commonplace be- and the Deconstruction of fore wholesale destruction and degradation Longleaf Pine Ecology of longleaf pine landscapes. Across much of the range of longleaf pine, ecological pro- It is not our purpose here to review the long cesses such as fire and the myriad species com- history of longleaf pine utilization and ex- pletely or partially associated with longleaf ploitation, but we have concerns about the pine strongly influenced the ecological struc- usual accounts of longleaf pine history and ture and function of the SECP. Wiregrass (Aris- their effects on perceptions of its ecology. It tida stricta and A. beyrichiana), for example, is well accepted that, since European settle- along with other grasses and the duff (fallen ment, the extent of longleaf pine has declined needles) of longleaf pine, promoted the spread to less than 2 or 3% of its former area (Means of frequent, low-intensity ground fires across and Grow 1985; Noss 1988, 1989; Ware et al. the landscape (Platt 1988; Noss 1989), which 1993; Means 1996). However, it is important in turn shaped vegetation structure and spa- to recognize the great length of time and in- tial patterns at multiple scales. Interactions be- creasing intensification of use and destruction tween longleaf pine-dominated uplands, scrub of longleaf pine resources since European ex- (mostly in Florida, with a somewhat lower fire ploitation began. By the year 2108 Florida will frequency), hardwood hammocks on sites to- have been part of the United States for the pographically protected from fire such as is- same length of time that it was under Span- lands, peninsulas, and karst grottos (Harper ish rule. Thus, human impact on southeastern 1911; Platt and Schwartz 1990; Means 1996), forest resources has occurred over a substan- and various types of riparian wetlands were tial period of time and has taken place under critical for generating and maintaining the re- the influence of several cultural groups (Harris gion’s biodiversity. Importantly, the archety- 1980; Myers 1990; Frost 1993, this volume). pal longleaf pine ecosystem—the open, park- Review of the historical literature regarding like forests of the Coastal Plain maintained the use and exploitation of forest resources by very high fire frequencies—does not ap- in the SECP indicates that significant impacts ply to all longleaf pine communities. In par- have been ongoing for most of the period ticular, the montane longleaf pine communi- since Europeans arrived. Although human im- ties of northern Georgia and Alabama have pact is not a particularly modern influence been poorly studied but seem to differ in on this ecosystem, the intensive site prepara- disturbance regime, species composition, and tion and other practices of modern industrial 12. Spatial Ecology and Restoration 381 silviculture may indeed have more lasting ef- the following sections we focus on the primary fects on the ecology of longleaf stands and aspects of the spatial ecology of longleaf pine landscapes than did more primitive forms of and on the need for restoration that accounts exploitation (Noss 1988). for this spatial heterogeneity and landscape- Although the types of forest products ex- level processes. tracted and forms of exploitation changed over the centuries, the trend of resource use has been one of continually increasing intensity, The Spatial Ecology of at least until very recently when clearcutting Longleaf Pine Natural of longleaf pine on public lands diminished (Elliott 1912; Greeley 1920; Tyson 1956; Communities: From Basic Adams 1976; Frost 1993; Stout and Marion Biology to Biogeography 1993; Harris 1999). Both the forest and wildlife populations reflect these changes. Relevant Basic Longleaf Pine The long history of exploitation in longleaf pine forests in the SECP is an essential consid- Biology eration for restoring the ecological function of Longleaf pine and wiregrass (and other herba- this natural community. Although the usual ceous species in some cases) are the ecolog- definition of longleaf pine forests as open- ically pivotal species of these pyrophilic sys- canopy “savannas” certainly is supported by tems, and frequent, low-intensity fire is the anecdotal accounts and the ecology of remain- single most critical ecological process neces- ing longleaf pine ecosystems on less productive sary for their maintenance (Noss 1988; Harris sites within the SECP, the distinctive charac- et al. 1996a; Whitney 1999). Longleaf pines teristics of longleaf pine forests on more pro- convert lightning strikes to ground fires (Platt ductive soils and outside the SECP have been et al. 1988). Fallen pine needles, with their obscured by the trends and pattern of histori- volatile oils and resins, combined with highly cal exploitation of this species. Longleaf pines flammable grasses promote frequent fires that on bottomland and other productive soils were eliminate most hardwoods and other poten- frequently cleared first because of easier access tially competing plants (Noss and Harris 1990; and conversion to agricultural uses. And fire Harris et al. 1996a; Platt 1999). This situation suppression has resulted in hardwood dom- supports the hypothesis of Mutch (1970), who inance in alluvial and other productive sites suggested that the flammable properties of that once supported impressive forests of lon- some plant species have been favored by natu- gleaf pine (Harris 1999). Hence, few examples ral selection because they reduce competition of these communities remain. This pattern is with other plant species. In addition, longleaf not exceptional. Across the United States and pine’s entire life cycle is linked to fire, and seed in other countries, the proportion of protected germination and seedling establishment are areas is lowest on low-elevation lands with enhanced if the seed falls on sparsely vegetated highly productive soils (Harris 1984; Scott et or bare mineral soil (Myers 1990). Volatile al. 2001). In other cases, longleaf pine on high- resins produced by longleaf pine not only pro- productivity sites was likely selectively cut, so mote fire but also deter pest outbreaks that are that the density and average size of trees de- more common in less fire-adapted yellow pine clined over time (Fig. 2). Because longleaf pine species (Harris et al. 1996a; Whitney 1999). communities occurred on a variety of sites and Lighterwood in the form of standing and have been affected by numerous forms of ex- prostrate longleaf pine logs existed in large ploitation over a long period of time, consider- volumes in the original forest (Harris 1999). ation of the landscape ecology of longleaf pine Settlers could use this readily available “down communities on regional, landscape, and local wood” for many years before they had to do scales is critical for developing restoration and actual logging. Indeed, the turpentine indus- management strategies. With this in mind, in try got its start by using down literwood and 382 IV. Restoration

FIGURE 2. Low-density stands of longleaf pine have now become so commonplace that many writers consider them to be characteristic of original conditions. Rich sites, such as those that were settled first, and alluvial sites that have now become dominated by hardwoods once supported dense stands of large longleaf pines. Source: Mohr, C. 1896. Timber pines of the southern United States. Bulletin No. 13, USDA Forest Service, Washington, DC. dragging it into piles on the ground and simply unique niches for various plants and animals letting the various grades of tars flow into the in the ground cover (Hermann 1993). barrels (Butler 1998). Such wood played a key Nutrient cycling is also an important is- role in virgin longleaf pine forests by providing sue on sandy sites where nutrients quickly 12. Spatial Ecology and Restoration 383 leach downward from surface layers. Gopher frequent, fires. In Marion and Alachua coun- tortoise (Gopherus polyphemus), pocket gopher ties in north-central Florida a prime exam- (Geomys pinetis), and scarab beetles (Peltotrupes ple is the southernmost extension of upland sp.) all help reverse the direction of nutrients pine forest (sometimes referred to as “red oak leaching downward (Kalisz and Stone 1984; woods”) found on richer soils including on Myers 1990). An endemic scarab beetle from the edges of ravines, sinkholes, or other rela- north-central Florida (Peltotrupes youngi) digs tively steep slopes (Harper 1915; Duever 1983; 360 cm or deeper and can transport as much Myers 1990). These communities had higher as 8 mg of subsoil to the surface per hectare productivity and harbored some plant species per year (Kalisz and Stone 1984). [including canopy co-dominants southern red oak and mockernut hickory (Carya alba)] that Heterogeneity: Between- and were absent or less common in adjacent com- among-Stand Considerations, Ecotones, munities both up- and downslope. The greater and Mosaics production of food resources was likely significant for herbivorous species such as fox squir- Spatial heterogeneity among stands of lon- rels (Sciurus niger), Florida mice (Podomys flori- gleaf pine and other natural communities em- danus), and gopher tortoises (Gopherus polyphe- bedded within the longleaf pine matrix was mus), and for carnivores including the eastern a key characteristic of these landscapes. Tree indigo snake (Drymarchon corais couperi) that form, density, and related characteristics gov- would benefit from increased prey densities. ern the within-stand heterogeneity, while the Fires in longleaf pine communities are also size, shape, and arrangement of the stands important for maintaining fire-adapted plant in the larger forest management unit affect communities such as pitcher plant bogs. Fre- landscape heterogeneity (Harris 1980, 1984). quent fires prune back wetland shrubs that Within stands of longleaf pine, elevation, soils, would otherwise encroach upslope, maintain- and disturbance history all influenced stand ing an open herbaceous bog community in structure. Various disturbances, including hur- some cases (Fig. 3; Noss and Harris 1990; ricanes, tornadoes, floods, and diseases, also Means 1996). This is particularly pronounced strongly influenced stand structure (Harris et on seepage slopes in the Florida Panhandle and al. 1996a). These patterns and processes are southern Alabama. Ecotones in longleaf pine essential for creating and maintaining the re- landscapes tend to be very high in plant di- quired environmental conditions for many versity, and their degradation or outright de- species. In addition, because many species re- struction results in significant erosion of bio- quire different community types during dif- diversity and ecological integrity. Fires moving ferent seasons or life stages, and must move downhill from longleaf pine communities also between communities on a regular basis, the help maintain graminaceous vegetation within arrangement of community types within the ephemeral wetlands needed or preferred by landscape is of critical importance (Noss 1987a; flatwoods salamanders (Ambystoma cingulatum) Harris 1989; Forman 1995; Harris et al. 1996a). and other amphibians, many of which are im- Gradual ecotones and abrupt edges between periled or declining (Huffman and Blanchard longleaf pine and other natural communities 1990; Dodd and LeClaire 1995; Palis 1996; were both common in intact landscapes (Barry Dodd and Cade 1998; Cox and Kautz 2000; 1980; Noss and Harris 1990; Means 1996; Platt Kautz and Cox 2001). 1999). Ecotones reflect changes in edaphic In Florida, longleaf pine sandhills are fre- conditions and interactions between processes, quently sharply juxtaposed with scrub com- especially fire and flooding, in longleaf pine munities. These communities are dependent landscapes. Ecotones included what could be on starkly different fire regimes, with scrub, considered unique natural communities that which is dominated by short-stature oaks or were dependent on more productive soils or sand pine (Pinus clausa), typically burning wetter soils along with frequent, or somewhat catastrophically on 15- to 80-year intervals 384 IV. Restoration

FIGURE 3. Fire moving downslope from longleaf pine communities maintains herbaceous wetland communities along ecotones that would otherwise be dominated by shrubby or forested wetlands. Such “ecotone communities maintained by a combination of wetter or richer soils (including herb bogs and red oak woods) are often the ﬁrst to disappear through ﬁre suppression. Source: Means (1996).

(Myers 1990). The transition between these ridge top to valley bottom, is also relevant two communities is often sharply demarcated (Milne 1935, 1947; Woodmansee 1990; For- and is thought to represent changes in historic man 1995). Though most energy and matter- fire regimes that then tend to be perpetuated flows in a catena may be downslope, pro- by the vastly different vegetative physiognomy cesses such as flooding, fire, movement of nu- and flammability (Myers 1990). trients, animal movements, and windflows can Spatial mosaics are an integral part of func- move upslope as well (Boerner and Kooser tional landscapes (Forman 1995). The discus- 1989; Woodmansee 1990; Forman 1995; Har- sion regarding ecotones touches on this is- ris et al. 1996a). Therefore, ecosystem clus- sue, but the interactions between different ters and catenas can be seen as networks con- patches within a longleaf pine landscape tran- nected by ecological processes flowing across scend edge or ecotonal phenomena. Forman’s landscapes. (1995) concept of the “ecosystem cluster” is Spatial mosaics have a strong influence on relevant. An ecosystem cluster “is recognized fire patterns within longleaf pine landscapes. as a spatial level of hierarchical organization Linear riparian networks, wetland patches, between the local ecosystem and the land- sinkholes, and ravines all can serve as signifi- scape. It describes a group of spatial elements cant barriers that create fire shadows of vari- connected by a significant exchange of en- ous sizes (Harper 1911; Noss and Harris 1990; ergy or matter” (Forman 1995:287). This is Platt and Schwartz 1990; Means 1996). Fire similar to the “functional landscape” concept shadows, from scales ranging from individual of Noss (1987a) where multiscalar interac- down lighterwood (Hermann 1993) to wet- tions between patches are considered essential land patches and riparian strips, include signif- for maintaining ecological integrity and bio- icant spatial heterogeneity and often support diversity (Harris 1984; Noss and Harris 1986; natural communities intolerant of fire, or fre- Noss 1987a). The concept of catena, used to quent fire, and in turn provide habitat for addi- describe connected sequences of soils from tional species and produce additional resources 12. Spatial Ecology and Restoration 385

FIGURE 4. The often interdigitated spatial relationship between longleaf pine forest and adjacent riparian forests and wetland communities is a critical ecological characteristic of intact longleaf pine landscapes. Many species require both upland and wetland communities to meet their life history needs, interactions between ﬁre and ﬂooding along these juxtapositions maintain unique natural communities, and riparian swaths provide corridors for many species. Reprinted from Harris, L.D., Hoctor, T. S., and Gergel, S. E. 1996a, with permission from the University of Chicago Press.

Hardwood Forest

Longleaf Pine Upland Forest

such as mast (e.g., acorns) for wildlife. Fire also tween natural communities, and provided cor- affects longleaf pine landscape mosaics beyond ridors for wetland species (Fig. 4; Delcourt and the boundaries of fire-adapted communities. Delcourt 1977; Harris 1988; Noss and Harris For example, Torreya taxifolia, an endangered 1990; Delcourt et al. 1993; Harris et al. 1996a): tree found only along bluffs of the Apalachicola River in the Florida Panhandle, may be imperiled because, due to fire suppression, smoke The role of riparian, bottomland forests in the land- from adjacent longleaf pine communities no scape is also critical. Between-community qualities longer controls fungal diseases (Schwartz et al. include their high contrast with conifer-dominated 1995). uplands, and a phenology of mast and other re- The significance of the relationship between source production that is non-synchronous with longleaf pine uplands and wetlands extends upland communities (Harris 1989). . . . Finally, the landscape-level functions of ripar- beyond interactions between flooding and fire. ian forests include dispersal of plants and animals In many landscapes formerly dominated by along both terrestrial and aquatic pathways. The lin- longleaf pine, riparian strips bisected uplands ear, dendritic pattern of riparian forests makes them in interdigitated patterns that resulted in fire- ideal collectors and transporters of wildlife across breaks, facilitated functional juxtapositions be- regional landscapes. (Noss and Harris 1990: 133) 386 IV. Restoration

30 Bottomlands

Uplands

Mast as % Natural Foods Mast as % Natural 10

Winter Spring Summer Fall

FIGURE 5. Florida black bear (Ursus americanus ﬂoridanus) in longleaf pine landscapes require both uplands that produce saw palmetto fruit (Serenoa repens) and other soft mast and riparian hardwood forests that produce acorns and other hard mast.

Spatial mosaics are also essential for main- Other species found in longleaf pine land- taining viable populations of many species scapes are also dependent on functional spa- found in longleaf pine landscapes. For exam- tial mosaics (Fig. 5). Without downplaying ple, Weigl et al. (1989) argued that manage- the significance of frequently burned, open- ment prescriptions for the red-cockaded wood- canopy longleaf pine stands, less frequently pecker (Picoides borealis) calling for extensive, burned communities are critical components frequent burning and removal of almost all of the broader landscape mosaic for many hardwoods throughout longleaf pine forests species. For example, sandhills and other open would not necessarily provide ideal habitat for pinelands are considered only secondary habi- fox squirrels, which are also imperiled: tat for the Florida black bear (Ursus americanus floridanus), and frequent fires that remove too much shrubby cover can reduce habitat qual- any action that removes all or most of the larger ity (Cox et al. 1994; Maehr 1997; Maehr et al. oaks (or hickories) from among the pines of the pre- 2001a; Hoctor 2003). In particular, saw pal- ferred pine-oak habitat will have a devastating ef- metto fruit (Serenoa repens) is an exceptionally fect on the food supply and nest cavity availability important food resource, and fruit production for the fox squirrel and many other kinds of wildlife. is highest in palmetto stands that are at least 5 The continuous thinning of dense small hardwoods years old (Hilmon 1969; Maehr 1997; Maehr by fire or other means would be highly beneficial in et al. 2001a). To facilitate mosaics that could most forests, but the management goal should be an provide high-quality bear habitat while meet- open stand of large pines and scattered 30+ year-old ing the habitat needs of species requiring more oaks or oak groves . . . [W]hile the mature pine-oak forest represents the major habitat of the fox squir- frequent fire, Maehr et al. (2001a: 43) recom- rel, this is not the only vegetation type which this mended: animal can use, nor does such prime habitat have to occur in single large units. A mosaic of habitats Ecosystem approaches to management on public with substantial pine-oak representation, large ar- lands may dictate prescribed fire regimes that are eas of edge, some open land, and access to bottom- not always optimal for bears (i.e. some public lands land seems to support squirrels as well as larger pine include timber harvesting as a primary objective, tracts. (Weigl et al. 1989: 79–80) and others are very close to urban areas). In these 12. Spatial Ecology and Restoration 387 cases, a multiple use module (MUM) approach Theoretically, a few ignitions in each state would could be applied to create natural, heterogeneous be sufficient to burn over most of the longleaf pine landscapes that also provide high quality habitat landscape. That this happened as recently as the nodes, such as old-growth forest for high quality 1880s was reported by Hough (1882), who said den sites, and altered fire regimes to provide dense that fires burned for weeks at a time over several stands of infrequently burned saw palmetto for food counties (when counties were much larger than at and cover. present). (Means 1996: 213) Under current conditions, use of prescribed Other species that either favor or require habi- fire to restore and manage longleaf pine is eas- tat mosaics in longleaf pine landscapes include ier in larger, rural landscapes than in smaller the eastern indigo snake, which prefers up- patches surrounded by intensive development land/wetland mosaics and uses gopher tortoise (Harris et al. 1996a). Red-cockaded woodpeck- burrows as refugia (Moler 1992), and vari- ers, fox squirrels, and indigo snakes all need ous amphibians that require ephemeral ponds large areas to support viable populations (Cox for breeding but live in natural pinelands as et al. 1994; Cox and Kautz 2000). In addition, adults (Huffman and Blanchard 1990; Dodd amphibians that breed in ephemeral ponds and LeClaire 1995; Palis 1996; Dodd and Cade also require extensive, intact uplands around 1998; Cox and Kautz 2000; Kautz and Cox suitable breeding ponds. Dispersal distances 2001). from natal ponds are frequently in excess of 2000 m for various species including gopher frog (Rana capito), striped newt (Notophthalmus Area and Connectivity perstriatus), and flatwoods salamander (Ashton 1992; Means et al. 1996; Palis 1996; Dodd and The former extensivity and connectivity of Cade 1998; Kautz and Cox 2001; Semlitsch longleaf pine landscapes suggests that myr- 2000). iad functional processes are dependent on Connectivity is a ubiquitous attribute of nat- these characteristics. Within natural longleaf ural landscapes. It occurs at many spatial and pine forests, large swaths of connected up- temporal scales and allows for flows of distur- lands were necessary for maintaining the high- bances, water, energy, and nutrients, as well frequency, low-intensity fire regime needed to as organisms and their genes (Harris 1984, perpetuate the system (Table 1): 1985; Forman and Godron 1986; Noss and

TABLE 1. The probability of fire and the effects of fire are scale dependent.a Probability of lightning Significance of Scale of focus ignition in a decade occurrence

Individual tree Very small Lethal (0.001 acre) Remnant patch Modest Serious (0.01–1 acre) Isolated tract High Important (10–100 acres) Forest Certainty Essential (10,000–100,000 acres)

a Habitat fragmentation has resulted in much smaller patches of longleaf pine forest that are much less likely to harbor lightning ignitions. Large, connected areas are much more likely to experience the ﬁres needed to maintain functional longleaf pine landscapes. 388 IV. Restoration

Harris 1986; Noss 1987b; Noss and Harris 1990; will inevitably lead to indirect changes in the struc- Harris and Scheck 1991; Noss and Cooper- tural content of these habitat fragments . . . [O]nly rider 1994; Harris et al. 1996b). Connectivity habitat assessments that evaluate the internal char- in the sense of intact ecotones and functional acteristics of these very large habitats in the con- juxtapositions between natural communities text of their location and surrounding environments within longleaf pine landscapes has been dis- will lead to reasonable conclusions about their adequacy. cussed above. Nevertheless, regional-scale reserve design and conservation planning must Hence, natural resource conservation and also consider the functional connectivity be- land-use planning must consider the effects of tween landscapes to maintain viable metapop- actions within their largest spatial and tem- ulations of sensitive, wide-ranging species and poral contexts (Forman 1987, 1995; Harris to provide opportunities for species to respond et al. 1996a). Landscape ecologists and con- to climate change. Long-term survival for servation biologists recognize habitat loss and many species will depend on interconnected fragmentation as the primary threats to biodi- landscapes that are sufficiently integrated to versity and many ecological processes and ser- allow functional dynamics in both time and vices (Wilcox and Murphy 1985; Harris and space (Peters and Darling 1985; Hunter et Silva-Lopez 1992; Meffe and Carroll 1997). al. 1988; Peters and Lovejoy 1992; Noss and Fragmentation occurs at multiple spatial scales, Cooperrider 1994; Harris et al. 1996a,b). Im- from the regional-scale separation of major portant, narrower corridors include riparian blocks of longleaf pine by extensive agricul- networks that serve as dwelling habitat and tural and urban development; to landscape- travel routes for many species (Harris 1988, scale fragmentation by agriculture, major 1989) and pinelands that may allow the dis- roads, clearcuts, and residential/commercial persal of red-cockaded woodpeckers (Walters development; to site-level fragmentation by et al. 1988), fox squirrels (Weigl et al. 1989), forest roads, trails, and plowed fire lines. and other species. Larger landscape linkages The understanding and maintenance of critical would ideally be wide enough to encompass landscape functions, including interactions be- a diversity of habitat types, including com- tween natural communities in time and space, plete topographic gradients, and to allow sen- are central considerations in landscape ecology sitive wildlife species to travel undisturbed by and are essential for long-term monitoring and human activities (Harris and Gallagher 1989; conservation of biodiversity. Fragmentation Noss and Harris 1990; Harris and Atkins 1991; ultimately leads to landscape dysfunction and Harris and Scheck 1991; Noss 1992, 1993; Noss the erosion of biodiversity; thus, strategies to and Cooperrider 1994; Harris et al. 1996b). foster landscape heterogeneity and connectivity, and maintain and/or restore critical landscape processes, are essential to both landscape Landscape-Level Strategies, viability and biodiversity conservation (Harris Opportunities, and Challenges 1984; Weigl et al. 1989; Turner et al. 1995; Harris et al. 1996a; Gordon et al. 1997; Poiani One of the primary lessons of landscape ecol- et al. 2000). Such comprehensive landscape ogy is that spatial context matters (Harris planning could maximize extensivity and con- 1984; Harris et al. 1996a). Harris and Kangas nectivity by linking reserves and multiple-use (1988:141–142) recognized the significance of conservation lands into functional networks. contextual challenges for restoring and man- The goal is to protect and manage a land- aging biodiversity: scape to preserve and/or mimic natural pro- As humans modify land use—and thus the compo- cesses and evolutionary forces. Only in such a sition of landscapes—the matrix and context of indi- landscape can we expect to conserve the re- vidual habitats are changed. Regardless of whether maining biodiversity and functional processes the content of the surviving habitat fragments is that will allow further evolution and adapta- directly altered, changes in the contextual setting tion (Harris et al. 1996a). Its spatial ecology 12. Spatial Ecology and Restoration 389 demands a restoration and conservation strat- old growth, connectivity among patches, in- egy that will return longleaf pine communi- tact ecotones, and functional environmental ties to their role as the matrix ecosystem. Such and community gradients. a strategy must include efforts to restore and The restoration and maintenance of viable maintain intact ecotones, functional juxtapo- metapopulations of longleaf-associated animal sitions of uplands and wetlands needed by species (e.g., fox squirrel, red-cockaded wood- many species, and the spatial extent and con- pecker) within connected networks of core nectivity needed by wide-ranging and other protected areas on public conservation lands, species adapted to longleaf pine forests. Har- augmented by multiple-use public and pri- ris et al. (1996a: 341) describe a vision for lon- vate lands managed to contribute to ecolog- gleaf pine landscape restoration that is relevant ical integrity, is an overarching goal (Harris here: 1984; Noss and Harris 1986; Noss 1987b; Noss and Cooperrider 1994; Harris et al. [S]imply saving some islands of old-growth longleaf 1996b; Hoctor et al. 2000). Such networks pine is not the issue; saving the red-cockaded wood- provide the opportunity to reestablish land- pecker is not the issue; and providing an intercon- scapes where fragmentation is minimized and nected habitat system that protects both the longleaf natural connectivity, habitat juxtapositions, pine and the red-cockaded woodpecker should not be the issue. Rather, we see the issue as being that of ecological gradients, and spatial and temporal restoring and maintaining a spatially integrated lon- heterogeneity are enhanced (Harris 1984; Noss gleaf pine ecosystem that can and will maintain the and Harris 1986; Noss and Cooperrider 1994; full suite of landscape ecological processes includ- Harris et al. 1996a; Soule´ and Terborgh 1999; ing fire that is ignited in one place but allowed to Margules and Pressey 2000; Poiani et al. 2000). disperse across the system; a system that can with- How do we restore longleaf pine landscapes stand the effects of major hurricanes and still re- starting from the current situation of primar- main viable and resilient because of its extensive ily small, isolated tracts of longleaf, virtually nature; a system that is capable of sustaining nat- all of which are degraded to some degree? Al- ural outbreaks of beetles and fungi; and a system though urbanization is increasing rapidly in that is interdigitated with other community types parts of the SECP, rural land uses and forestry that provide seasonally important services for the longleaf pine community and vice versa. still dominate. Even in Florida, 80% of the state is currently classified as rural. Therefore, the protection and restoration of functional networks of longleaf pine and associated com- Landscape Conservation munities is still feasible. Significant opportu- Strategies nities to restore and maintain longleaf pine landscape-types exist on large public lands, Given the current status of longleaf pine com- including national forests and military lands. munities as remnant islands in a sea of in- In fact, large-scale restoration to longleaf pine tensive forestry, hardwood stands that devel- and increased use of prescribed fire is under- oped after fire suppression, agriculture, and, way on various public conservation lands and increasingly, urban areas, opportunities to re- military installations, including Eglin Air Force store and manage at landscape scales are rare Base in northwest Florida (Gordon et al. 1997). and probably will be fleeting. However, aware- Even the national forests, where longleaf pines ness of longleaf pine biology combined with were clear cut and replaced with plantations of planning guidelines from conservation biology slash, loblolly, and sand pines, have reduced and landscape ecology can be used as a sound large-scale logging, have moved toward more basis for restoring ecological processes in lon- uneven-aged management, and are now re- gleaf pine landscapes that will support native planting longleaf or allowing natural regener- biodiversity. Overall, regional-scale restoration ation on many sites. and management of longleaf pine landscapes Restoration of functional longleaf pine must include reestablishing large patches of landscapes must include consideration of 390 IV. Restoration between-stand characteristics, including man- the scale of tens to hundreds of thousands of agement for old-growth longleaf within forests hectares, in addition to the more typical site- also managed for timber production (Harris level restoration projects. Although ecologi- 1984). Reestablishment of longleaf pine diver- cal function might be restored on single, very sity within multiple-use landscapes should in- large public land holdings, regional landscape clude rehabilitation of sandhills and clayhills restoration requires design and management using prescribed fire, replanting of longleaf on across ownerships, which in turn will require sites that originally supported longleaf, and multiagency and public–private cooperation restoration of longleaf on other selected sites, (Harris 1984; Noss and Harris 1986; Noss and including bottomland ecotones. Cooperrider 1994). Context management also The importance of considering ecological requires consideration of adjacent land uses context for restoring and maintaining func- and the utilization of buffer-zone principles to tional landscapes cannot be overstated. Ecol- encourage compatible land uses that can pro- ogy has long suffered from a focus on within- vide habitat for wide-ranging species, facilitate system attributes and functions (Harris et al. critical ecological fluxes, minimize external 1996a). Landscape ecology has made clear threats and negative edge effects, and poten- that multiscale spatial patterns and processes tially provide functional connectivity among across heterogeneous landscapes control key landscapes. Such context management also is ecological functions. Another crucial design critical for restoring landscapes that can sup- principle for longleaf pine landscapes is the port more natural fire regimes and other land- restoration of functional ecotones and habitat scape processes: juxtapositions between uplands dominated by longleaf and associated smaller patch upland The prospect of restoring an interconnected land- communities with wetlands including bogs, scape system of conservation areas where ecologi- marshes, shrub swamps, bottomland forests, cal processes such as naturally ignited and naturally and forested swamps. Fires should be allowed propagated fire exist (that is, a LET-BURN POLICY) to burn into wetland edges to reestablish eco- appears imminent but problematic in at least certain tonal communities shaped by the interaction states. These problems may be overcome by managing entire landscapes to emphasize compatibility of fire and flood and to restore breeding habi- of adjacent land uses in a manner so that conserva- tat for the myriad plant, invertebrate, and tion lands will be shielded from the negative impacts amphibian species characteristic of intact lon- of adjacent intensive development, and developed gleaf pine landscapes (Noss 1988; Huffman and areas will be shielded from the potential negative Blanchard 1990; Noss and Harris 1990; Means impacts of conservation land management. (Harris 1996; Palis 1996; Dodd and Cade 1998; Platt et al. 1996a:334) 1999; Cox and Kautz 2000; Kautz and Cox 2001). Interactions, juxtaposition, and inter- Therefore, protection and restoration coordi- digitation with forested wetlands are essen- nated with private partners will be essential. tial characteristics of longleaf pine landscapes, Red-cockaded woodpeckers are the flagship and wetlands provide key functions including species needing inclusion of, and incentive mast production, drought refugia, and breed- for, private landowners to restore longleaf ing habitat (Clewell et al. 1982; Hart 1984; pine landscapes. One project, discussed below, Harris and Vickers 1984; Vickers et al. 1985; involves efforts to reestablish red-cockaded Gross 1987; Harris 1988, 1989; Forman 1995). woodpeckers on lands around military bases Restoration and management of ecologi- to increase the viability of regional populations cal context require spatially extensive ap- and reduce the management burden that cur- proaches. We again emphasize that, in the rently lies almost exclusively within military case of longleaf pine, restoration of context installations. means reestablishing the role of longleaf pine Opportunities for restoring connectiv- as the landscape matrix for other commu- ity among widely separated longleaf pine nities. This will often require restoration on landscapes are particularly challenging. 12. Spatial Ecology and Restoration 391

Nevertheless, current land uses could facilitate Gordon et al. 1997). The DOD has developed the restoration of longleaf pine across larger integrated natural resource management plans areas over time. that include restoration of longleaf communities. One major effort involves Eglin Air Force Base in the Florida Panhandle, which seeks to Opportunities restore extensive longleaf pine forests through prescribed fire and other efforts “utilizing in- A key fact is that there are still millions of tegrated natural resources management and hectares of forested land in the SECP.Although principles of ecosystem management to ensure much of this forest has been converted to pine ecosystem viability and biodiversity while pro- plantation or otherwise degraded by fire sup- viding compatible multiple uses” (Means 1996: pression and other significant changes in eco- 219). logical processes (Ware et al. 1993), emerg- The DOD has also been working with part- ing trends toward landscape and regional- ners in the North Carolina sandhills around scale conservation planning, protection, and Fort Bragg and other project areas to protect restoration provide significant opportunities additional habitat for federally listed species to reestablish longleaf pine as the dominant (especially the red-cockaded woodpecker), matrix vegetation. The following examples of connect existing conservation lands, and re- conservation initiatives include landscape and duce the potential for further urban encroach- regional-scale projects and other programs that ment around bases that could interfere with may facilitate restoration of longleaf pine land- military training operations (Goodison and scapes: Hoctor 2003). These efforts will now be facilitated by recent authorization for DOD to work with partners to protect lands near military Cross-Florida Greenway bases (Robert Lozar personal communication). This conservation area is an approximately An additional project involves cooperation 190-km-long corridor of more than 40,000 ha between DOD, the state of Florida, The Na- that runs from the Florida Gulf Coast north ture Conservancy, the University of Florida, of Tampa to the St. Johns River in northeast- and private landowners (e.g., the Nokuse Plan- ern Florida. Conservation efforts include the tation) to protect a 90-km landscape linkage restoration of remaining longleaf pine forests between the Apalachicola National Forest and degraded by fire suppression and replanting Eglin Air Force Base in northwest Florida. This of longleaf pine on former agricultural lands. landscape linkage will protect existing lon- Ecotonal communities including degraded red gleaf pine forests and provide opportunities to oak woodlands are also being restored. One of restore longleaf pine while protecting an es- the primary goals of this project is to re-create sential flight-training corridor for DOD. This a continuous corridor of longleaf pine and re- project illustrates the concept of “sharing the lated natural communities from the Ocala Na- burden” in the management of area-sensitive tional Forest to the Gulf Coast and to pro- species such as the red-cockaded woodpecker tect and restore populations of Sherman’s fox and, more importantly, in the restoration of in- squirrel (Sciurus niger shermani ), gopher tor- tact landscapes that will require public–private toises, indigo snakes, red-cockaded woodpeck- cooperation on a regional scale. ers, and many other species dependent on longleaf pine landscapes (Harris and Hoctor 1992). The Nature Conservancy’s Department of Defense Projects Ecoregional Planning U.S. Department of Defense (DOD) installa- The Nature Conservancy has developed tions contain some of the most significant ecoregion-based biodiversity plans across the longleaf pine forest remaining (Means 1996; range of longleaf pine. Ecoregional planning 392 IV. Restoration is a biodiversity assessment effort attempting EPA Southeastern Ecological to comprehensively identify all sites needed Framework to conserve biodiversity within all ecoregions in the United States (Groves et al. 2000, The EPA Southeastern Ecological Framework 2002). The plans for ecoregions in the south- is a cooperative effort between EPA Region 4 eastern United States identify various areas and the University of Florida that was delin- for protecting existing longleaf pine sites and eated using a GIS-based analysis to identify pri- landscape-scale restoration opportunities to ority land-conservation areas across an eight- restore longleaf pine as a matrix community. state region including several states within the In the Peninsular Florida Ecoregion, sites ca- range of longleaf pine: North Carolina, South pable of supporting potential matrix-quality Carolina, Georgia, Florida, Alabama, and Mis- longleaf pine landscapes were defined as hav- sissippi. The framework is meant to encour- ing at least 2000 ha of sandhill or flatwoods age federal and state interagency cooperation cover. Twenty-seven such sites were identified and coordinate protecting and restoring land- for flatwoods and eight for sandhills within scapes efficiently (Hoctor et al. 2002a). As with the ecoregion (Hilsenbeck et al. 2001; Hoctor the Florida Ecological Network, millions of 2003). hectares of pineland are included, which could provide significant opportunities for longleaf The Florida Ecological Network pine restoration (Fig. 6). The Florida Ecological Network is part of the Forest Certification/Landowner Florida Greenways Program administered by Programs the Florida Department of Environmental Pro- tection. The Florida Ecological Network was Forest certification is a developing, incentives- developed using a GIS-based regional land- based strategy to facilitate ecologically sound scape analysis to delineate the best opportu- management of forest resources. Forest nities to protect large, connected landscapes certification involves an independent, third- across the state (Hoctor et al. 2000; Hoctor party assessment of field-level forest manage- 2003). Existing longleaf pine communities and ment practices against specified social, ecologi- focal species dependent on longleaf pine were cal, and economic standards (http://www.sfrc. included as priorities in the analysis (Hoctor et ufl.edu/Extension/ffws/fc.htm#ctcs). In the al. 2001, 2002b; Kautz and Cox 2001). Other Southeast, forest certification is beginning to areas of degraded pineland cover that could be used to restore longleaf pine on private contribute to protecting large landscapes also lands that have been converted to other were included, and such areas may provide sig- uses or other tree species. The success of nificant opportunities for longleaf pine restora- certification will be based on consumer de- tion in the future. An implementation plan mand for wood products that are certified as has been developed for the Florida Green- being produced using management practices ways Program that includes the identification that promote healthy forests. If successful, of Critical Linkages within the Florida Ecologi- forest certification could provide significant cal Network (Hoctor et al. 2002b). These Criti- economic incentives for private landowners cal Linkages are now the primary focus of land- to restore longleaf pine forests. However, scape protection opportunities in Florida and whether sufficient demand exists is unknown. include projects relevant to longleaf pine con- Also, restoration of functional longleaf pine servation, including the Apalachicola National landscapes will require the restoration and Forest–Eglin Air Force Base project mentioned better management of very large areas. above and a 60,000-ha landscape linkage be- Whether forest certification will produce tween the Camp Blanding military training sufficient incentive to restore hundreds of site and Osceola National Forest in northeast thousands of hectares or more remains to be Florida. seen. 12. Spatial Ecology and Restoration 393

FIGURE 6. The EPA Southeastern Ecological Framework (SEF) covers almost half of EPA Region 4. Over 80% of the SEF is forested and approximately 25% is within existing conservation lands or open water. Although much of the forest cover on private land is intensively managed, the SEF suggests that there are still very good opportunities to restore large longleaf pine-dominated landscapes across much of its historic range.

Challenges vationists will likely face the need to transplant many species to new sites in desperate efforts Climate Change to avoid extinctions. Such efforts, including the reestablishment of functional natural commu- Although connected networks of public and nities and landscapes, will obviously be diffi- private conservation lands may provide some cult (Peters and Lovejoy 1992). opportunity for species to respond to climate change, the rate and magnitude of climate change combined with the current level Water Mining and Aquifer of landscape-scale habitat fragmentation may likely exceed the capacity of many species to Impacts adapt (Peters and Lovejoy 1992). In coastal ar- In Florida most water for drinking, agriculture, eas, especially in very low elevations such as and industry comes from the Floridan aquifer. Florida, sea-level rise could inundate hundreds Greatly increased demand combined with in- of thousands of hectares. In the future, conser- creasing drought have resulted in significantly 394 IV. Restoration lower aquifer levels in recent years (Bacchus could be accommodated by infrastructure im- 2000). Though longleaf pine is adapted to tol- provements and direct growth to areas most erate xeric conditions, the lack of recruitment suitable (Nicolas and Steiner 2000). Good com- and the death of mature trees during pro- prehensive plans would limit sprawling devel- longed droughts may be a significant concern opment and would give land acquisition ef- (Richard Franz personal communication). In forts more time to protect lands critical for addition, aquifer drawdowns can result in sig- biodiversity (Hoctor 2003). Finally, good plan- nificantly reduced hydroperiods (or conver- ning also is needed to contribute to the core sion to uplands) of ephemeral ponds and other area-buffer model of reserve design, where wetlands (Bacchus 2000). Therefore, contin- intensive development is separated from bio- ued lowering of aquifer levels especially in diversity reserves by rural lands including silvi- combination with drought could seriously im- culture, agriculture, and other uses more compact the ecological integrity of remaining lon- patible with conservation objectives (Harris gleaf pine landscapes and significantly hinder 1984; Noss and Harris 1986; Soule´ 1991). restoration. However, it is now widely acknowledged that growth management in Florida has largely failed to stop sprawling development (Nicolas Funding for Land Conservation, and Steiner 2000; Hoctor 2003). With signifi- Growth Management, cant private property rights issues and political and Regional Conservation leanings that currently impede regional conservation planning (Hocter et al. 2004), efforts Planning to enact or enforce effective growth manage- Florida has spent over 4 billion dollars to ac- ment will be challenging. quire conservation lands and establish con- Nevertheless, there is a growing recogni- servation easements on ecologically significant tion of the essential ecological services that private lands over the past 14 years. Never- natural landscapes provide to humans such theless, even with Florida’s existing commit- as clean drinking water, storm water man- ment, millions of unprotected hectares impor- agement, flood control, particulate matter re- tant for conservation remain (Hoctor et al. moval, and carbon sequestration, as well as 2000; Hoctor 2003). Though land acquisi- food and shelter for native species (Daily 1997, tion for conservation purposes is generally 2000; Benedict and McMahon 2001). One pos- popular, and various local and state govern- sibility for restoring longleaf pine landscapes ments and the federal government are spend- is the development of a “landscape utility” ing significant amounts on land conserva- concept that recognizes the services provided tion, needs greatly exceed available funds by intact ecosystems and compensates pri- (Hoctor et al. 2004). In Florida and other sun- vate landowners for ecologically sound man- belt states, development induced by rapidly agement that provided such services or pur- growing human populations will continue to chases land from them for public use (Maynard erode the available private land base that Hiss personal communication). Florida’s water could support the restoration of longleaf pine management districts already purchase lands landscapes. or acquire conservation easements to ensure Florida enacted growth management leg- that critical hydrological functions are pro- islation in the 1970s and 1980s in response tected or restored. Broader representation of to rapid development. This legislation cre- ecosystem services (including carbon seques- ated state oversight for local future land-use tration) that could be funded both through plans and large developments with regional mitigation funds and utility taxes (Maynard impacts. Drafters of the legislation felt that Hiss personal communication) would be chal- state oversight would expose the link between lenging but could provide a significant tool growth and infrastructure costs and conse- for restoring and managing longleaf pine land- quently might control growth in a manner that scapes. 12. Spatial Ecology and Restoration 395 Wildland–Urban Interface The Inverse of Ecological Related to issues regarding increasing develop- Networks: Road Networks ment and urban/suburban sprawl is the con- Highway and other linear transportation comitant increase in the wildland–urban in- projects (such as proposed high-speed train terface, which can be defined as “an area routes) are a prominent threat to biodiver- where increased human influence and land- sity (Trombulak and Frissell 2000). Existing use conversion are changing natural resource highways need to be retrofitted to facilitate goods, services, and management” (Macie and the movement of focal species and to reduce Hermansen 2002:2). Although the threat of roadkills (Hoctor et al. 2000; Hoctor 2003; wild fire ideally may be used as a tool for pro- Smith 2003). Furthermore, systemwide plan- moting prescribed burning to restore and man- ning should be conducted to avoid ecolog- age longleaf pine communities in the urban– ically sensitive areas (especially large, intact wildland interface, efforts to use prescribed landscapes and corridors) when planning new fire, especially in the growing season, are hin- transportation routes, and to enhance efforts dered by concerns about smoke drift and as- to minimize and mitigate impacts that may sociated declines in air quality and highway be unavoidable. Currently, such efforts are safety. As the public becomes increasingly ur- usually limited to individual projects, where ban and unfamiliar with the natural role of politics, road design constraints, and limited fire in southeastern ecosystems, the difficulty budgets are frequently invoked to avoid spend- of implementing prescribed burning is bound ing sufficient money and time to truly miti- to increase. Other important issues within the gate ecological impacts. Systemwide planning wildland–urban interface include continued should ensure that an appropriate budget is habitat loss and increasing habitat fragmenta- developed to minimize and mitigate impacts. tion and increased demands for resource-based State departments of transportation typically recreation activities (Macie and Hermansen have very large budgets (the Florida Depart- 2002). ment of Transportation’s budget for infrastructure improvements between 2001 and 2020 is $108 billion; Florida Department of Trans- Making Longleaf Pine a portation 2001). Even if only 5% of these bud- Competitive Economic Benefit gets could be used to counteract the ecological costs of roads by building bridge spans, widen- on Private Lands ing bridges, building wildlife underpasses and Efforts to restore and maintain longleaf pine overpasses, and protecting large areas of habi- landscapes would clearly benefit if longleaf was tat, biodiversity could be protected while considered an economic resource competitive transportation capacity is increased (Smith with other uses of land, or at least if tax ben- 1999; Hoctor et al. 2000; Hoctor 2003; Smith efits or other economic returns were provided 2003). (Macie and Hermansen 2002). Forest certification may promote this situation, especially in large rural landscapes. However, whether Conclusions longleaf pine communities can compete with the economic gains from development is ques- Restoring the great longleaf pine forest of the tionable, especially as land values continue to Southeast is, essentially, an effort to re-create rise. Paying landowners for providing ecolog- G. Evelyn Hutchinson’s ecological theater and ical services such as carbon sequestration and evolutionary play (Hutchinson 1965) in one aquifer recharge could help increase the value of the most biologically outstanding regions of of longleaf pine beyond selective timber har- the world. Saving a few patches of longleaf vesting, hunting rights, or other traditional pine and focusing attention on a few legally economic uses. protected species—the course we have been on 396 IV. Restoration in conservation planning for several decades should go beyond the minimalist goals of ex- in the region—is not sufficient. As of 1993, isting species recovery plans toward the goal there were 27 federally listed species and 99 of restoring viable populations and metapop- candidates for listing that were closely associ- ulations of longleaf-associated species across ated with or dependent on longleaf pine and their historic distributions. In most cases, this related (i.e., wiregrass) ecosystems (Noss et al. involves protecting and restoring large tracts 1995). Recently, many of the candidate species of old-growth longleaf pine that are well have been removed from the list for essentially connected at local, landscape, and regional political reasons (i.e., misplacement of the bur- scales. r den of proof; Noss et al. 1997), but the status Longleaf-associated species that are not cur- of biologically imperiled species generally con- rently listed under the U.S. Endangered tinues to decline. Species Act, but which are declining and What is needed is a multifaceted and com- vulnerable to existing and future threats, prehensive program to protect, restore, and should be identified and used as focal species manage longleaf pine landscapes. Such a pro- (e.g., Lambeck 1997) for conservation plan- gram has several essential components: ning. Modeling of habitat and population viability of focal species addresses specific issues of habitat area, distribution, and con- r A detailed gap analysis of all the variants of figuration (including connectivity) that are longleaf pine communities across the range poorly addressed by conventional conserva- of the species should be performed to assess tion assessments. For example, “area-limited how well each type is represented in var- species” are those that occur in low densities ious categories of conservation areas. This or require large areas, and “dispersal-limited assessment should include the associations species” include those that are sensitive to described under the Longleaf Pine Alliance barriers or sources of mortality, such as roads, in the National Vegetation Classification Sys- when attempting to move across the land- tem (Anderson et al. 1998, Grossman et al. scape. Lambeck (1997) hypothesized that 1998). However, for a more comprehen- the most sensitive species in these groups sive assessment of representation needs, a may serve as effective umbrellas for other rangewide map of existing longleaf pine species with similar vulnerabilities. A suite cover should be overlaid on a map of abi- of such species, intelligently selected, can otic habitats, i.e., defined by soils, elevation, serve as focal species for conservation pro- geologic substrate, and so on. grams on landscape to regional scales (Lam- r An interorganizational recovery team, in- beck 1997; Carroll et al. 2001; Noss et volving government and private partners, al. 2002). An example of a potential focal should be created to develop a Longleaf Pine species closely associated with longleaf pine Ecosystem Recovery Plan. Such a plan must is the fox squirrel. In general, the south- incorporate goals to restore the full range of eastern subspecies of fox squirrels appear longleaf pine community variation (i.e., us- to be long-lived, with low adult mortal- ing the results of the gap analysis described ity and few, small litters each year. These above), along with fire and other natural dis- life history characteristics may explain why turbances within historic ranges of variabil- most populations of southeastern fox squir- ity in space, time, and other attributes. The rel are declining and have failed to recover Ecosystem Recovery Plan would not replace even after preservation of potential habitats the need for more detailed, species-level (Tappe and Guynn 1998). Other potential plans; in fact, there is evidence that ecosys- focal species that currently receive no pro- tem plans produced to date provide less tection or planning under the Endangered assurance of species recovery than single- Species Act include several upland snakes species plans (Clark et al. 2002). To the ex- such as the eastern diamondback rattlesnake tent feasible, the Ecosystem Recovery Plan (Crotalus adamanteus), pine snake (Pituophis 12. Spatial Ecology and Restoration 397

melanoleucus), and kingsnake (Lampropeltis Bacchus, S. 2000. Uncalculated impacts of unsus- getula). tainable aquifer yield including evidence of sub- r The longleaf pine ecosystem must be con- surface interbasin flow. J Am Water Resour Assoc sidered together with other vegetation types 36(3):457–481. in the region. As noted previously, ecotones Barry, J.M. 1980. Natural Vegetation of South Carolina. are sites of high diversity in the longleaf pine Columbia: University of South Carolina Press. Beckett, S., and Golden, M.S. 1982. Forest vegeta- landscape, and these extend from the mon- tion and vascular flora of Reed Brake Research tane peaks in the extreme north of the re- Natural Area, Alabama. Castanea 47(4):386–392. gion to the low swamps and coastal marshes. Benedict, M.A., and McMahon, E.T. 2001. Green in- Ideally, the entire complex of environmental frastructure: Smart conservation for the 21st cen- gradients would be restored to intactness tury. Sprawl Watch Clearinghouse Monograph across much of the region. When planning Series, Washington, DC. across vegetation types and other gradi- Boerner, R.E.J., and Kooser, J.G. 1989. Leaf litter ents, wide-ranging species that utilize mul- redistribution among forest patches within an Al- tiple habitats, for example the Florida pan- legheny Plateau watershed. Landscape Ecol 2:81– ther/eastern cougar and black bear, are ideal 92. focal species (Maehr et al. 2001a,b). Impor- Butler, C. 1998. Treasures of the Longleaf Pines: Naval Stores. Shalimar, FL: Tarkel Publishing. tantly, we urge the active reestablishment Carroll, C., Noss, R.F., and Paquet, P.C. 2001. Car- of Florida panther populations—and their nivores as focal species for conservation planning connection into functional, self-sustaining in the Rocky Mountain region. Ecol Appl 11:961– metapopulations—across the historic distri- 980. bution of the subspecies in the Southeast. Clark, J.A., Hoekstra, J.M., Boersma, P.D., and r We emphasize the importance of “sharing Kareiva, P. 2002. Improving U.S. Endangered the burden” in the management of area- Species Act recovery plans: Key findings and rec- limited species such as the red-cockaded ommendations of the SCB Recovery Plan Project. woodpecker and fox squirrel, as well as in Conserv Biol 16:1510–1519. the restoration and management of longleaf Clewell, A.F. 1981. Natural setting and vegetation pine communities generally. A fully restored of the Florida Panhandle. USACE, Contract No. DACWO1-77-C-0104. Mobile, AL. and intact landscape requires public–private Clewell, A.F., Goolsby, J.A., and Shuey, A.G. 1982. cooperation across counties and states. Riverine forests of the South Prong Alafia River System, Florida. Wetlands 2:21–72. References Cox, J., and Kautz, R. 2000. Habitat conservation needs of rare and imperiled wildlife in Florida. Of- Adams, A.L. 1976. The Fourth Quarter, autobiogra- fice of Environmental Services, Florida Fish and phy by same (as told to Tom Dunkin). Published Wildlife Conservation Commission, Tallahassee. privately 379 pp. Cox, J., Kautz, R., MacLaughlin, M., and Gilbert, Anderson, M., Bourgeron, P.,Bryer, M.T., Crawford, T. 1994. Closing the Gaps in Florida’s Wildlife Habitat R., Engelking, L., Faber-Langendoen, D., Gally- Conservation System. Tallahassee: Florida Game and oun, M., Goodin, K., Grossman, D.H., Landaal, Freshwater Fish Commission. S., Metzler, K., Patterson, K.D., Pyne, M., Reid, Daily, G.C., ed. 1997. Introduction: What are M., Sneddon, L., and Weakley, A.S. 1998. Inter- ecosystem services? Nature’s Services: Societal De- national Classification of Ecological Communities: Ter- pendence on Natural Ecosystems. Washington, DC: restrial Vegetation of the United States. Volume I. The Island Press. National Vegetation Classification System: List of Types. Daily, G.C. 2000. Management objectives for the Arlington, VA: The Nature Conservancy. protection of ecosystem services. Environ Sci Policy Anon. 1987. Pensacola, the Lumbering Era (1887). 3(6):333–339. Pensacola, FL: The John Appleyard Agency, Inc. Delcourt, H., and Delcourt, P. 1977. Preset- Ashton, R.E., Jr. 1992. Flatwoods salamander. In tlement magnolia-beech climax of the Gulf Rare and Endangered Biota of Florida, Vol. 3, Amphib- Coastal Plain: Quantitative evidence from the ians and Reptiles, 2nd Edition, ed. P.E. Moler, pp. Apalachicola River Bluffs, north-central Florida. 39–43. Gainesville: University Press of Florida. Ecology 58:1085–1093. 398 IV. Restoration

Delcourt, P.A., Delcourt, H.R., Morse, D.F., and Goodin, K., Landaal, S., Metzler, K., Patterson, K. Morse, P.A. 1993. History, evolution, and organi- D., Pyne, M., Reid, M., and Sneddon, L. 1998. zation of vegetation and human culture. In Biodi- International Classification of Ecological Communities: versity of the Southeastern United States: Lowland Ter- Terrestrial Vegetation of the United States. Volume II. restrial Communities, eds. W.H. Martin, S.G. Boyce, The National Vegetation Classification System: Develop- and A.C. Echternach, pp. 47–79. New York: John ment, Status, and Applications. Arlington, VA: The Wiley & Sons. Nature Conservancy. Dodd, C.K., Jr., and Cade, B.S. 1998. Movement pat- Groves, C., Valutis, L., Vosick, D., Neely, B., terns and the conservation of amphibians breed- Wheaton, K., Touval, J., and Runnels, B. 2000. ing in small, temporary wetlands. Conserv Biol Designing a Geography of Hope: A Practitioner’s Hand- 12:331–339. book for Ecoregional Conservation Planning. Arling- Dodd, C.K., Jr., and LeClaire, L.V. 1995. Biogeogra- ton, VA: The Nature Conservancy. phy and status of the striped newt (Notophthalmus Groves, C.R., Jensen, D.B., Valutis, L.L., Redford, perstriatus) in Georgia, USA. Herpetol Nat Hist 3:37– K.H., Shaffer, M.L., Scott, J.M., Baumgartner, 46. J.F., Higgins, J. V., Beck, M.W., and Anderson, Duever, L.C. 1983. Natural communities of Florida’s M.G. 2002. Planning for biodiversity conserva- inland sand ridges. Palmetto. Winter Park, FL: tion: Putting conservation science into practice. Florida Native Plant Society 3(3):1–3, 10. BioScience 52:499–512. Elliott, S.B. 1912. The Important Timber Trees of the Harper, R.M. 1911. The relation of climax vegeta- United States. Boston: Houghton Mifflin. tion to islands and peninsulas. Bull Torrey Bot Club. Florida Department of Transportation. 2001. 2020 38:515–525. Florida transportation plan. Florida Department Harper, R.M. 1915. Vegetation types. In Natural of Transportation, Tallahassee. Resources of an Area in Central Florida. Florida Forman, R.T.T. 1987. The ethics of isolation, the Geological Survey, 13th Annual Report, eds. E.H. spread of disturbance, and landscape ecology. In Sellards, R.M. Harper, E.N. Mooney, W.J. Latimer, Landscape Heterogeneity and Disturbance, ed. M.G. H. Gunter, and E. Gunter, pp. 135–188. Turner, pp. 213–229. New York: Springer-Verlag. Harris, L.D. 1980. Forest and wildlife dynamics in Forman, R.T.T. 1995. Land Mosaics: The Ecology of the Southeast. Trans North Am Wildl Nat Resourc Landscapes and Regions. London: Cambridge Uni- Conf. 45:307–322. versity Press. Harris, L.D. 1984. The Fragmented Forest: Island Bio- Forman, R.T.T., and Godron, M. 1986. Landscape geography Theory and the Preservation of Biotic Diver- Ecology. New York: John Wiley & Sons. sity. Chicago: University of Chicago Press. Frost, C. 1993. Four centuries of changing landscape Harris, L.D. 1985. Conservation corridors: A high- patterns in the longleaf pine ecosystem. Proc Tall way system for wildlife. ENFO Report 85-5, Timbers Fire Ecol Conf 18:17–44. Florida Conservation Foundation, Winter Park. Goodison, C., and Hoctor, T.S. 2003. Defining the Harris, L.D. 1988. The nature of cumulative impacts Southeastern Ecological Framework for Military on biotic diversity of wetland vertebrates. Environ Installations. Report to the U.S. Army Corps Manage 12:675–693. of Engineers Construction Engineering Research Harris, L.D. 1989. The faunal significance of frag- Laboratory, Champaign, IL. mentation of southeastern bottomland forests. In Gordon, D.R., Provencher, L., and Hardesty, J.L. Proceedings of the Symposium: The Forested Wetlands of 1997. Measurement scales and ecosystem man- the Southern United States, Orlando, FL. USDA For- agement. In The Ecological Basis of Conservation: Het- est Service, General Technical Report SE-50, pp. erogeneity, Ecosystems, and Biodiversity, eds. S.T. A. 126–134. Asheville, NC. Pickett, R.S. Ostfeld, M. Shachak, and G.E. Likens, Harris, L.D. 1999. Remembering Florida’s ancient pp. 262–273. New York: Chapman & Hall. forests: Old Florida in words and pictures. Florida Greeley, W.B. 1920. Forester. USDA report to U.S. Defenders of the Environment Special Bulletin, Senate. Gainesville, FL, Summer 1999. Gross, F.E.H. 1987. Characteristics of small stream Harris, L.D., and Atkins, K. 1991. Faunal movement floodplain ecosystems in north and central corridors in Florida. In Landscape Linkages and Bio- Florida. Master’s thesis, University of Florida, diversity, ed. W. E. Hudson, pp. 117–134. Defend- Gainesville. ers of Wildlife. Washington, DC: Island Press. Grossman, D.H., Faber-Langendoen, D., Weakley, Harris, L.D., and Gallagher, P.B. 1989. New ini- A.S., Anderson, M., Bourgeron, P., Crawford, R., tiatives for wildlife conservation: The need for 12. Spatial Ecology and Restoration 399

movement corridors. In Defense of Wildlife: Pre- Hilsenbeck, R., Hoctor, T., Goodison, C., Hernan- serving Communities and Corridors, ed. G. Macin- dez, P., Caster, W., and Hernandez, J. 2001. The tosh, pp. 12–34. Washington, DC: Defenders of Peninsular Florida Ecoregional Plan. The Florida Wildlife. Chapter of The Nature Conservancy, Tallahassee. Harris, L.D., and Hoctor, T.S. 1992. Cross Florida Hoctor, T.S. 2003. Regional landscape analysis and Greenbelt State Recreation and Conservation reserve design to conserve Florida’s biodiver- Area Management Plan. Volume IV: Report on Bi- sity. Ph.D. dissertation, University of Florida, ological Issues. University of Florida, Gainesville. Gainesville. Harris, L.D., and Kangas, P. 1988. Reconsideration Hoctor, T.S., Carr, M.H., and Zwick, P.D. 2000. Iden- of the Habitat Concept. Trans North Am Wildl Nat tifying a linked reserve system using a regional Resour Conf 53:137–144. landscape approach: The Florida ecological net- Harris, L.D., and Scheck, J. 1991. From implications work. Conserv Biol 14:984–1000. to applications: The dispersal corridor approach to Hoctor, T.S., Teisinger, J., Carr, M.H., and Zwick, P.D. the conservation of biological diversity. In Nature 2001. Ecological Greenways Network Prioritiza- Conservation 2: The Role of Corridors, eds. D.A. Saun- tion for the State of Florida. Final Report. Office ders and R. J. Hobbs, pp. 189–220. Chipping Nor- of Greenways and Trails, Florida Department of ton, New South Wales, Australia: Surrey Beatty Environmental Protection, Tallahassee. and Sons. Hoctor, T.S., Carr, M.H., Goodison, C., Zwick, P.D., Harris, L.D., and Silva-Lopez, G. 1992. Forest frag- Green, J., Hernandez, P.,McCain, C., Whitney, K., mentation and the conservation of biological di- and Teisinger, J. 2002a. Final report: Southeastern versity. In Conservation Biology: The Theory and Ecological Framework. Environmental Protection Practice of Nature Conservation, eds. P. Fielder and Agency Region 4, Atlanta, GA. S. Jain, pp. 197–237. New York: Chapman & Hoctor, T.S., Teisinger, J., Carr, M.H., and Zwick, P.D. Hall. 2002b. Identification of Critical Linkages Within Harris, L.D., and Vickers, C.R. 1984. Some fau- the Florida Ecological Greenways Network. Final nal community characteristics of cypress ponds Report. Office of Greenways and Trails, Florida and changes induced by perturbation. In Cypress Department of Environmental Protection, Talla- Swamps, eds. K. C. Ewel and H. T. Odum, pp. 171– hassee. 185. Gainesville: University Press of Florida. Hoctor, T.S., Carr, M., Zwick, P., and Maehr, D.S. Harris, L.D., Hoctor, T.S., and Gergel, S. E. 1996a. 2004. Identification and realization of a Florida Landscape processes and their significance to bio- Ecological Network: the Florida Statewide Green- diversity conservation. In Population Dynamics in ways Project and its political context. In Ecologi- Ecological Space and Time, eds. O. E. Rhodes, Jr., K. cal Networks and Greenways: Concept, Design, Imple- Chesser, and M.H. Smith, pp. 319–347. Chicago: mentation, eds. R.H.G. Jongman and G. Pungetti, University of Chicago Press. pp. 222–250. London: Cambridge University Harris, L.D., Hoctor, T.S., Maehr, D., and Sanderson, Press. J. 1996b. The role of networks and corridors in Hough, F.B. 1882. Report on forestry, submitted enhancing the value and protection of parks and to Congress by the Commissioner of Agriculture. equivalent areas. In National Parks and Protected U.S. Department of Agriculture, Washington, DC. Areas: Their Role in Environmental Areas, ed. R. G. Huffman, J.M., and Blanchard, S.W. 1990. Changes Wright, pp. 173–198. Cambridge, MA: Blackwell in woody vegetation in Florida dry prairie and Science. wetlands during a period of fire exclusion, and Hart, R. 1984. Evaluation of methods for sampling after dry-growing-season fire. In Fire and the En- vegetation and delineating wetlands transition vironment. Ecological and Cultural Perspectives, pp. zones in coastal west-central Florida. Technical 75–83. General Technical Report. SE-69. USDA Report Y-84-2, U.S. Army Engineer Waterways Forest Service, SE Forest Experiment Station, Experiment Station, Vicksburg, MS. NTIS No. AD Asheville, NC. A144 677. Hunter, M.L., Jr., Jacobson, G.L., Jr., and Webb, T., Hermann, S.M. 1993. Small-scale disturbances in III, 1988. Paleoecology and the coarse-filter ap- longleaf pine forests. Proc Tall Timbers Fire Ecol Conf proach to maintaining biological diversity. Conserv 18:265–274. Biol 2: 375–384. Hilmon, J.B. 1969. Autecology of saw palmetto Hutchinson, G.E. 1965. The Ecological Theater and the (Serenoa repens (Batr.) Small). Ph.D. dissertation, Evolutionary Play. New Haven, CT: Yale University Duke University, Durham, NC. Press. 400 IV. Restoration

Kalisz, P.J., and Stone, E.L. 1984. Soil mixing Mohr, C. 1896. Timber pines of the southern United by scarab beetles and pocket gophers in north- States. Bulletin No. 13, USDA Forest Service, central Florida. Soil Sci Soc Am J 48(1):169–172. Washington, DC. Kautz, R.S., and Cox, J.A. 2001. Strategic habitats Moler, P. 1992. Eastern indigo snake. In Rare and for biodiversity conservation in Florida. Conserv Endangered Biota of Florida, Volume III: Amphibians Biol 15:55–77. and Reptiles, ed. P.Molar, pp. 181–186. Gainesville: Lambeck, R.J. 1997. Focal species: A multi-species University Press of Florida. umbrella for nature conservation. Conserv Biol Mutch, R. W. 1970. Wildland fires and ecosystems— 11:849–856. A hypothesis. Ecology 51:1046–1051. Macie, E.A., and Hermansen, L.A., eds. 2002. Myers, R.L. 1990. Scrub and high pine. In Ecosystems Human influences on forest ecosystems: The of Florida, eds. R. Myers and J. Ewel, pp. 150–193. southern wildland–urban interface assessment. Orlando: University of Central Florida Press. General Technical Report SRS-55. USDA Forest Nicolas, J. C., and Steiner, R.L. 2000. Growth man- Service, Southern Research Station, Asheville, agement and smart growth in Florida. Wake Forest NC. Law Rev 35(3):645–670. Maehr, D.S. 1997. The comparative ecology of bob- Noss, R.F. 1987a. From plant communities to land- cat, black bear, and Florida panther in south scapes in conservation inventories: A look at The Florida. Bull Fla Mus Nat Hist 40:1–176. Nature Conservancy (USA). Biol Conserv 41:11– Maehr, D.S., Hoctor, T.S., Quinn, L.J., and Smith, 37. J.S. 2001a. Black bear habitat management Noss, R.F. 1987b. Protecting natural areas in frag- guidelines for Florida. Florida Fish and Wildlife mented landscapes. Nat Areas J 7:2–13. Conservation Commission, Tallahassee, FL. Noss, R.F. 1988. The longleaf pine landscape of the Maehr, D.S., Hoctor, T.S., and Harris, L.D. 2001b. Southeast: Almost gone and almost forgotten. En- The Florida panther: A flagship for regional dangered Species Update 5 (5):1–8. restoration. In Large Mammal Restoration: Ecolog- Noss, R.F. 1989. Longleaf pine and wiregrass: Key- ical and Sociological Challenges in the 21st Century, stone components of an endangered ecosystem. eds. D.S. Maehr, R.F. Noss, and J. L. Larkin, pp. Nat Areaes J 9:211–213. 293–312. Washington, DC: Island Press. Noss, R.F. 1992. The Wildlands Project: Land con- Margules, C.R., and Pressey, R.L., 2000. Systematic servation strategy. Wild Earth (Special Issue):10– conservation planning. Nature 405:243–253. 25. Means, D.B. 1987. Impacts on diversity of the 1985 Noss, R.F. 1993. Wildlife corridors. In Ecology of Land and Resource Management Plan for Na- Greenways, eds. D.S. Smith and P.C. Hellmund, tional Forests in Florida. Unpublished report to pp. 43–68. Minneapolis: University of Minnesota the Wilderness Society. Press. Means, D.B. 1996. Longleaf pine forest: going, go- Noss, R.F., and Cooperrider, A. 1994. Saving Na- ing, ...InEastern Old-growth Forest: Prospects for Re- ture’s Legacy: Protecting and Restoring Biodiversity. discovery and Recovery, ed. M. Davis, pp. 210–229. Washington, DC: Defenders of Wildlife and Island Washington, DC: Island Press. Press. Means, D.B., and Grow, G. 1985. The endangered Noss, R.F., and Harris, L.D. 1986. Nodes, networks, longleaf pine community. ENFO (Florida Conser- and MUMs: Preserving diversity at all scales. En- vation Foundation) Sept: 1–12. viron Manage 10:299–309. Means, D.B., Palis, J.G., and Baggett, M. 1996. Ef- Noss, R.F., and Harris, L.D. 1990. Connectivity and fects of slash pine silviculture on a Florida popula- conservation of biological diversity: Florida as tion of flatwoods salamander. Conserv Biol 10:426– a case history. In Proceedings of the 1989 Society 437. of American Foresters National Convention, Spokane, Meffe, G.K., Carroll, C.R., and contributors. 1997. WA, September 24–27, pp. 131–135. Society of Principles of Conservation Biology, 2nd edition. Sun- American Foresters, Bethesda, MD. derland, MA: Sinauer Associates. Noss, R.F., LaRoe, E.T., and Scott, J.M. 1995. En- Milne, G. 1935. Some suggested units of classifica- dangered Ecosystems of the United States: A Prelimi- tion and mapping for East African soils. Soil Res nary Assessment of Loss and Degradation. Biological 4:183–198. Report 28. Washington, DC: USDI National Bio- Milne, G. 1947. A soil reconnaissance journey logical Service. through parts of Tanganyika Territory, December Noss, R.F., O’Connell, M.A., and Murphy, D. D. 1935 to February 1936. J Ecol 35:192–265. 1997. The Science of Conservation Planning: Habitat 12. Spatial Ecology and Restoration 401

Conservation under the Endangered Species Act. Wash- Smith, D.J. 2003. The ecological effects of roads: ington, DC: Island Press. Theory, analysis, management, and planning Noss, R.F., Carroll, C., Vance-Borland, K., and considerations. Ph.D. dissertation, University of Wuerthner, G. 2002. A multicriteria assessment Florida, Gainesville. of the irreplaceability and vulnerability of sites in Soule,´ M.E. 1991. Land use planning and wildlife the Greater Yellowstone Ecosystem. Conserv Biol maintenance: Guidelines for conserving wildlife 16:895–908. in an urban landscape. J Am Plann Assoc 57:313– Palis, J.G. 1996. Element stewardship abstract: flat- 323. woods salamander (Ambystoma cingulatum Cope). Soule,´ M.E., and Terborgh, J., eds. 1999. Continen- Nat Areas J 16:49–54. tal Conservation: Scientific Foundations of Regional Re- Peet, R.K., and Allard, D.J. 1993. Longleaf pine veg- serve Networks. Washington, DC: Island Press. etation of the southern Atlantic and eastern Gulf Stout, I.J., and Marion, W.R. 1993. Pine flatwoods Coast regions: A preliminary classification. Proc and xeric pine forests of the southern (lower) Tall Timbers Fire Ecol Conf 18:45–81. coastal plain. In Biodiversity of the Southeastern Peters, R.L., and Darling, J.D.S. 1985. The green- United States: Lowland Terrestrial Communities, eds. house effect and nature reserves. BioScience 35: W.H. Martin, S.G. Boyce, and A.C. Echternach, 707–717. pp. 373–446. New York: John Wiley & Sons. Peters, R.L., and Lovejoy, T.E., eds. 1992. Global Tappe, P.A., and Guynn, D.C., Jr. 1998. South- Warming and Biological Diversity. New Haven, CT: eastern fox squirrels: R- or K-selected? Implica- Yale University Press. tions for management. In Ecology and Evolution- Platt, W.J. 1999. Southeastern pine savannas. In Sa- ary Biology of Tree Squirrels, eds. M.A. Steele, J. F. vannas, Barrens, and Rock Outcrop Plant Communities Merritt, and D. A. Zegers, pp. 239–247. Virginia of North America, eds. R. C. Anderson, J.S. Fralish, Museum of Natural History Special Publication 6. and J.M. Basin, pp. 23–51. London: Cambridge Martinsville, VA. University Press. Trombulak, S.C., and Frissell, C.A. 2000. Review Platt, W.J., and Schwartz, M.W. 1990. Temperate of ecological effects of roads on terrestrial and hardwood forests. In Ecosystems of Florida, eds. R. aquatic communities. Conserv Biol 14:18–30. Myers and J. Ewel, pp. 194–229. Orlando: Uni- Turner, M.G. 1989. Landscape ecology: The effect versity of Central Florida Press. of pattern on process. Annu Rev Ecol Syst 20:171– Platt, W.J. , Glitzenstein, J.S., and Rathbun, S. K. 197. 1988. The population dynamics of a long- Turner, M.G., Gardner, R.H., and O’Neill, R.V. 1995. lived conifer (Pinus palustris). Am Nat 131:491– Ecological dynamics at broad scales: Ecosystems 525. and landscapes. BioScience Suppl S29–S35. Poiani, K.A., Richter, B.D., Anderson, M.G., and Tyson, W.K. 1956. History of the utilization of lon- Richter, H.E. 2000. Biodiversity conservation at gleaf pine (Pinus palustris Mill.) in Florida from multiple scales: Functional sites, landscapes, and 1513 until the twentieth century. Master’s thesis, networks. BioScience 50:133–146. University of Florida, Gainesville. Rerick, R. H. 1902. Memoirs of Florida, 2 vols. Atlanta. Varner, J.M., III, Kush, J.S., and Meldahl, R.S. Schwartz, M.W., Hermann, S.M., and Vogel, C.S. 2003a. Vegetation of frequently burned old- 1995. The catastrophic loss of Torreya taxifolia: growth longleaf pine (Pinus palustris Mill.) savan- Assessing environmental induction of disease nas on Choccolocco Mountain, Alabama, USA. hypotheses. Ecol Appl 5:501–516. Nat Areas J 23:43–52. Scott, J.M., Davis, F.W., McGhie, G., and Groves, C. Varner, J.M., III, Kush, J.S., and Meldahl, R.S. 2001. Nature reserves: Do they capture the full 2003b. Structural characteristics of frequently- range of America’s biological diversity? Ecol Appl burned old-growth longleaf pine stands in the 11:999–1007. mountains of Alabama. Castanea 68:211–221. Semlitsch, R. D. 2000. Principles for management Vickers, C.R., Harris, L.D., and Swindel, B.F. 1985. of aquatic-breeding amphibians. J. Wildl Manage Changes in herpetofauna resulting from ditching 64:615–631. of cypress ponds in coastal plains flatwoods. For Smith, D.J. 1999. Highway–wildlife relationships Ecol Manage 11:17–29. (Development of a decision-based wildlife under- Wahlenberg, W.G. 1946. Longleaf Pine: Its Use, Ecol- pass road project prioritization model on GIS with ogy, Regeneration, Protection, Growth and Manage- statewide application). Technical report, Florida ment. Washington, DC: Charles Lathrop Pack Department of Transportation, Tallahassee. Forestry Foundation. 402 IV. Restoration

Walker, J., and Peet, R.K. 1983. Composition and tions for survival in the southeast. Tall Timbers Bull species diversity of pine-wiregrass savannas of the 24. Green Swamp, North Carolina. Vegetatio 55:163– Wells, B.W., and Shunk, I. V. 1931. The vegeta- 179. tion and habitat factors of the coarser sands of Walters, J.R., Hansen, S.K., Carter, J.H., III, Manor, the North Carolina plain: An ecological study. Ecol P.D., and Blue, R.J. 1988. Long-distance dispersal Monogr 1:465–521. of an adult red-cockaded woodpecker. Wilson Bull Whitney, K. 1999. Spatial structure affects land- 100:494–496. scape ecology function. Master’s thesis, University Ware, S., Frost, C., and Doerr, P.D. 1993. South- of Florida, Gainesville. ern mixed hardwood forest: The former longleaf Wilcox, B.A., and Murphy, D. D. 1985. Conserva- pine forest. In Biodiversity of the Southeastern United tion strategy: The effects of fragmentation on ex- States: Lowland Terrestrial Communities, eds. W.H. tinction. Am Nat 125:879–887. Martin, S.G. Boyce, and A.C. Echternacht, pp. Woodmansee, R. G. 1990. Biogeochemical cycles 447–493. New York: John Wiley & Sons. and ecological hierarchies. In Changing Landscapes: Weigl, P.D., Steele, M.A., Sherman, L. J., Ha, J., and An Ecological Perspective, eds. I. S. Zonneveld and Sharpe, T. 1989. The ecology of the fox squir- R. T. T. Forman, pp. 57–71. New York: Springer- rel (Sciurus niger) in North Carolina: Implica- Verlag. Chapter 13 Longleaf Pine Restoration Economics and Policy

Janaki R. R. Alavalapati, G. Andrew Stainback, and Jagannadha R. Matta

Introduction policy and economic opportunities and challenges associated with the restoration of this Public preference for native forest ecosystems ecosystem and make an important observa- is on the rise throughout the world because tion that private landowners are key because a of their valuable market outputs, i.e., tim- large percentage of land in the Southeast lies in ber and nontimber products, and nonmarket their hands. Next, we review some of the ma- outputs such as biodiversity, ecological ser- jor policy initiatives that have been either tried vices, and aesthetics. As a result, restoration of or suggested toward restoring longleaf pine on native forest ecosystems has become an impor- private lands. Finally, we provide some con- tant component of sustainable forest manage- clusions along with future directions for policy ment (Stainback and Alavalapati 2004). Lon- research and development. gleaf pine (Pinus palustris) forests are one of the most biologically diverse native ecosystems in North America, supporting hundreds of plant Longleaf Pine: Past and and animal species. When Europeans ﬁrst col- onized North America, forests dominated by Present longleaf pine covered vast areas of the southeastern Coastal Plain. At that time longleaf The history of longleaf pine forests has been pine forests may have existed on close to 36 described in detail by Frost (this volume). Dur- million hectares (Landers et al. 1995). Due to ing colonial times longleaf pine was harvested landscape changes brought on by colonization, mostly for its valuable wood. However, large- agricultural expansion, and population growth scale logging occurred in the nineteenth and over the past several centuries, longleaf pine early twentieth centuries. By 1935 approxi- today covers only a small fraction of its histor- mately 8 million hectares of the original 36 ical range. million hectares of longleaf pine forests re- This chapter focuses on the economics and mained (Landers et al. 1995). These forests policy aspects of longleaf pine restoration on were further reduced during the 1950s when private lands. In particular, we discuss the timber and pulp industries started to convert

r Janaki R. R. Alavalapati, G. Andrew Stainback, and Jagannadha R. Matta School of Forest Resources and Conservation, University of Florida, Gainesville, Florida 32611.

403 404 IV. Restoration land that previously supported longleaf pine species. When an RCW is on private land, to the fast-growing loblolly pine (Pinus taeda) management activities can be significantly re- and slash pine (Pinus elliottii). Lack of refor- stricted. These restrictions can result in a signif- estation and government policies to exclude icant loss of timber revenue to the landowner. fire also contributed to the decline of longleaf Thus, landowners have a strong incentive to pine forests. Finally, the conversion of forests avoid managing land in a way that may attract to agriculture led to further reductions. Today, RCWs. This includes restoring longleaf pine. virgin longleaf stands exist only in a few iso- Thus, many landowners and resource profes- lated areas (Abrahamson and Hartnett 1990). sionals believe that economic incentives that Over two-thirds of remaining longleaf pine encourage landowners to plant longleaf pine forests in the United States is found on pri- and actively manage land to produce habitat vate lands, with most of this on nonindustrial for the RCW and other species may be more private land. The only exception is the state of effective than relying solely on command and Florida where the majority of longleaf pine is in control regulations. public ownership (Kelly and Bechtold 1983). Most longleaf pine stands are natural in origin as historically very little was planted on Economics of Longleaf cutover sites. Longleaf pine forests support a Pine Restoration diverse collection of plant and animal species (see chapters by Peet and Means in this vol- Much of the old-growth longleaf pine forests ume) many of which have declined with the that exist today are on public land. This in- loss of their habitat. More than 30 species asso- cludes National Forest land, military bases such ciated with this forest are listed as endangered as Eglin Air Force Base, state-owned land and or threatened under the Endangered Species to a lesser extent land owned by the Depart- Act, with the red-cockaded woodpecker (Pi- ment of Interior. As part of its stated objec- coides borealis) being the most notable example. tive to protect and enhance biodiversity on The red-cockaded woodpecker (RCW) re- national forests the U.S. Forest Service has quires large tracts of mature pine stands, made conserving the remaining longleaf on preferably longleaf, with a relatively open units land a top priority. In addition, the U.S. derstory free of midstory vegetation. RCWs Forest Service has planted tens of thousands live in groups that consist of a breeding pair, of hectares of longleaf pine forests. The mil- the current year’s offspring, and sometimes itary has made it a priority to maintain ex- helpers that usually are the male offspring isting stands on its land, and the Department from the previous breeding season (Kennedy of the Interior has set a goal of maintaining et al. 1996). The RCW excavates cavities in ma- and restoring longleaf pine. States and local ture living pine trees for its nest and roosts. It governments have also recognized the value also feeds on the insects that live under the of longleaf and implemented policies for its bark of the trees. A group of cavity trees that restoration. For example, the Blackwater State are used by a group of RCWs is referred to as Forest in Florida maintains extensive longleaf a “cluster.” pine stands and longleaf pine forests on pri- Currently there are fewer than 5000 RCW vate lands have been purchased by Alachua groups scattered throughout the southeastern County, Florida, as part of an effort to protect United States and populations are still being sensitive ecosystems. All of these efforts can lost (Kennedy et al. 1996). The largest pop- make important and unique contributions to ulations occur on public lands where exten- the restoration of longleaf pine and its associ- sive tracts of mature longleaf pine are actively ated environmental benefits. However, in the maintained. In 1973 Congress passed the En- traditional range of longleaf pine only 10% of dangered Species Act (ESA) to prevent the the land base is in public ownership (Johnson loss of vulnerable species, including the RCW. and Gjerstad 1999). The remaining lands are The ESA prohibits harm or taking of a listed privately owned. 13. Longleaf Pine Restoration 405

Because of the large amount of land in pri- layers: the pine canopy, and a rich ground vate ownership, private landowners will play a cover of grasses, legumes, and other flowering crucial role in restoring longleaf forests. There herbs. The understory species of longleaf pine are several advantages that longleaf pine offers forests are fire-adapted and they will resprout landowners. Longleaf is generally more resis- quickly after fire. Reduction in fire frequency tant to fire, hurricane damage, and pine bark to intervals longer than 5 years leads to beetle attacks than other common commercial elimination of the herb layer, which provides pine species in the region. It grows relatively seeds and legumes for small game species straight and firm, making it a good choice such as rabbits and quail, as well as hun- for producing sawtimber and poles. Longleaf dreds of nongame birds and small mammals pine needles make an excellent straw, which like mice and voles which, in turn, sup- can provide periodic income for landowners. port hawks, owls, and mammalian predators It can produce excellent habitat for game such (http://www.forestry.auburn.edu/sfnmc/class/ as bobwhite quail and is generally perceived longleaf.html). Much of this complex food as more aesthetically pleasing than other pine web collapses when fire is eliminated from species. Despite these advantages, except for woodlands. Direct costs and liability issues very xeric sites, longleaf has generally proved associated with managing fire-dependent to be less profitable than other timber species longleaf pine forest ecosystem are additional in the region. hurdles that discourage landowners from Investigating various site indices, thinning restoring longleaf pine. schedules, rotation ages, and discount rates, There have been some recent improvements Yanquoi and Flick (1994) found longleaf in establishing stands of longleaf pine. Arti- pine to be less profitable than loblolly pine. ficial regeneration is being advanced by con- Alavalapati et al. (2002) found longleaf tainer seedlings and herbicide control of com- pine to be less profitable than slash pine for peting vegetation. For instance, harvesting and producing sawtimber and pulpwood from regeneration using the shelterwood method, unthinned even-aged plantations. Stainback where mature pines are left during initial har- and Alavalapati (2004) found that longleaf vest to provide a seed source, are well suited pine in the form of silvopasture (growing for longleaf pine (Demers and Long 2003). longleaf pine along with pasture) can be There is also an increasing interest in pro- more profitable than growing pure longleaf viding incentives to private landowners to pine. There are several reasons for the lower plant longleaf pine. Many environmental ben- returns of longleaf pine. First, longleaf pine efits such as biodiversity, carbon sequestra- has historically been difficult to regenerate. tion, recreational and aesthetic benefits that In addition, it is not a prolific seed producer are commonly associated with longleaf pine and its seeds are relatively large, resulting are at least partly external to the landowner. in lower dispersal rates. Second, longleaf Thus, although these attributes of longleaf pine is generally intolerant of competition may provide significant benefits to society, the (Boyer and Peterson 1983). Thus, natural landowner rarely receives financial compensa- regeneration is typically more difficult than tion for producing them. If public preferences with other common pine species of the region. for hunting in longleaf pine ecosystems, lon- Third, because of the early grass stage, longleaf gleaf pine straw, and longleaf sawtimber reflect pine typically exhibits slower growth at young in the market, in the form of price premiums stand ages (Boyer 1996). Finally, managing for example, competitiveness of longleaf rela- understory in longleaf pine forests is a major tive to slash or loblolly might improve. hurdle that landowners face. Maintaining Currently, longleaf pine restoration and sunny and open areas is required in order to its associated benefits may be underpro- ensure native plant and animal diversity in duced from a social perspective. Several pol- longleaf pine forests. A healthy longleaf pine icy mechanisms have been suggested and some community will have two principal vegetation have been implemented to internalize the 406 IV. Restoration environmental benefits associated with lon- Policy Options to Restore gleaf pine management. The most notable to date have been centered around inducing Longleaf Pine on Private landowners to create or improve habitat for the Land federally endangered RCW. These policies include Habitat Conservation Plans (HCPs), Safe Endangered Species Act and Harbor programs, and marketable Transfer- able Endangered Species Certificates (TESCs). Incentives to Protect RCW Other policies include cost sharing programs, The ESA, which was enacted by Congress in efforts to disseminate information to landown- 1973, is to protect and recover species that are ers about the potential benefits of longleaf close to extinction. The Act prohibits any indi- pine and conservation easements purchased by vidual to “take” a species, where “take” is de- governments or nonprofit organizations. For fined as “to harass, harm, pursue, hunt, shoot, example, the Florida Fish and Wildlife Conser- wound, kill, trap, capture, collect, or attempt vation Commission encourages management to engage in any such conduct.” Under this and enhancement of RCW foraging habitat Act, the U.S. Fish and Wildlife Service oversees and nesting sites through its Landowner Incen- species that live on land, while the National tive Program (LIP). Similarly, USDA’s Natural Marine Fisheries Service focuses on marine Resources Conservation Service provides both species. The RCW was one of the first species technical assistance and up to 75% cost-share listed as endangered under the ESA. In 1995 assistance to establish and improve wildlife in a landmark case, Sweet Home v. Babbitt, the habitat under its Wildlife Habitat Incentives U.S. Supreme Court ruled that it is permissible Program (WHIP). Longleaf pine areas have for the U.S. Fish and Wildlife Service to define been accorded a priority ecological site status “take” to include any modification of habitat for this program in Florida. Potential policies of a listed species that would impair behavior for restoring longleaf pine also include setting essential for the survival of the species. Such up markets that allow landowners to sell cred- behavior potentially includes mating, feeding, its for the additional carbon that would be se- sheltering, or any other behavior essential to questered as a result of switching to longleaf survival. pine from other species that have shorter ro- The ESA can require public agencies, such tations. These credits could potentially be sold as the U.S. Forest Service, not only to prevent to governments or private companies if a cap taking a listed species but also to implement and trade system for CO2 emissions were im- plans to recover the species. However, private plemented similar to the one set up for sul- landowners have a responsibility only to pre- fur dioxide emissions and acid rain. Changes vent a taking of a listed species. This is an im- can be made to the Conservation Reserve Pro- portant distinction, as over 80% of the species gram that would provide greater incentives listed under the Act have at least some habitat for landowners to plant and maintain biolog- on private land (Zhang and Mehmood 2002). ically important ecosystems such as longleaf As mentioned before, more than 90% of the pine forests. Finally, research and information potential longleaf pine (preferred RCW habi- about alternative land uses that may make lon- tat) occurs on private land. Thus, implement- gleaf pine more attractive to landowners can be ing polices that effectively restore longleaf pine disseminated. For example, using longleaf pine on private land is essential for the recovery of in agroforestry systems such as silvopasture or the RCWs. selling pine straw can generate an annual cash The benefits of biological diversity and en- flow to help defray the disadvantage of the dangered species conservation may be large long time periods involved in producing saw- but are generally diffuse throughout society. timber and poles (Stainback and Alavalapati Conversely, the cost of protecting biodiversity 2004). In the remainder of this paper we will on private land is concentrated on a relatively discuss these policy options in greater detail. small number of landowners. In addition, the 13. Longleaf Pine Restoration 407 benefits of land management aimed at max- (Based on testimony of Benjamin Cone, Jr. before imizing private goods are received mostly by the U.S. Congress, March 31, 1997) the individuals who own the land, whereas the costs associated with the degradation of public In 1982 the ESA was amended in order to goods such as RCW habitat are diffuse. Thus, lessen the economic burden of the Act on pri- the interests of the private landowner often vate landowners. The amendments allow a are not the same as those of the public when landowner to incidentally take a listed species it comes to biodiversity. The ESA restrictions if the take is appropriately mitigated. “Inci- on private lands are largely aimed at mitigat- dental” is defined as an action that is not in- ing the latter type of market failure. However, tended to take a species but nonetheless results addressing the private cost of protecting biodi- in a take. Residential development or timber versity is also very critical. harvests are examples of actions that can be The restrictions placed on private land use considered incidental. Appropriate mitigation through the ESA to protect RCWs can be involves the development of a Habitat Conser- costly to the landowner. For example, around vation Plan (HCP) that may include relocating each cavity tree at least 60 acres of forag- populations or creating new habitat elsewhere. ing habitat must be maintained within a half- These provisions allow much greater flexibil- mile of the tree. In addition, there are restric- ity for landowners to comply with the act. tions placed on the harvesting of large pines, HCPs involving many landowners or statewide pesticide use, and road construction within HCPs allow for further reduction in costs. a 200-foot radius of a cavity tree. These re- The flexibility of HCPs can be enhanced to strictions can significantly reduce the value of provide financial incentives for landowners the land to the landowner. For example, Lan- to enhance and expand habitat for listed cia et al. (1989) found that the opportunity species. cost of providing foraging habitat for RCWs One example of this involving longleaf pine can be as high as $155 per acre. This cost is is the Safe Harbor program developed by the an incentive for landowners to manage their U.S. Fish and Wildlife service in the late 1990s. land in a way that prevents RCWs from in- In this program landowners sign a contract habiting their property. A well-publicized ex- obligating them to protect a baseline popu- ample often touted by those seeking reform lation of RCWs. In addition, landowners in is Ben Cone, a private landowner in North this program must agree to provide additional Carolina. habitat by planting more trees, removing understory, and providing artificial nesting cavities (Zhang and Mehmood 2002). In exchange Ben Cone owns large tracks of longleaf pine in North Carolina. He managed his land using longer har- the landowner will not be subject to further vesting rotations and frequent low-intensity fires. land use regulations if more RCWs inhabit This management regime gave him a stream of in- their land. The landowner thus eliminates the come from quail hunting leases and timber harvests risk that the property will be further regu- as well as creating habitat for RCWs. In 1991 the lated under the ESA. The Safe Harbor program U.S. Fish and Wildlife Service prevented Cone from started in North and South Carolina but it is ex- harvesting timber on more than 1,500 acres of his panding in other states. Landowners are gen- land in order to protect 29 RCWs that resided on his erally supportive of the Safe Harbor program. property. Cone now clearcuts around the restricted Zhang and Mehmood (2002) found through RCW habitat not allowing any trees to mature to an surveys that participants in Safe Harbor are age that would attract more woodpeckers. He claims pleased overall and favor the program over a that he is motivated to prevent further RCWs from 1 inhabiting his land to prevent further regulation of strict command and control regime. However, his property. Thus, while the ESA is efficient in pro- the authors conclude that the expansion of the tecting the existing RCWs on private property, it is program could be greatly helped by tying the also creating a perverse incentive discouraging the program with more direct financial incentives production of more habitat. to the landowners. 408 IV. Restoration

Along these lines, Kennedy et al. (1996) Carbon Credits, Pasture (Cattle), proposed developing marketable Transferable and Longleaf Pine Endangered Species Certificates (TESCs) for RCWs. In this program a landowner would re- Global climate change, induced by combus- ceive TESCs for engaging in activities outlined tion of fossil fuels and land use change, is a in a statewide HCP. The landowner could then growing concern among many policymakers use these certificates to incidentally take RCWs and environmentalists. Forests are thought to or sell them to other landowners. The mar- play a crucial role in the global carbon cycle ket created in RCW habitat would have sev- by sequestering carbon dioxide as they grow eral distinct advantages. First, allowing TESCs and emitting it when biomass decays or burns. to be traded would lower the overall cost of As such, planting trees on marginal agricul- mitigation activities. Landowners who have a tural land, growing trees on ranchlands (sil- low cost for the mitigation required by the vopasture), and changing forest management HCP would have an incentive to acquire these regimes to increase biomass production may be certificates and sell them to landowners with desirable options to sequester additional car- higher costs. Landowners who have no RCWs bon. This potential influenced participants in on their property can also participate. The HCP the Kyoto protocol to allow countries to count baseline for such a landowner would be zero. carbon sequestered in forests toward obliga- Any additional RCW habitat produced would tions under the protocol. The United States be extra and the landowners could sell any has been a strong proponent of this idea. Thus, TESCs they received. Such a program could even though the United States has pulled out provide a strong incentive for landowners to of Kyoto, President Bush’s alternative proposal produce RCW habitat because it would now includes over $3 billion for agricultural and have a marketable value. If the program re- forestry carbon sequestration activities (Bush quired more than 1 hectare of RCW habi- 2002). Research indicates that forestry can be a tat to be created for every permit to destroy cost-effective option for reducing atmospheric 0.4 hectare, it would result in a net increase in carbon dioxide concentrations. Dixon (1997), RCW habitat. for example, estimated carbon could be se- Similar programs have been implemented questered in U.S. forests for between $2 and on small scales to protect other endangered $56 per ton. species and wetlands on private land. For ex- Research focusing on sequestering carbon ample, in 1995 the Bank of America created a in southeastern U.S. pine forests suggests that 73-hectare conservation bank in southern Cal- private landowners could be induced to se- ifornia to preserve threatened and endangered quester additional carbon for relatively modest species. The bank then received credits and prices. Stainback and Alavalapati (2002) inves- sold those credits to developers who needed tigated the impact of internalizing carbon ben- them to develop other lands. The Florida efits onto private forest landowners on the op- Wetlands’ Bank has restored 140 hectares of timal management of slash pine plantations. wetlands and sold credits for $40,000 each The results indicate that carbon prices of less under provisions of the Clean Water Act than $50 per metric ton can induce landown- (Shogren 1998). In order for these programs ers to lengthen their rotations to sequester ad- to work effectively they need careful govern- ditional carbon. Huang and Kronrad (2001) ment oversight to ensure no net loss or even found that private landowners would be will- require a gain in habitat or wetlands. How- ing to shift from the financially optimal tim- ever, by providing an economic value for eco- ber rotation to a rotation that maximizes se- logical services they can provide incentives questered carbon for prices of carbon less than for landowners to more explicitly consider $70 per metric ton in loblolly plantations. environmental benefits in land management Stainback and Alavalapati (2004) conducted decisions. an in-depth economic analysis of restoring 13. Longleaf Pine Restoration 409

TABLE 1. Land expectation values ($/hectare) under different management regimes and carbon prices ($/metric ton) for longleaf pine in the southeastern United States.a,b Management regime

Carbon price p tf sp tf (rot = 60) tf (rot = 80) sp (rot = 60) sp (rot = 80) 0 1700.18 2088.04 2804.64 1507.34 1186.10 2471.05 1927.42 10 1700.18 3088.81 3805.42 2644.02 2372.21 3558.31 3014.68 20 1700.18 4151.36 4843.26 3756.00 3558.31 4695.00 4200.79 30 1700.18 5201.56 5893.45 4867.97 4744.42 5806.97 5386.89 40 1700.18 6276.47 6956.01 5979.94 5905.81 6918.94 6572.99 50 1700.18 7376.08 8055.62 7116.62 7091.91 8043.27 7882.65 a Reprinted from Stainback and Alavalapati (2004) with permission from Elsevier. b p, tf, and sp represent traditional pasture, traditional forestry, and silvopasture, respectively, “rot” stands for a fixed rotation at either 60 or 80 years. longleaf pine. They developed a stand-level initial tree density for silvopasture would be economic model of a silvopasture (pasture 1236 and 988 trees per hectare, respectively. and longleaf pine) and compared the prof- Increasing the rotation ages to create forag- itability of silvopasture with traditional pas- ing and nesting RCW habitat was found to ture (no trees) and traditional forestry (only decrease the land value under both tradi- trees). Within the silvopasture and traditional tional forestry and silvopasture. The opportu- forestry regimes they estimated the initial tree nity cost of extending the rotation for tradi- density and rotation age that maximize land tional forestry and silvopasture is shown in value. They then incorporated carbon bene- Fig. 1. The opportunity cost of extending ro- fits into the model and investigated how pay- tation to 60 years is smaller than that of tra- ments to the landowner for sequestering car- ditional longleaf pine forests. Results suggest bon in trees impact the profitability of silvopas- that the opportunity cost of extending the ro- ture and traditional forestry with longleaf pine. tation, however, declines with an increase in Specifically, they modeled that a landowner is carbon price. In the face of carbon payments paid for sequestering carbon as the stand grows and policy support to ensure longer rotations, and the landowner pays for carbon emissions longleaf pine restoration through silvopasture from the decay of sawtimber, pulpwood, and can be a viable option. With more than 2 mil- wood waste left at harvest. They incorporated lion hectares of ranchlands, Florida provides revenues associated with pine straw. They sim- unique opportunities to restore longleaf pine ulated the model to estimate the opportunity through silvopasture. cost of fixing the rotation age at 60 and 80 There may be other payment mechanisms years to provide habitat for the RCW for both that are easier to administer and have lower silvopasture and traditional forestry with and transaction cost. For instance, a lump sum without carbon payments. could be paid to landowners to plant longleaf Stainback and Alavalapati’s (2004) results pine on marginal pasture or other land with no suggest that silvopasture is more profitable trees with the payer receiving the carbon cred- than both traditional forestry and traditional its. Alternatively, payments could be made on pasture with and without carbon payments the basis of temporarily sequestering carbon. (Table 1). As expected, land values for silvopas- In this type of program, a landowner could ture and traditional forestry increase with in- be paid a smaller amount to sequester carbon creasing carbon prices. If the price of carbon for a year. As long as the carbon remains se- is $0 and the rotation age were to be fixed to questered, the landowner continues to receive 60 and 80 years, respectively, to create forag- this rental payment. Even though these mech- ing and nesting habitat for RCW, the optimal anisms exclude the carbon that is potentially 410 IV. Restoration

1000

800

600

400

200

opportunity cost ($/hectare) 0 0 10 20 30 40 50 carbon price ($/metric ton) traditional forestry (rot=60) traditional forestry (rot=80) silvopasture (rot=60) silvopasture (rot=80)

FIGURE 1. Opportunity cost of fixing the rotation age at 60 and 80 years under traditional forestry and silvopasture with longleaf pine in the southeastern United States. Reprinted from Stainback and Alavalapati (2004) with permission from Elsevier. sequestered in end products, they reduce the place an even greater emphasis on environ- need to have reliable estimates of product de- mentally sensitive lands. Even though erosion cay and resulting carbon emissions, thus mak- is still the top priority, other environmental ing it easier to deal with risk and changing mar- benefits such as generation of wildlife habi- ket conditions. tat receive greater attention in calculating the new EBI. Landowners who receive contracts Other Potential Policy are given annual payments of 10 to 15 years and cost sharing assistance to establish a per- Mechanisms manent ground cover such as trees. The Conservation Reserve Program (CRP) has Because of its high environmental value and been responsible for a significant portion of ability to grow on marginal pastureland, lon- the longleaf pine that has been planted on pri- gleaf pine is a good candidate for CRP partici- vate lands to date. This can be furthered by ex- pation. In 1998 a longleaf pine national CRP panding this program or by combining it with priority area was established to help restore other programs. The CRP emerged from the longleaf pine on private lands in the South- Soil Bank Act of the 1950s. It was originally east. This approach has proved effective. For conceived as a means to take highly erodible instance, since 1998 over 10,000 hectares of agricultural land out of production. Usually longleaf pine have been planted in South Car- land managed under the program is planted olina as a result of this program (USDA 2003). in trees, although sometimes other vegetative Expanding the acreage allowed under the CRP cover can be used. There has been a trend in and possibly combining it with other incen- recent years to expand the mission of the CRP tive programs such as those discussed earlier to address other environmental concerns as to encourage RCW habitat improvement may well. For instance, basing on an Environmental be very effective in restoring longleaf on pri- Benefits Index (EBI), landowners make offers vate land. Programs such as LIP and WHIP are to enroll their land through a competitive pro- particularly worth exploring. cess that chooses land with the highest EBI. In addition to expanding the CRP, efforts can The area enrolled each year has traditionally be made to disseminate information regarding been capped around 14.5 million hectares. In nontimber products that can be marketed from 2002, the CRP was reauthorized to expand the longleaf pine. These include pine straw, hunt- program to over 15.6 million hectares and to ing leases, and silvopasture. Roise et al. (1991) 13. Longleaf Pine Restoration 411 found that pine straw from a longleaf pine for the RCW and a carbon market as well as plantation can yield $32 to $64 per hectare participating in the Conservation Reserve Pro- per year. This translates into an increase in gram. This type of co-benefit approach is inland value of more than $80 per hectare.2 Lon- creasingly recognized as a necessity for making gleaf forests provide good habitat for quail and market mechanisms work for the conservation turkey and are compatible with grazing in a sil- of ecologically valuable lands (Daily and Elli- vopasture system. These nontimber products son 2002). Furthermore, an increase in straw not only provide an additional source of in- price, demand for quail hunting, and premi- come but also provide intermediate income ums for graded lumber from longer rotation while waiting for a stand of trees to mature. might stimulate landowners to restore lon- This can be especially important for longleaf gleaf pine. However, it is doubtful that market pine because of the longer rotations associated mechanisms are going to prove to be a panacea with sawtimber and pole production as well as for longleaf pine restoration. Instead, a diver- RCW habitat production. sity of approaches is needed. This includes restoration of longleaf on public lands, public purchase of private lands with longleaf pine Conclusions ecosystems, especially those under a threat of conversion to other uses, the use of conserva- Longleaf pine is regarded by many as one of the tion easements and appropriate regulation of most treasured forest resources in the south- private land, as well as market incentives for ern United States. Yet, because of past poli- landowners. With a growing interest in lon- cies and the fact that many of the benefits of gleaf pine, and innovative policies, this species longleaf pine are diffuse, this forest has dra- may once again be a substantial component of matically declined since European coloniza- our southern forests. tion. Recent efforts of landowners, resource specialists, and governments have made some progress in restoring this once dominant for- Endnotes est on private lands. However, more efforts are 1. Overreliance on command and control policies to needed to ensure the perpetuation of longleaf regulate private lands may be politically unsus- pine and all of its benefits into the future. tainable. At the time of writing, a bill (HR 4840) The increased interest in harnessing market is being considered by Congress that would al- forces to help solve environmental problems low landowners to participate more effectively may be particularly useful in the restoration and provide input in developing recovery plans. of longleaf pine. The vast majority of poten- 2. Reported figures on pine straw revenues may tial longleaf pine sites are located on private not account for additional fertilization that is property. Thus, traditional command and con- needed to offset nutrient removal associated with straw collection. This means additional cost to the trol mechanisms may not be conducive for its landowner and lower profits. restoration. Further, focusing on just one of the many environmental services associated with longleaf pine may not provide a strong References enough incentive for its restoration. For example, Alavalapati et al. (2002) found that car- Abrahamson, W.G., and Hartnett, D.C. 1990. Pine bon benefits alone were not enough to make flatwoods and dry prairies. In Ecosystems of Florida, longleaf pine as profitable as slash pine. A eds. R.L. Myers and J.J. Ewel, pp. 103–149. Orlando: University of Central Florida Press. combination of subsidizing landowners for car- Alavalapati, J.R.R., Stainback, G.A., and Carter, D.R. bon sequestration, RCW habitat, and other 2002. Restoration of the longleaf pine ecosystem amenities was necessary. Landowners grow- on private lands in the U.S. South: An ecological ing longleaf pine in the future may be able to economic analysis. Ecol Econ 40:411–419. participate in several markets such as a Trans- Boyer, W.D. 1996. Longleaf pine can catch ferable Endangered Species Certificate market up. In Proceedings of the First Longleaf Alliance 412 IV. Restoration

Conference—Longleaf Pine: A Regional Perspective on Lancia, R.A., Roise, J.P., Adams, D.A., and Lennartz, Challenges and Opportunities, ed. J. S. Kush, pp. 28– M.R. 1989. Opportunity costs of red-cockaded 29. Mobile, AL, September 17–19. woodpecker foraging habitat. South J Appl For 13: Boyer, W.D., and Peterson, D.W. 1983. Longleaf 81–85. pine. In Silvicultural Systems for the Major Forest Landers, J.L., Van, L., David, H., and Boyer, W.D. Types of the United States. Agricultural Handbook 1995. The longleaf pine forest of the Southeast: No. 445, ed. R.M. Burns, pp. 153–156. Washing- Requiem or renaissance? J For 93:39–43. ton, DC: USDA Forest Service. Roise, J.P., Chung, J., and Lancia, R. 1991. Red- Bush, G.W. 2002. Global Climate Change Policy cockaded woodpecker habitat management and Book. The White House, Washington, DC. longleaf pine straw production: An economic Daily, G.C., and Ellison, K. 2002. The New Economy analysis. South J Appl For 15:88–92. of Nature: The Quest to Make Conservation Proﬁtable. Shogren, J.F. 1998. A political economy in an eco- Washington DC: Island Press. logical web. Environ Resour Econ 11:557–570. Demers, C., and Long, A. 2003. Longleaf pine Stainback, G.A., and Alavalapati, J.R.R. 2002. Eco- regeneration. Retrieved 11 April 2003 from nomic analysis of slash pine forest carbon seques- http://edis.ifas.uﬂ.edu/BODY FR064. tration in the southern U.S. J For Econ 8:105– Dixon, R.K. 1997. Silvicultural options to conserve 117. and sequester carbon in forest systems: Prelim- Stainback, G.A., and Alavalapati, J.R.R. 2004. inary economic assessment. Crit Rev Environ Sci Restoring longleaf pine through silvopasture Technol 27:139–149. practices: An economic analysis. For Policy Econ Huang, C.H., and Kronrad, G.D. 2001. The cost of 6(3–4): 379–390. sequestering carbon on private forest lands. For USDA. 2003. Farmers save environmentally- Policy Econ 2:133–142. sensitive farmland with CRP. Retrieved Johnson, R., and Gjerstad, D. 1999. Restoring the 15 April 2003 from http://www.scda.state.sc.us/ longleaf pine forest ecosystem. Alabama’s Trea- AgLinks/ANRCESITE/fsa enviro sens CRP.html. sured Forests Fall:18–19. Yanquoi, R., and Flick, W.A. 1994. Economics of Kelly, J.F., and Bechtold, W.A. 1983. The longleaf longleaf pine management. In Forest Economics on pine resource. In Proceedings of the Symposium on the Edge: Proceedings of the 24th Annual Southern the Management of Longleaf Pine, ed. R.M. Farrar, Forest Economics Workshop, eds. D.H. Newman and Jr., pp. 11–22. Long Beach, MS, April 4–6. M.E. Aronow, p. 89. Savannah, GA, March 27– Kennedy, E.T., Costa, R., and Smathers, W.M., Jr. 29. 1996. New direction for red-cockaded wood- Zhang, D., and Mehmood, S.R. 2002. Safe harbor pecker habitat conservation: Economic incen- for the red-cockaded woodpecker: Private forest tives. J For 94:22–28. landowners share their views. J For 100:24–29. Chapter 14 Role of Public–Private Partnership in Restoration A Case Study

Vernon Compton, J. Bachant Brown, M. Hicks, and P. Penniman

Introduction landowners and managers shared concerns and challenges regarding the decline of the With today’s increasing challenges in restoring longleaf pine ecosystem in northwest Florida the longleaf pine ecosystem, land managers, and south Alabama. In 1996, seven public both public and private, need innovative man- and private landowners formed a partner- agement solutions. Since most challenges are ship to address common land- and water- shared across the landscape and desired end conservation concerns and challenges, and to results are similar for land managers, one in- utilize the opportunity to act collaboratively novative approach that is proving effective is and cooperatively. Currently, there are ten working in partnership with multiple orga- partners in GCPEP that share landscape-scale nizations, agencies, and stakeholders. Within conservation goals in the region. a partnership, members share the risks and This chapter will describe how the frame- the challenges of managing the longleaf pine work and function of GCPEP may provide a ecosystem, as well as the beneﬁts, such as “blueprint” for other partnerships, and will ex- healthier, more functional ecosystems. Focus plain the ecological rationale behind the cre- and emphasis on collaboration, cooperation, ation of GCPEP. In addition, some of the early and consensual goals provide the foundation and current successes as well as challenges the for positive and productive partnership ac- partnership has experienced will be discussed. tions, which usually result in successful attain- The chapter will also examine how the partment of partnership and member goals and nership maintains a common focus on, and objectives. kinetic progress toward, conservation goals The Gulf Coastal Plain Ecosystem Partner- through planning and prioritization methods. ship (GCPEP) is an example of a partner- The chapter is approached in sequence ship that has been able to frequently attain beginning with the concepts that lead to challenging and ambitious landscape-scale the formation of the partnership, including conservation goals and objectives through landscape-scale conservation and the ad- positive, result-oriented action and collabo- vantages of ecosystem management through ration. GCPEP was formed because several partnerships. The following sections explain in

r Vernon Compton, J. Bachant Brown, M. Hicks, and P. Penniman The Nature Conservancy, Jay Florida Ofﬁce, The Gulf Coastal Plain Ecosystem Partnership of Jay, Florida 32565.

413 414 IV. Restoration detail the various aspects of GCPEP, including will only happen through support of and par- inception, discussion about each individual ticipation in conservation planning and imple- partner, and many of the conservation projects mentation by the local community. that have been identified as priorities by the partners. Partnerships in Landscape-Scale Conservation Conservation When forming a landscape-scale partnership, consideration of many different factors is es- Landscape-scale conservation served as the sential. Partnerships require a clear under- primary purpose for establishing GCPEP. Suc- standing of the purpose of the individual orga- cessful landscape-scale conservation usually nizations interested in becoming enrolled, as involves actions that affect large and numer- well as the manner in which the coalition of ous parcels of land, typically owned by multi- organizations will operate. A successful part- ple persons or organizations. Conserving func- nership will become an entity of its own that tional landscapes improves the likelihood of ideally will be greater than the sum of its parts. achieving sustainable conservation of biodi- This best occurs when each organization is well versity. According to Low (1999), emphasis on established and committed to remaining in- conserving functional landscapes dramatically volved in the partnership for the long term. improves efficiency and effectiveness for the Partnerships are often guided by a Steering following reasons: Committee, the method used by GCPEP. Ide- r ally, the Steering Committee has agreed-upon Conservation actions that simultaneously af- operating guidelines to ensure efficient opera- fect ecological systems, communities, and tion of the partnership. During Steering Com- species at multiple scales within a single in- mittee meetings, and day-to-day operations tact landscape provide a more ecologically and interactions, it is important to approach integrated conservation strategy that better all topics and issues with the utmost respect for protects functional landscapes and biodiver- members and their respective organizations, as sity. r well as to minimize preconceived expectations Functional landscapes typically include pri- and conceptions. Negotiations are most suc- vate and public lands both of which are fre- cessful when all partners view one another as quently needed to protect and restore eco- equal. When the playing field is level for ev- logical processes. r eryone involved, it provides an effective envi- Landscape-scale conservation requires an ronment for cooperation, communication, and ecosystem approach involving multiple understanding. Greatest potential for success strategies to abate critical threats driven by is realized when goals for far-reaching coop- incompatible human uses of the lands and erative restoration projects are shared and in- waters. r volvement for partners is maximized. The end Landscape-scale conservation focuses on results of such an approach can be extremely restoration of conservation targets. positive and may produce widespread benefits Ecologically important natural systems and re- that may never have been imagined when ini- sources are typically embedded within a large tially planning meetings and projects. working landscape, which includes the people who live and work in these places. Except for isolated wilderness areas, threats to con- GCPEP: An Example of an servation targets often involve incompatible Effective Partnership human uses and economic development. So- lutions invariably require working with local The Gulf Coastal Plain Ecosystem Partnership landowners, community leaders, and govern- is a successful collaboration among ten pub- ments. Long-term conservation of these places lic and private organizations that collectively 14. Public–Private Partnership in Restoration 415

FIGURE 1. Gulf Coastal Plain Ecosystem Partnership lands and surrounding landscape in northwest Florida and south Alabama. manage more than 425,859 ha of land in one cooperative restoration methods, which have of the most biologically signiﬁcant regions in been successful for GCPEP.The partnership has North America (Fig. 1). The GCPEP landscape proven to be more effective and productive has the vast majority of the world’s remain- than expected, achieving goals that no one or- ing old-growth longleaf pine ecosystems, con- ganization could individually accomplish. taining some longleaf pine trees that are over 500 years old. GCPEP partners include the Departments of Defense, Florida Department How GCPEP Began of Environmental Protection, Florida Division of Forestry, Florida Fish and Wildlife Con- The GCPEP began with an idea. One agency servation Commission, International Paper, contacted another to discuss the possibility National Forests in Alabama, National Park of combining efforts to create a contiguous Service, Nokuse Plantation, Northwest Florida landscape for recovery efforts for the federally Water Management District, and The Nature endangered red-cockaded woodpecker (RCW; Conservancy (Compton et al. 2002a and The Picoides borealis). By reconnecting the longleaf Nature Conservancy 2005). pine ecosystem, northwest Florida and south Explaining the inception of GCPEP may pro- Alabama lands could provide enough contigu- vide guidelines for initiating a partnership. An ous forest to aid in the recovery of the RCW understanding of the partnership framework and other rare species, such as Florida black and projects may offer measures of success for bears (Ursus americanus ﬂoridanus). The original 416 IV. Restoration

GCPEP landscape consisted of connected lands jority is charged with finding an alternative that were primarily undeveloped, but became solution acceptable to all. The goal is to always fragmented by roads and increasing develop- maintain productivity while keeping the con- ment. Reconnecting these lands through shar- sensus process efficient. Decisions are based ing resources, cooperating on management ac- upon Steering Committee voices only—the tivities, and protecting important conservation GCPEP staff does not vote. The Steering Com- lands would potentially restore the landscape mittee functions best when everyone partici- to establish a more functional metapopulation pates and ensures input from their respective of RCWs and other species requiring broad organizations in all decisions. and largely intact longleaf pine and associated The Steering Committee established ecosystems. GCPEP’s mission: to develop a set of long-term Since GCPEP originally formed there have strategies to abate the critical threats and to been several changes. New partners have joi- improve regional ecosystem health; to recover ned GCPEP, while existing partners have en- listed species of plants and animals and avoid rolled additional lands into the partnership new listings; to restore and protect large, landscape. Steering Committee representa- connected, functional examples of native tives have changed due to shifting responsibil- ecosystems; and to provide ecosystem goods ities, relocations, and retirements. The GCPEP and services compatible with the above to staff, which is explained later in more detail, surrounding communities. plays an important role of providing continu- At each GCPEP Steering Committee meet- ity over time as changes occur both within ing, a research, scientific, and general infor- and surrounding the partnership. Additional mation manual that highlights all the partners’ GCPEP staff has been added to support strate- progress since the last meeting is distributed gies and actions set by the Steering Committee. by the GCPEP staff to each representative (Compton et al. 2002b). The manual is then disseminated to the widest audience possible, The GCPEP Framework particularly within and to the partners’ agencies and supervisors. The Steering Commit- It was decided that a Steering Committee tee recognizes the importance of exporting the would allow the GCPEP to function best be- lessons learned from the partnership to other cause each partner would have equal rep- landowners, organizations, community lead- resentation and decision-making power. The ers, and the general public. Scientific research GCPEP is guided by the Steering Committee, and knowledge gained remains limited in which is composed of two representatives from value if not shared with either those who man- each of the partner organizations. Each part- age land or influence the management of land. ner organization chooses the representatives, which include one primary and one alternate contact. Representation at the Steering Com- Early Successes of GCPEP mittee meetings by one of the representatives from each partner organization is encouraged. GCPEP Memorandum of Occasionally when a representative is unable Understanding to attend the meeting a designee chosen by the primary contact may represent the organi- The wording of the GCPEP Memorandum zation. The GCPEP Steering Committee, which of Understanding (MOU) was established meets biannually, has established guidelines to through a series of meetings to discuss the el- ensure efficient operation of the partnership. ements that each agency could agree upon, Consensus is desired in reaching agreements which would also fit within legal and inner- among the partners during the Steering Com- agency requirements. The MOU recognizes mittee meetings to ensure an equal voice for that the individual public and private agen- all. If there is minority dissent, then the ma- cies have legitimate and varied management 14. Public–Private Partnership in Restoration 417 goals. The MOU is in no way intended to limit the media. Continuous and careful planning or constrain the individual goals and missions assists with the challenge of allotting the of each partner’s organization. amount of time, staff, and resources necessary The purpose of the GCPEP MOU is to de- to manage required tasks, while maintaining velop and implement a voluntary and cooper- and balancing the prioritization of crucial con- ative stewardship strategy to sustain the long- servation opportunities that may be lost if not term viability of native plants and animals, the promptly addressed. integrity of ecosystems, the production of com- Challenges are experienced during the pro- modities and ecosystem services, and the hu- cess of receiving approval and submitting man communities that depend upon them. funding proposals with numerous partner or- The goals of the GCPEP MOU are to assist, ganizations. Clear and constant communica- share information, and coordinate efforts with tion with each department involved is required the member partners in fulfilling the purposes to complete proposal submissions. Ensuring of the MOU; to provide a model for local, state, each partner involved in the agreement has re- federal, and private entities working together viewed and approved the proposal typically re- to fulfill the purposes of the MOU; and to com- quires additional time. Close attention must be municate to the public the success in meeting paid to tracking the progress and the reporting both individual and common goals related to requirements for each project that is awarded the MOU. funding to ensure the deliverables stated in the agreement are routed and received in a timely Conservation Area Planning manner. Conservation Area Planning, originally known GCPEP Benefits Individual as Site Conservation Planning, which is discussed in further detail later in this chapter, Partners represents a tremendous partnership accom- In addition to collective accomplishments, plishment by going beyond thinking within each GCPEP partner has achieved outstand- individual boundaries to thinking at a landing individual conservation successes. The scape level. The completion of a Conservation partnership has played an important role Area Plan allows for more effective manage- in projects by providing assistance, scientific ment and restoration across large landscape ar- expertise, funding, in-kind donations, and eas, according to Compton et al. (2002a). public education support. Important contributions include facilitation of projects us- GCPEP Challenges ing unconventional methods such as cross- boundaries projects like RCW translocation GCPEP has encountered many unexpected between forests. Additional in-kind contribu- challenges, some of which required extensive tions to assist the partners include sharing sup- cooperation to reach solutions. An initial chal- plies, equipment, and personnel required for lenge for the partnership was bringing together support of landscape-scale conservation, such a committed, well-established nucleus of or- as office space and staff, sharing GIS data, burn ganizations. The establishment early on of the prioritization modeling, endangered-species GCPEP Steering Committee to set overall goals management, and road maintenance. and priorities was a challenge but led to a The following section highlights each indi- stronger partnership and much faster success vidual GCPEP partner and briefly describes the on the ground. lands they manage in the partnership. When working through any challenging process it is important that each partner be Department of Defense cognizant of language to remain positive and solution-oriented in conversations, written At 187,548 ha, Eglin Air Force Base holds the documents, and while communicating with largest amount of land of all partners in the 418 IV. Restoration

GCPEP. Undeveloped lands serving as buffers barrier island habitats to the partnership, for military operations contain old-growth lon- which are surrounded by urban development gleaf pine forests and red-cockaded wood- posing significant challenges with prescribed pecker clusters, along with other unique nat- burning and roads leading into the proper- ural communities and species. Eglin projects ties. Ongoing projects in the parks include bird include biodiversity restoration, native plant counts, protection of bird nesting areas, and demonstration areas, and native plantings beach mouse habitat restoration. along roads and streams for erosion control. Eglin has led the way with developing a burn prioritization model and assisting with export- Florida Division of Forestry ing it to other partner lands. Blackwater River State Forest is one of the Naval Air Station Pensacola manages 3409 largest state forest in Florida with 78,779 ha ha of forest, wetlands, shoreline, and outdoor managed in the partnership. Working with recreation areas in the GCPEP. Naval Air Sta- GCPEP, the Blackwater River State Forest has tion Pensacola leadership highlights include increased erosion control efforts by using na- maintaining the regional osprey population tive plants to protect the entire Blackwater with 20 new fledglings produced annually, River watershed. The Division of Forestry has honeybee relocation programs, sea oat plant- improved road management programs that ings for shoreline stabilization, International construct stream crossings to protect water Coastal Cleanups, and Tree City USA designa- quality and aquatic habitat. Successful red- tion on the base. cockaded woodpecker recovery programs have Naval Air Station Whiting Field man- also been implemented such as translocation of ages 3795 ha in the partnership and natu- birds and installation of cavity inserts. ral resource efforts include gopher tortoise Pine Log State Forest with 2797 ha is the old- (Gopherus polyphemus) and flatwoods salaman- est state forest in Florida containing sandhills, der (Ambystoma cingulatum) protection, public flatwoods, cypress swamps, and titi forests lo- nature trails, Tree City USA designation, agri- cated on the eastern border of the GCPEP.Point cultural and timber projects, and regional sup- Washington State Forest with 6170 ha borders port for conservation land purchases. an area of rapid residential and commercial growth along spectacular, fragile coastal areas on the Gulf of Mexico and contains rare species Florida Department of such as white-topped pitcher plants (Sarracenia Environmental Protection leucophylla). The Florida Department of Environmental Protection (DEP) manages 23,176 ha in Florida Fish and Wildlife GCPEP. The Coastal and Aquatic Managed Conservation Commission Areas, a Division of DEP, manages four aquatic preserves. Beneficial efforts include Gulf stur- The Florida Fish and Wildlife Conservation geon (Acipenser oxyrinchus desotoi) studies, Commission (FWCC) manages the parcel of shoreline vegetation restoration, and coastal land and habitat between the Northwest cleanups. The Blackwater River State Park Florida Water Management District and Eglin and the Yellow River Marsh Preserve State Air Force Base known as Escribano Point. Es- Park, also managed by DEP, maintains and re- cribano Point is comprised of 472 ha, which is a stores lands and waters that provide a variety mosaic of habitats including pine and scrubby of recreational opportunities including swim- flatwoods, inshore marine habitat, oak ham- ming, canoeing, hiking, birding, botanizing, mocks, wet prairies, and wetlands. Escribano and camping. Point also includes high-quality submerged Big Lagoon, Tarkiln Bayou Preserve, and plant communities with many rare plant Perdido Key State Parks bring coastal and species. The FWCC has technical knowledge 14. Public–Private Partnership in Restoration 419 of endangered species, animal population sparkling blue waters, fertile coastal marshes, records, bear information, game and nongame and maritime forests all of which are important ecology, wildlife expertise, and prescribed fire, GCPEP conservation targets. This is the most and provides assistance to landowners as well highly visited GCPEP natural area and the Na- as operational support for the partnership. tional Park Service uses the opportunity to focus on resources interpretation and education. International Paper Nokuse Plantation International Paper is a private timber and paper products company that has dedicated 9819 Nokuse Plantation is an ambitious and excit- ha of crucial conservation lands to GCPEP, in- ing project by a private conservation buyer. cluding an important connector parcel link- The project includes 21,448 ha east of Eglin ing Eglin Air Force Base and the Blackwa- Air Force Base that was chosen for the bio- ter River State Forest. This parcel serves as logical significance and the importance of con- a critical wildlife corridor for wide-ranging nectivity to GCPEP as a wildlife corridor and species such as the Florida black bear and pro- to restore highly degraded lands. The objective vides important habitat for the rare Florida bog of the visionary project is to protect region- frog (Rana okaloosae) and flatwoods salaman- ally significant areas that may serve as a critical der. With GCPEP support, the company imple- wildlife habitat for species such as the Florida mented a cooperative gully restoration project black bear and to restore the historical longleaf that helped to protect Florida bog frog habitat. pine ecosystems. Additionally, International Paper has included other important conservation lands within the Northwest Florida Water Blackwater River watershed to GCPEP. Management District National Forests in Alabama One of the five water management districts in Florida, the Northwest Florida Water Manage- Conecuh National Forest in south Alabama is ment District (NWFWMD) is charged with pro- composed of 33,909 ha of longleaf pine habi- tecting watersheds, providing drought control, tat. Conecuh has received recognition for con- and maintaining drinking water supplies in the tinually meeting annual prescribed burning Florida Panhandle. The NWFWMD manages goals. Conecuh sets examples through success- 45,715 ha in the GCPEP area, which serve to ful longleaf pine restoration and monitoring protect rivers, associated floodplains, estuar- projects and the use of native grasses for road ine systems, and wildlife habitat. With GCPEP maintenance and erosion control. The forest assistance the NWFWMD has added to its also protects crucial Gulf sturgeon spawning landholdings, conducted important prescribed areas and red-cockaded woodpecker nesting burns on wetland savannas, and constructed habitats. trail systems for public use and education.

National Park Service The Nature Conservancy More than 80 percent of Gulf Islands National The Nature Conservancy manages 2056 ha Seashore is under water, but the barrier is- with its Perdido River Nature Preserve lands are the most outstanding features to and Choctawhatchee River Delta Preserve. those who visit. The Seashore stretches 170 km Participation in GCPEP has helped The Nature from Cat Island in Mississippi to the east- Conservancy advance its mission to conserve ern tip of Santa Rosa Island in Florida, but biodiversity in northwest Florida and south only the portions within the Florida Panhan- Alabama through community involvement, dle are enrolled within GCPEP which consists landscape-scale conservation and restoration, of 10,034 ha. There are snowy-white beaches, and land acquisitions. 420 IV. Restoration New Partners their organizations. The GCPEP staff does not vote on any topics. The Conservation Area Other important conservation lands may be Plan, which was completed by the Steering added to the GCPEP with unanimous agree- Committee, provides guidance for the staff. ment from the Steering Committee and the To enable adequate operations, a partner- landowners. The Steering Committee has es- ship staff is recommended whose primary fo- tablished the following criteria to admit new cus is the overall partnership. These posi- partners to the GCPEP: tions may include a project director to lead important meetings and coordinate multiple 1. Understands and supports the purposes of projects; scientists to lead restoration, research, the GCPEP and can clearly articulate both and monitoring; and a program manager to what their organization has to gain from facilitate a wide range of administrative and and what they plan to contribute to the financial tasks. This office provides a central partnership. location for functional communication among 2. Meets one or both of the following criteria: the different partners. (a) Manages or owns significant land or wa- Approaching all interactions with flexibility ter holdings in the GCPEP geographic and accommodating such a diverse group en- area with strong preference given to ables the staff to take advantage of vast op- those sharing a border with one or more portunities. Remaining focused on priorities existing GCPEP partners, or that have been identified, while at the same (b) Can offer significant expertise in one time incorporating unexpected changes, pro- or more of the following management vides a dynamic forum for coordinating mul- or conservation disciplines: forestry, wa- tiple projects. While attending to the requests ter and watersheds, wildlife, biodiversity, of one individual partner, it is also essential to prescribed fire, endangered species, or ensure that all of the partners receive a timely nature-based recreation. response when requesting assistance. Depend- 3. Commits to appointing and sending at least ing on the size of the landscape there may be a one, and preferably two, representatives considerable amount of travel involved to ad- to all GCPEP Steering Committee meetings dress all of the partners’ needs. and other functions as needed. Low (1999) stated that the local project staff, 4. Agrees to lead or co-lead one or more coop- particularly the project director, is the single erative GCPEP projects per year. most important element of success and the lo- 5. Agrees in principle to provide financial or cal partnership staff is possibly the most im- operational support to the GCPEP, either portant factor that determines the success of a as direct funds or as in-kind support, and partnership for landscape-scale conservation. agrees to seek additional resources to sup- An ability to multitask and attend to numer- port cooperative projects. ous issues with various degrees of prioritiza- 6. Understands and agrees to adhere to the tion is essential. Some of the qualities include GCPEP operating guidelines. commitment to the future; ability to handle 7. Agrees to keep all appropriate people within risk and uncertainty; ability to form construc- their organization informed and knowl- tive relationships with all kinds of people; and edgeable about the GCPEP purposes and aptitude for problem solving. activities. Partnership Staff Support Staff

Multiple partners contribute the necessary Good project support for the GCPEP has staff to facilitate the GCPEP. The GCPEP staff been critical and extensive. According to Low provides assistance, support, coordination, and (1999), no local partnership, particularly in information to the Steering Committee and the early stages of development, should be an 14. Public–Private Partnership in Restoration 421 island. Each project needs high-level assistance the planning process becomes as important as from a support team. A local project needs to the information used in planning (The Nature be able to call upon experienced ecosystem Conservancy 1998). conservation practitioners to serve as sounding The Conservation Area Plan is broken down boards for ideas, to provide advice and coun- into a Five-S Framework (The Nature Conser- sel, to provide contacts with outside sources of vancy 2000). The Five S’s are: assistance, and to provide hands-on help. The r Systems: the conservation targets at a site local GCPEP staff has received tremendous ad- and the natural processes that maintain ditional support from the partners’ regional of- them fices, providing assistance in numerous areas r Stresses: the causes of destruction, degrada- including conservation science, land protection, or impairment of the systems at a site tion, government relations, communications, r Sources: the agents or activities generating and operations. the stresses r Strategies: the types of conservation activi- Conservation Area ties deployed to abate sources of stress (threat abatement) and enhance or restore the sys- Planning tem (restoration) r Success: measures that monitor the effective- Once a partnership is established, use of a plan- ness of implemented strategies often involv- ning framework tool is highly recommended ing tracking of biodiversity health and threat to help maintain focus, assist with prioritiz- abatement at a site. ing, and make the best use of limited time and funding. Any type of conservation planning For the purposes of this chapter, the plan- must overcome many challenges, one of which ning process will be briefly described. In or- is the need to simultaneously accommodate der to gain a thorough understanding of the many different, sometimes competing, goals, process, it is highly recommended to refer to only one of which may be conserving biodiver- The Nature Conservancy’s “Landscape-Scale, sity. This planning tool also needs to be readily Community-Based Conservation: A Practi- available, reasonably fast, and cost-effective. tioner’s Handbook” (Low 1999). The GCPEP utilized The Nature Conservancy’s The first “S,” Systems, captures the conser- Conservation Area Plan process identified by vation targets at a site. Conservation targets Low (1999). With this planning framework include significant and possibly unique ecosys- tool, the partners were able to determine lo- tems, biological communities, and species. cal threats to the long-term persistence of con- Identifying the appropriate targets is the single servation targets, which include specific focal most important step, since it lays the foun- species and natural communities at a site, and dation for all subsequent steps in the plan- to identify the most important management ning process. The goal is to choose conserva- actions needed to conserve selected conservation targets that represent multiple levels of tion targets. biological organization, have different life his- The Conservation Area Plan approach ex- tory requirements, depend on different eco- plicitly recognizes that humans are part of logical processes, and encompass a variety ecosystems, ecosystems are complex moving of different spatial scales. In effect, planning targets, ecosystem structure and composition targets act as conservation umbrellas or sur- are controlled by processes operating at many rogates, however imperfectly, for all other different spatiotemporal scales simultaneously, target species and natural communities oc- and the scientific community has little un- curring in the geographic area. Thus, tar- derstanding of the structure and function or gets, whether community- or species-level, are life history needs of most of the ecosystems used to cumulatively address the ecological and species that they seek to conserve. Thus, requirements for all species and communities all knowledge is treated as provisional, and occurring at a site. 422 IV. Restoration

After the selection of conservation targets, can be defined as making substantial progress key ecological attributes associated with each toward the long-term abatement of critical of the targets are identified and defined. This threats and the sustained maintenance or en- important step allows for accurate assessment hancement of biodiversity health at sites. Com- of target viability, threats to the targets, and monly, it takes a long time for implementation subsequent strategies identified to abate the of a conservation strategy to manifest in the threats. It also allows the planners to better actual improvement or maintenance of target understand and identify the data gaps and un- viability and health, signified by desired per- certainties associated with the targets and their formance of indicators of biodiversity health. respective key ecological attributes. Therefore, indicators are needed and used to The next two “S’s” in the framework, account for incremental short-term success. Stresses and Sources, combine to determine These indicators can reflect the capacity to im- the threats to a system. In order to develop plement strategies. The three key factors that effective conservation strategies, one must un- can account for early success within a project derstand both the stresses affecting the con- such as a partnership are: project leadership servation targets and their key ecological at- and support, strategic approach, and adequate tributes, and the sources of stress. In this stage funding. of the planning process, after identifying the As a planning tool, the Conservation Area major stresses to the targets, the stresses are Plan should be adaptive in order to make ac- then ranked based upon the severity and scope commodations for individual circumstances. of damage. For each stress there may be one or For instance, GCPEP used the Five-S Frame- more sources of stress. After the major sources work as a foundation, which was modified and of stress are identified, they are ranked accord- built upon to develop the overall GCPEP Con- ing to specific guidelines. The source is what servation Area Plan. When the partners first managers must focus on for threat-abatement established the partnership, one of the first strategies. agenda items for planning was to share in- After having identified and ranked what the dividual land management conservation ob- primary critical threats are to the conservation jectives. From these individual objectives, the targets, conservation strategies are identified partners then collectively prioritized overall based on their ability to abate the threats to GCPEP objectives. It was these conservation key ecological attributes of the conservation objectives and the process used in identify- targets and ultimately improve and/or maining them, as well as the identification of com- tain target viability or health. Strategies can mon challenges and conservation issues that be either threat abatement, which focuses on laid the foundation for the Conservation Area preventing, diminishing, or removing one or Plan process. more sources of stress, or restoration, which The agreed-upon GCPEP objectives in prior- directly enhances or restores the viability of ity order were: the conservation target. When identifying and developing strategies, it is important to first 1. Conserve viable populations of target consider an array of strategic approaches and species then formulate a suite of potential strategies. 2. Introduce relatively natural fire regimes Next, evaluate and rank the potential strate- and protect key ecotypes gies as to their impact and feasibility in order 3. Protect urban interface and reduce to identify the top priorities for immediate fragmentation by use of conservation action. easements The last “S” in the framework is Success. 4. Control erosion in ecologically sensitive Measuring conservation success is an impor- areas tant step in order to monitor whether or not 5. Manage recreation and public access to actions or implemented strategies are having maximize compatibility with conservation the desired and anticipated outcome. Success objectives 14. Public–Private Partnership in Restoration 423 r 6. Increase communication, interaction, and Gulf sturgeon (T) r training among partners Longleaf pine sandhill matrix r 7. Increase inventory and monitoring to fur- Mainland sand pine scrub r ther adaptive management Okaloosa darter (E, e) r 8. Increase public education and stakeholder Pine flatwoods matrix r involvement Red-cockaded woodpecker (E) r 9. Share resources on priority projects Seepage slopes r 10. Secure outside funding and support Steephead stream/slope systems r 11. Inventory and control exotic species Upland game birds 12. Protect aquatic resources 13. Increase understanding of successful eco- T = federally threatened nomic management of longleaf pine t = state threatened 14. Restore and manage the longleaf pine E = federally endangered ecosystem e = endemic to GCPEP lands 15. Recover the red-cockaded woodpecker 16. Manage populations of game species In choosing these 18 primary conservation- 17. Conserve functional community types planning targets, GCPEP deviated from the recommendations in the Five-S Framework. From these objectives, a list of potential conser- The handbook recommends that no more than vation planning targets was identified. Then, eight focal targets be chosen; however, due to using the partners’ knowledge of the GCPEP the large land area that was enrolled in the area and the ecological analysis done by Hard- partnership and the large number of varying esty and Moranz (1999), the partners selected, partner needs, a larger number of targets was by consensus, a subset of targets. These 18 pri- deemed necessary. Also, based upon the part- mary conservation-planning targets included ners’ needs that were identified through the 8 species and 10 natural communities. The objectives, a species target (upland game birds) 8 species were chosen because they were de- was chosen that might not have apparent “eco- clining across their range, they had large area logical” significance. Game birds were chosen requirements (relative to their body size), they because several partners identified this species were found on the majority of GCPEP lands group as one for which they needed assistance or waters, and they would not necessarily be and guidance with regard to population man- well protected through appropriate manage- agement, in order to meet particular land man- ment of natural community-level targets. The agement objectives. 10 communities were chosen because they are Once the targets were identified, GCPEP staff important for facilitating functional ecological met with each partner individually to con- processes and each included many rare, threat- duct threats analyses for the targets that oc- ened, endangered, and/or ecologically signifi- curred on the lands they manage. During these cant species. sessions, which included the partners’ scien- The GCPEP conservation targets identified in tists and managers, partner comments regard- the Conservation Area Plan are: ing specific targets were also incorporated. The GCPEP staff combined each of the individual r Alluvial rivers/floodplains partner target threat analyses into an overall r Barrier island complex GCPEP target threat analysis. This allowed the r Blackwater rivers/floodplains partners to gain a sense of the threats per target r Depression wetlands across the landscape as well as on individual r Estuarine systems properties. The final step was to compile over- r Fish/mussel complex all stresses and sources (overall threats) for all r Flatwoods salamander (T) of the GCPEP conservation targets. This part r Florida black bear (t) of the planning process is a work in progress r Florida bog frog (e) since it is still being determined how to obtain a 424 IV. Restoration more accurate picture by weighting the rank- and maintenance, forestry, and urban devel- ings of the stresses and sources based on the opment (source). number of conservation targets and the num- The GCPEP staff then selected ten strate- ber of partners affected. gies considering all of the partners’ conser- Many threats were identified as being vation objectives, issues, and challenges, and directly or indirectly related to the burgeoning their ability to abate threats to the identified 18 growth of residential and commercial land conservation targets as explained through the uses in the region over the last decade. This threats analyses. The following were identified growth has been forecasted to increase even for each strategy: the overall goal, the partner more so over the next decade. With this contacts for whom GCPEP staff will work, the growth have come increased land and water potential expected accomplishments, and the supply demand, recreational pressures, water conservation targets addressed by the strategy. quality degradation, strain on infrastructure, The issues that the ten GCPEP strategies ad- and other pressures on public and natural dress are as follows: resources. r Inadequate/incompatible fire management According to Hiers et al. (2002), from 1990 r Incompatible development to 2000, the population of the seven-county r Inadequate/incompatible dirt roads, utility GCPEP area increased by 18.5%. In Florida, corridors, culverts, or clay pits management the populations of Escambia, Okaloosa, Santa r Surveying, mapping, and monitoring of con- Rosa, and Walton counties increased by 12, servation targets 19, 44, and 46, respectively. Walton and Santa r Incompatible recreation Rosa counties rank in the top ten fastest grow- r Invasive and native species management ing counties in Florida between 1990 and r Inadequate/incompatible agriculture man- 2000. agement The threats analysis process identified a r Inadequate/incompatible forestry manage- number of primary threats that endanger ter- ment restrial and aquatic targets. These threats can r Internal and community GCPEP communi- be considered “killer threats” to several of the cations and education targets. The biggest terrestrial “killer threat” r Illegal trash dumps identified by the partners was altered fire regime (stress) due to inadequate or incompat- After the strategies were selected, specific ac- ible fire management (source). These sources tion items were identified that would accom- included, but were not limited to, the partners’ plish the overall goals of the strategies. For ability to burn, the seasonality of burns, and each of the strategies, the partners priori- the placement of plow lines. Another terres- tized the actions, which served as the basis trial “killer threat” identified was decreased refor current and future GCPEP projects and productive fitness (stress) due to demographic activities. isolation (source). As was mentioned earlier for the last “S,” The identified aquatic “killer threat” con- Success, early success can be shown with lead- cerned the hydrological and ecological impacts ership and support, a strategic approach, and (stress) that a proposed dam (source) would adequate funding. Early on, the GCPEP part- have within one of the five watersheds within nership concept appeared to be potentially suc- GCPEP. Another aquatic “killer threat” was cessful. Since then, the partnership has proven alteration to the natural hydrologic, chemi- effective in minimizing and eliminating critical cal, and physical characteristics of aquatic sys- threats due to the commitment of the partners tems and subsequent degradation of aquatic to accomplish the top action priorities cho- ecological community and species integrity sen in the Conservation Area Plan. A few of (stress) due to incompatible land use practices the top action project categories include pre- in agriculture, recreation, road construction scribed fire, endangered species management, 14. Public–Private Partnership in Restoration 425 and land protection. They are described below. that will provide prescribed burning assistance Projects can range from assisting individual across GCPEP lands. partners at specific sites, to landscape-scale in A priority conservation objective identified scope, involving coordination with multiple in the GCPEP Conservation Area Plan is the partners. reintroduction of natural fire regimes to protect key ecosystems, embedded communities, and species. The challenges that led to in- Prescribed Fire compatible and inadequate fire management being identified as a “killer threat” included Longleaf-pine-dominated sandhills and flat- insufficient amount of area burned, insuffi- woods provide the matrix within which many cient return of fire intervals, and resistance other communities, such as seepage slopes to growing season burning due to public and depression wetlands, are embedded. These misconceptions. embedded communities require the same pre- Collaborative work at Eglin led to the scribed fire treatments as surrounding sand- development of an innovative landscape dis- hills and flatwoods. Others, such as baygalls, turbance computer model that simulates man- have a less frequent fire return interval, ap- agement of longleaf pine habitats in mod- proximately every 50–100 years. The ex- eled landscapes. The landscape disturbance ceptional diversity of animals and plants in model creates “movies” of expected landscape the GCPEP landscape is a result of frequent changes over time resulting from different fire. For instance, the federally endangered management scenarios. The model identified red-cockaded woodpecker depends on fire- the need to burn on a shorter return interval maintained longleaf pine sandhills and flat- than previously planned. Eglin and the GCPEP woods for foraging within the understory. are collaborating to continue the development The federally threatened flatwoods salaman- of a spatially explicit model that uses GIS data der also depends on fire to maintain the neces- layers to evaluate ecological condition of up- sary ecotone of the depression wetlands where land longleaf pine ecosystems, which will ulti- they breed (Hardesty et al. 1999). According mately help to prioritize management actions to Provencher et al. (2000), plant diversity on across the landscape. Another effort is the de- fire-maintained ecosystems in the GCPEP is velopment of a spatial model that will help pri- very rich: as many as 45 plant species have oritize limited prescribed fire resources to areas been found in 400-m2 plots and at least 293 where fire is most needed. species of plants have been identified within Development pressures are intense across sandhills on Eglin Air Force Base lands. Un- the GCPEP landscape and have led to increas- derstory species richness and cover have been ing wildland–urban interface challenges such positively correlated with insect species abun- as lack of prescribed fire near urban areas dance and biomass. Fire-adapted understory due to public misconceptions and concerns plant species also play an important role in about fires being conducted close to neigh- this ecosystem by carrying the fire that lim- borhoods. The wildland–urban interface chal- its the invasion of competing hardwoods and lenge was highlighted when in 1998 the wild- sand pines. fire season proved to be devastating in Florida: Significant partnership support for fire nearly 2300 wildfires burned almost 202,500 management exists, as evidenced by: pre- ha throughout the state, and more than 300 scribed burning on public lands and co- homes and 30 businesses were damaged. As operative GCPEP prescribed burns, annual a result, greater statewide emphasis has been smoke management meetings, completion of placed on managing the wildland–urban in- a peer-reviewed landscape-disturbance model, terface. The Division of Forestry, along with partner involvement in a fire council, and the GCPEP has created several fire teams to be the start-up of an Ecosystem Support Team proactive in the prescribed fire management of 426 IV. Restoration these wildland–urban interface areas in order landscape the RCW population is increasing to decrease the high fuel loads. due to a cooperative and intensive habitat improvement program ranging from increased lightning season prescribed burning, installation of cavity inserts, and supplementing the Endangered Species population with females from other southeast- Management ern population strongholds. In addition, partners share equipment and training opportuni- Partners are working together to improve habi- ties. tat and recover populations of several rare The RCW population at Eglin Air Force Base and endangered species, including the flat- has increased significantly over the last 10 woods salamander, the Florida bog frog, the years (Moranz and Hardesty 1998). In the Gulf sturgeon, the Okaloosa darter (Etheostoma late 1980s Eglin lost a military test mission okaloosae), and the red-cockaded woodpecker. due to a jeopardy opinion from the U.S. Fish The red-cockaded woodpecker, a medium- and Wildlife Service. The main reason for the sized woodpecker that inhabits open, mature jeopardy opinion was the lack of informa- pine or pine-oak woodlands, was federally tion about the population of two endangered listed as endangered under the Endangered species: the RCW and the Okaloosa darter. This Species Act due to dramatic declines that had loss of a mission sparked the development of occurred across their range. During the previ- a management program that has produced the ous decade the population had also declined fastest growing large population of RCWs. Pop- in the three main population centers within ulations on Eglin Air Force Base lands have the GCPEP landscape: Eglin Air Force Base, grown from 217 active clusters of cavity trees Blackwater River State Forest, and Conecuh in 1994 to 308 active clusters currently. National Forest. These rare woodpeckers have A complete systematic survey of RCW habi- often been labeled “indicators” of a healthy tat was reported in 1993 and a monitor- ecosystem. They depend upon southern pine ing and banding program was established in forests managed well with prescribed fire, mid- 1992. Along with the continued survey and story management, and stand density control. monitoring program, Eglin has developed an Several research studies suggest that RCW intensive management program, which has productivity is directly related to the diver- included constructing over 800 artificial cav- sity and quality of the understory plant–insect ities, translocating over 40 birds, conduct- community. One study at the Savannah River ing growing-season fires, and protecting cav- Ecology Laboratory in South Carolina by Han- ity trees. Eglin has also completed the first ula and Franzreb (1998) observed that up to landscape-level research program to deter- 70% of the prey captured by red-cockaded mine the best combination of management woodpeckers below the canopy was mainly techniques to increase the population. from the soil/litter layer. In addition to healthy At Blackwater River State Forest currently ground cover maintained by regular prescribed all RCWs are banded using color leg bands for burning, the woodpeckers also require old identifying and monitoring the birds’ activity. pines for nest cavity construction. Very few All tree clusters are surveyed for activity yearly. old-growth pine trees remain, another limit- All nestlings and immigrating adults are color- ing factor across the range of the RCW. RCW banded yearly. During the 2001 breeding sea- family groups also need large habitat areas, de- son, 26 of 27 clusters had successful nests. Of fending home ranges of 61–202 ha. the 21 nests where chicks were banded, 42 The local recovery effort for the RCW has nestlings were produced. Of this total, 19 nests been a GCPEP success story. Several of the fledged 34 young, with 13 males and 21 fe- partners have worked cooperatively to re- males. The artificial cavity insert program has verse the RCW’s decline. Across the GCPEP also proven to be successful, with 67 out of 130 14. Public–Private Partnership in Restoration 427 installed inserts currently occupied by RCWs. systems, water resource development and sup- The population has also been augmented with ply, increased public access, public lands man- 24 juvenile RCWs translocated from other agement and maintenance, and increased pro- populations, including Apalachicola National tection of land by acquisition of conservation Forest and Eglin Air Force Base. These inten- easements. sive management efforts have succeeded in The Nature Conservancy, Florida Division of stopping the decline of the RCW population Forestry, and International Paper have worked on Blackwater. In a period of 4 years the RCW closely on numerous Florida Forever projects, population has increased from 13 active groups one of which is Yellow River Ravines, a 6500- and 6 single males, a total of 19 clusters in ha project. The purchase of this important par- 1998, to 33 active groups and no single males, cel would connect the two largest landholdings a total of 33 clusters. in GCPEP, Eglin Air Force Base and Blackwater Conecuh National Forest, in Covington and River State Forest, and also includes a 1600- Escambia counties of Alabama, manages 22 ac- ha inholding in the state forest. The GCPEP tive RCW clusters. Although a smaller popula- Steering Committee has long recognized the tion compared to that of its neighbors to the importance of the Eglin–Blackwater connec- south, it has increased steadily from the 14 tor parcel as a significant conservation land clusters that remained after Hurricane Opal in and as a buffer for Eglin Air Force Base. The 1995. In keeping with the RCW Recovery Plan, property is a Stage 1 Priority Site identified the U.S. Forest Service maintains an inventory in The Nature Conservancy’s East Gulf Coastal of both active and inactive nesting sites, mon- Plain Core Team (1999). The Florida Game and itors nesting activity, bands fledglings each Freshwater Fish Commission has also identi- spring, provides “recruitment” habitat by in- fied it as an important conservation area in the stalling artificial cavities, and maintains ex- report “Closing the Gaps in Florida’s Wildlife isting habitat by prescribed burning approxi- Habitat Conservation System” by Cox et al. mately 10,000 ha each year. In addition, the (1994). The property includes important trib- agency is developing future habitat for the utaries of the Yellow River that protect water RCW by actively working to restore the native quality and species diversity and which pro- longleaf pine ecosystem. vide habitat for several rare species. Other important parcels in the project area buffer Black- water River State Forest and Naval Air Station Land Protection (NAS) Whiting Field. The following are significant reasons to pro- Land protection was chosen as a high-priority tect this and other GCPEP area lands long term: strategy by the GCPEP Steering Committee due to the large number of inholdings, buffers, and 1. Military Mission—Protecting the mili- connectors needed to protect the biological di- tary mission is dependent upon ensuring versity of the GCPEP landscape over the long adequate acreage for military training. term. In addition, the partners recognized the Encroachment along the boundaries of a benefit these lands would provide concern- military base can have negative impacts ing prescribed burning, especially by reduc- on mission capacity through restrictions ing smoke management concerns and urban– on low-level flights over developed areas wildland interface issues. or noise concerns from neighbors. The The Florida Forever Program is the state’s connector parcel is adjacent to an im- blueprint for conservation of the unique nat- portant Ranger training area along the ural resources and is the largest program of Yellow River and an outlying field for NAS its kind in the United States. The program Whiting Field. Protecting this land would encompasses a wide range of goals, includ- also provide long-term habitat for several ing: restoration of damaged environmental rare species. Increasingly, as habitat around 428 IV. Restoration

bases is developed, more habitat demands vation corridor will link Eglin Air Force Base fall upon the bases themselves. It then to Apalachicola National Forest and the Gulf becomes more and more difficult to meet of Mexico. All of these projects were strongly the military mission while supporting the supported by GCPEP, other state and federal rare species displaced from surrounding agencies, the local county commission, envi- developed habitats. ronmental and recreational organizations, and 2. Conservation Significance—The Yellow the general public. River Ravines land would serve as a critical GCPEP has been instrumental in moving wildlife corridor between Eglin Air Force land and water management and land protec- Base and Blackwater River State Forest. tion from being controversial community is- Rare species include the Florida bog frog, sues, historically, to the present, being more flatwoods salamander, and Florida black strongly supported issues. This has, in part, bear. Natural communities of significance been due to a tremendous education effort include steephead stream/slope systems aimed at community leaders and politicians. and depression wetlands. The rare species The GCPEP staff has served as an impor- and communities found on the connector tant communication and support link be- parcel are also being managed for recovery tween the partnership and the surrounding on Eglin Air Force Base. communities. 3. Water Quality/Quantity—Several creeks in the corridor, including Weaver, Garnier, and Julian Mill creeks, feed the Yellow River. Conclusion This area serves as a water recharge area for Santa Rosa County. Planning for water Large-scale restoration of the longleaf pine recharge is important in an area that is de- ecosystem may be more effective if public and pendent upon water from the shallow sand private landowners choose to work together and gravel aquifer. Protecting water qual- in landscape-level partnerships. When part- ity and quantity is important for biodiver- ners who restore and manage longleaf pine sity protection, providing water supply, and use science-based planning as a common goal, protecting military training that is depen- partnerships can succeed though individual dent upon adequate water flow in the Yel- partner missions may vary widely. Comple- low River watershed. tion of a Conservation Area Plan allows for 4. Recreation and Hunting—As the region more effective and efficient management and continues to grow and develop, recreational restoration across large landscape areas. Given space will become more limited and delimited funding for personnel, equipment, and mands on remaining space will increase. projects, this method of planning is recognized The Yellow River Ravines connector par- by many to leverage strategies to accomplish cel, located in Santa Rosa County, could short- and long-range goals. decrease future recreational demands on The success of any partnership depends on Eglin and provide increased recreational op- respect and cooperation and may operate more portunities for area residents and visitors efficiently with a staff dedicated to the effort. alike. More may be accomplished when combining expertise and resources to effectively man- The Yellow River Ravines project was ap- age individual lands, while at the same time proved by the State of Florida as an “A-ranked” meeting the challenges of sustaining larger Florida Forever Project, assuring funding for ecosystems. By doing this, GCPEP serves as an the project. Additional projects, which protect example of how organizations can work to- and buffer important conservation lands, were gether to achieve common and important goals also approved by the State of Florida including such as restoring and maintaining the longleaf the Northwest Florida Greenway Project. This pine ecosystem. This chapter may provide a 100-mile-long and 5- to 10-mile-wide conser- “blueprint” for partnerships to set conservation 14. Public–Private Partnership in Restoration 429 and restoration objectives and priorities for Coastal Plain Ecoregional Plan. Chapel Hill, NC: both the individual partners and the collective The Nature Conservancy. partnership. Hanula, J.L., and Franzreb, K.E. 1998. Source, distribution and abundance of macroarthropods on the bark of longleaf pine: Potential prey of the red- cockaded woodpecker. For Ecol Manage 102:89– Acknowledgments 102. Hardesty, J.L., and Moranz, R.A. 1999. Lon- We extend gratitude to each of the GCPEP gleaf pine ecosystem restoration in northwest partners’ Steering Committee Representatives Florida sandhills: Issues and recommendations. and to their organizations for their commit- Gainesville, FL: The Nature Conservancy. ment to, and support of, the GCPEP. The time Hardesty, J.L., Moranz, R.A., Woodward, S., and Compton, V. 1999. The Gulf Coastal Plain Ecosys- dedicated by each of the partners is vital to tem Partnership: An assessment of conservation the strength and success of the partnership. opportunities. Gainesville, FL: The Nature Con- We are also grateful to the many individu- servancy. als, the governmental and nongovernmental Hiers, S., Bachant, J., Compton, V., and Penniman, agencies, and the other essential organizations, P. 2002. The Gulf Coastal Plain Ecosystem Part- which contribute generous amounts of time, nership: Freshwater ecosystem demonstration. expertise, and resources to make the GCPEP an Jay, FL: The Nature Conservancy. effective conservation tool. Low, G. 1999. Landscape-scale, community-based conservation: A practitioner’s handbook. Arling- ton, VA: The Nature Conservancy. References Moranz, R.A., and Hardesty, J.L. 1998. Adap- tive management of red-cockaded woodpeckers Compton, V. 2002. Gulf coastal plain ecosystem in northwest Florida: Progress and perspectives. partnership—Achieving results through cooper- Gainesville, FL: The Nature Conservancy. ation. Florida Forests 6 Issue 3:22–25. Provencher, L., Litt, A., Gordon, D., and Tanner, G. Compton, V., Bachant, J., Hiers, S., and Penniman, 2000. Reference condition variability: Product to P. 2002a. Gulf Coastal Plain Ecosystem Partner- Eglin Air Force Base, Natural Resources Division. ship: Site conservation plan. Jay, FL: The Nature Gainesville, FL: The Nature Conservancy. Conservancy. The Nature Conservancy. 1998. An approach for Compton, V., Bachant, J., and Penniman, P. 2002b. conserving biodiversity at portfolio sites: Site con- Gulf Coastal Plain Ecosystem Partnership: Steer- servation planning. Arlington, VA: The Nature ing committee manuals. Jay, FL: The Nature Con- Conservancy. servancy. The Nature Conservancy. 2000. The five-s frame- Cox, J., Kautz, R., MacLaughlin, M., and Gilbert, T. work for site conservation. Arlington, VA: The 1994. Closing the gaps in Florida’s wildlife habi- Nature Conservancy. tat conservation system. Florida Game and Fresh The Nature Conservancy. 2005. gulf coastal plain Water Fish Commission, Tallahassee, FL. ecosystem partnership. Altamonte Springs, FL: East Gulf Coastal Plain Core Team. 1999. East Gulf The Nature Conservancy. Index

2,4-D, 307, 309 Augmentation, 6, 336, 338 Bluestem, 4, 68, 71, 98, 135, 138, 3-in-1 plow, 273 definition, 336 141, 298 Autecology, 4, 231 Broomsedge bluestem, 68 AccordTM, 284 Mohr’s bluestem, 68 Adaptive management, 8, 297, 320, Bare-root nursery seedlings, 112 Bobwhite quail, 124, 193, 199, 217, 346, 353, 360, 423 Bark beetles, 43, 153 231, 237, 246, 271, 300, 405 Age classes, 124, 220, 221, 228, 242, Basal area, 104, 105, 109, 117, 120, Bottlebrush, 105 243, 246, 248, 278, 279, 281, 123, 141–150, 220, 222, Broomsedge bluestem, 68 291 225–238, 243, 246, 247, Brown spot needle blight, 276, 326 Alabama, 3, 9, 10, 23, 25, 26, 28, 31, 253–256, 262–264, 272, 277, Bulldozing, 273 36, 54, 55, 78–80, 90, 95, 97, 279, 282, 286, 304, 308 Burning, 15, 20, 34, 43, 44, 106, 102, 171, 172, 181, 199, 200, BD-q, 120–124, 225, 226, 230, 233, 110, 112, 114, 118, 122, 125, 251, 254, 257, 271, 277, 302, 235, 248, 279, 282 139, 179, 187, 194, 196, 200, 328, 359, 364, 365, 377, 379, Bedding, 193, 196, 273, 274, 285 233, 234, 243, 248, 274, 276, 380, 383, 392, 413, 415, 419, Beech, 11, 16 280, 286, 287, 289, 290, 300, 427 Ben Cone, 407 303–319, 326–328, 330–332, Allelopathy, 136, 138 Biodiversity, 4–7, 51, 52, 190, 200, 339, 346, 348, 357, 369, 383, American bison, 179, 185, 198 300, 305, 306, 311, 339, 340, 386, 395, 418, 419, 425–427 Amphibians, 158–161, 166–169, 369, 379, 380, 383, 384, 174, 175, 179, 183, 189, 193, 388–395, 403–407, 414, Cabbage palmetto, 71 194, 196–199, 337, 338, 358, 418–422, 428 Candle stage, 328 362–364, 383, 387 Biogeography, 381 Canis rufus floridanus, 198 resident amphibians of longleaf Biological legacy, 302 Canopy gaps, 98, 110, 111, 115, pine savannas, 161 Birds, 32, 33, 38, 109, 112, 118, 158, 120–123, 125–127, 139, 289 Andropogon mohrii,68 159, 171–177, 179, 183–191, Carbon credits, 408, 409 Andropogon virginicus, 68, 73 197–199, 271, 272, 280, 289, Carolina Sandhills National Wildlife Angelina National Forest, 198 331, 336–365, 405, 418, 423, Refuge, 304, 316, 317, 357 Aquifer impacts, 393 426 Carrying capacity, 31–33 Arboreal, 158, 166, 172, 174, 178, characteristic birds of longleaf pine Carry-log, 27, 38 191, 194 savannas, 176 Carya tomentosa,16 Aristida beyrichiana, 4, 15, 57, Bison bison, 179, 185, 198 Castanea dentata,9 60–68, 72–76, 78, 81–88, Black bear, 180, 181, 184, 338, 363, Catastrophic events, 72, 340, 343 135, 217, 307, 310, 315, 386, 397, 415, 419, 423, 428 Catena, 384 330 Black Water River State Forest, 418, Catkins, 99, 100 ArsenalTM, 283, 284 419, 426–428 Cattle, 3, 18, 31–34, 107, 109, 186, Artificial regeneration, 111, 135, Blackgum, 70, 161 337, 408 272, 279, 289 Blackjack oak, 71, 284, 380 Chapman oak, 70 direct seeding, 111 Bluejack oak, 71, 284 Charcoal root rot, 283

431 432 Index

Chestnut, 9, 71 area planning, 417, 421 biodiversity, 4–7, 51, 52, 190, 200, Chestnut oak, 71 landscape-scale, 414 300, 306, 311, 339, 340, 379, Chopper, 285, 303, 307 partnerships in, 414 388, 391, 392, 394, 395, 403, Classification, 51–91, 226, 367, 396 strategies, 58, 335–337 404, 418–422 ecological, 51–91 Conservation biology, 345, 378, faunal, 5, 157–201, 337 national vegetation, 52, 58, 66, 90, 389 plant, 37, 136, 383 396 Conservation Reserve Program understory, 37, 327 Clayey and rocky uplands, 59 (CRP), 153, 406, 410, 411 vertebrate, 5 Clearcutting, 95, 115, 116, 124–219, Containerized nursery seedlings, Donor population, 336, 342, 344, 227, 290, 305, 381 113 347–349, 353, 357–359, 361, alternative-strip, 115 Cornus florida, 62, 69, 77, 380 364 progressive-strip, 115 Cotton mouse, 165, 180 Drum chopping, 273, 303, 306–308 Climate change, 187, 188, 388, 393, Cotyledons, 102 Dwarf live oak, 71 408 Cougar, 180, 182, 184, 397 Climax, 194, 303–306, 310, 337 Cronartium quercuum,99 Eastern Indigo snake, 168, 171, 185, Coastal Plain, 3–15, 21, 26, 31, 44, Crossett Experimental Forest, 223, 199, 338, 363, 365, 383, 387 51–56, 59–67, 72–90, 97, 101, 224, 226 Ecological functions, 125, 390 107, 108, 157–163, 166, 169, Cruise, 288 Economic benefit, 395 171–176, 179, 181, 185, 187, Ctenium aromaticum, 57, 64, 65, 66, Economics and policy, 403 188, 190, 194, 195, 197, 198, 69, 81, 85, 87, 88, 89, 315 Ecoregions, 53–55, 59, 338, 340, 200, 218, 223, 228, 229, 238, Cutting cycle, 121–124, 219–223, 348, 392 242, 246, 252, 256, 258, 298, 227–238, 277, 280 Atlantic Coastal Plain, 53 315, 330, 343, 359, 364, 365, Eastern Gulf Coastal Plain, 55 377, 380, 403, 413, 414, 415, Darling National Wildlife Refuge, Fall-line Sandhills, 53, 54 427 198 of the longleaf ecosystem, 53 Cogon grass, 310, 317 Darlington oak, 71 Piedmont and Montane Uplands, Colinus virginianus, 124, 176, 217 DBH class, 224, 226 55 Columbian mammoth, 185, 187 Demographic factors, 352, 355 Southern Coastal Plain, 54, 55 Common names, 68, 72, 91, 160 Demographic isolation, 424 West Gulf Coastal Plain, 55 Community succession, 6, 116 Dendroctonus spp., 153 Ecosystem cluster, 384 Competition, 5, 32, 46, 95, 98–124, Density management diagram, 263 Ecotones, 195, 377, 379, 383, 384, 135–154, 195, 221, 228, 233, Department of Defense, 391, 417 388–390, 397 236, 238, 246, 251, 252, 261, DeSoto, 17, 30 Eglin Air Force Base, 73, 75, 306, 272–274, 276–278, 281, 283, Dionaea muscipula, 15, 69, 81 389, 391, 392, 404, 417–419, 285, 286, 288, 303, 309, 311, Direct seeding, 111, 112, 272, 280, 425–428 316, 319, 320, 331, 332, 354, 305, 310, 311, 316, 318, 320 Endangered Species Act, 6, 335, 338, 381, 405 Disking, 273, 317 396, 404, 406, 426 apparent, 136–138, 151 Dispersal, 95, 100, 102, 108, 112, Endangered species management, asymmetrical, 136, 142 115, 119, 165, 174, 180, 189, 424, 426 Composition, 4, 6, 7, 14, 46, 51, 52, 302, 304, 313, 331, 336, Endemic species, 90, 159, 162, 169, 57, 59, 67, 80–87, 97, 115–117, 340–346, 349, 354, 358, 298 138–141, 158, 159, 178, 188, 360–362, 385, 387, 388, 396, Entisol, 56 221, 298, 299, 300, 303, 305, 405 Environmental benefits index (EBI), 306, 310–313, 327, 330, 331, seed, 95, 100, 108, 116, 186 410 350, 355, 369, 379, 380, 388, Disturbance, 7, 95–99, 110, 119–127, Environmental stochasticity, 421 135, 138, 196, 199, 218–221, 339–341 altering species, 310 242, 273–275, 284, 299, 304, EPA Southeastern Ecological Cone production, 95, 100, 101, 108, 307, 377, 380, 383, 425 Framework, 392, 393 118, 232 natural, 124–127 Escambia Experimental Forest, 102, Conecuh National Forest, 419, 426, regimes, 95, 125, 135, 377 231, 234, 264, 277 427 silviculture that mimics, 124–127 European settlers, 13, 16 Connectivity, 341, 378, 380, Disturbance dynamics, 95, 97, 125 Evosystem, 377, 378 387–390, 396, 419 Diversity, 4, 5, 37, 38, 51, 52, 56, 57, Exotic plants, 124 Conservation, 3, 4, 6, 23, 52, 58, 90, 67, 73, 76, 80, 81, 84, 87, 90, 115, 125, 127, 153, 200, 242, 97–99, 122, 125, 126, 136, 142, Facilitation, 135–153, 417 245, 251, 305, 306, 311, 313, 144, 145, 154, 157–159, 161, Fagus grandifolia,16 335, 336, 338–340, 342, 344, 163, 165–201, 218, 242–246, Faunal diversity, 5, 157, 159, 161, 345, 359, 364, 365, 368, 369, 252, 297–380, 383, 388, 390, 163–201, 337 377–379, 388–396, 403–429 397, 405, 406, 411, 425–427 Felling, 236, 256, 306, 307 Index 433

Fence laws, 33 Florida Fish and Wildlife Grass, 4, 5, 16, 29, 32, 38, 51, 57, 59, Feral hogs, 3, 31, 32, 96, 365 Conservation Commission, 90 69–77, 86, 90, 99, 102–110, Feral livestock, 29 Florida mouse, 180, 181, 199 114, 115, 118, 135, 138–141, Fire, 4, 6, 10–16, 19, 23, 29–38, 44, Florida red wolf, 198 146–151, 161, 165, 166, 174, 46, 52–313, 319, 320, 326–330, Flowering dogwood, 69, 302, 380 178, 187, 231, 274, 276, 277, 337, 339, 350, 365, 377, Forest certiﬁcation, 392 280, 288, 304, 310, 315, 317, 379–381, 383–391, 419, 420, Forest structure, 117, 123, 242, 337, 330, 331, 405 422, 424–426 348, 350 Bahia, 274, 280, 305 dormant season, 152, 153, 286, Form class, 258, 259 cogon grass, 310 287, 290, 303, 307, 308 Fort Benning, 357 Indiangrass, 71 effects of, 13, 95, 105, 106, 159, Fort Bragg, 80, 87, 357, 391 lopsided Indiangrass, 71 194, 196, 387 Fort Stewart, 74, 316, 357 switchgrass, 70 exclusion, 4, 11, 32, 34–38, 97, Fossil record, 179, 183, 186 toothache grass, 69 108, 195, 290, 300–309, 365 Fossorial, 160, 169–173, 181, warm season, 261 frequency, 14, 15, 37, 52, 84, 158, 191–193, 198 wiregrass, 4, 15, 34, 38, 57, 59, 67, 298, 300–304, 326, 380 Fox squirrel, 179–181, 199, 271, 68, 73, 74, 76, 78, 80, 98, 135, growing season, 13, 35, 196, 338, 363, 383, 386–389, 391, 138, 141, 165, 166, 173, 174, 285–288, 290, 291, 303 396, 397 195, 200, 217, 284, 307–309, history, 178, 272, 288, 290–292, Frogs, 5, 159–167, 172, 173, 190, 313–320, 330–332, 380, 381, 303, 312 191, 193, 196, 198, 362–364 396 prescribed, 102, 108, 109, 112, list of frogs of longleaf pine Grass stage, 5, 29, 32, 99, 103–110, 118–124, 135, 141, 142, 153, savanna, 161 114, 115, 118, 139, 146, 150, 154, 196, 199, 200, 217, 231, Fuel, 15, 27, 29, 59, 98, 109, 110, 151, 231, 276, 277, 288, 405 232, 234, 235, 236, 243, 246, 114, 140, 194, 248, 277, Greenwood plantation, 244, 245 272, 277, 278, 280, 281, 285, 281–291, 297, 300, 307, 312, Ground layer restoration, 297 289, 305, 306, 307, 310, 319, 326–331, 426 Ground sloth, 184, 187 320, 327, 350, 386–391, 395, Functional landscape, 384, 390, 414 Group selection, 95, 111, 120–127, 419, 420, 424–426 141, 201, 221, 222, 225, 227, regime, 10 Gallberry, 69, 163, 166, 174, 195 230–238, 279, 280, 291 suppression, 13, 29, 34, 37, 52, 59, Gap-phase regeneration, 110, 111, Growth and yield, 5, 124, 235, 236, 73, 80, 86, 97, 178, 196, 284, 120 251–264 300, 307, 337, 339, 365, 381, Genetic considerations, 312, 341 of natural stands, 5, 253–256 385, 389, 391 Georgia, 4, 23–28, 32, 36, 37, 53–58, planted stands, 5, 256–259 Fire ant, 199 67, 68, 70, 76–87, 95, 97, 114, Gulf Coastal Plain Ecosystem Flatwoods, 3, 15, 56–59, 64, 73, 75, 124, 157, 159, 171, 178, 181, Partnership, 6, 413–415 81–86, 97, 103, 104, 123, 198–200, 237, 242, 246, 254, 157–166, 169, 171–174, 257, 260, 298, 303, 310, 313, Habitat conditions, 126, 231, 306, 178–181, 189, 193, 196–199, 316, 326, 328, 330, 332, 359, 330, 346, 348, 349 298, 307, 308, 310, 313, 327, 362, 365, 377, 379, 380, 392 Habitat Conservation Plans (HCPs), 328, 337, 338, 362–364, 377, Georgia oak, 70 406–408 379, 383, 387, 392, 418–428 Germination, 4, 95, 102, 103, Habitat fragmentation, 197, 337, Flatwoods salamander, 161–163, 106–108, 111–114, 140, 146, 344, 387, 393, 395 189, 193, 196, 198, 338, 362, 153, 175, 186, 222, 272, 278, Haliaeetus leucocephalus, 338, 366 363, 383, 387, 418–428 280, 291, 311, 313, 315–319, Heartwood, 18, 43, 48, 191, 192 Florida, 3, 4, 13, 16, 17, 23, 25, 26, 381 Herbaceous understory, 109, 271, 28, 33, 36, 43, 46, 47, 51, factors affecting, 315 285, 289, 291 54–58, 62, 68–78, 81–84, 87–90, Girdling, 18, 153, 306, 307 Hexazinone, 142, 143, 152, 275, 105, 110, 114, 138, 157, 159, Glyphosate, 142, 152, 275, 280, 284, 284–286, 307–309 163, 167–218, 242, 246, 251, 307, 309 High-grading, 124 254, 298, 307, 310, 313, Gopher frog, 161, 164–167, 189, Hunting, 25, 124, 184, 187, 242, 317–320, 335, 338, 339, 342, 191, 338, 362–364, 387 245, 248, 330, 365, 395, 405, 344–346, 356, 359, 361–369, Gopher tortoise, 5, 165, 168, 407, 410, 411, 428 377–397, 403–409, 413–428 171–175, 181, 185, 186, Hurricane damage, 405 Florida Department of 191–199, 246, 271, 331, 332, Hybridization, 99 Environmental Protection, 415, 338, 358, 363–365, 383, 387, Hypocotyl, 102 418 391, 418 Florida Division of Forestry, 415, Gopherus polyphemus, 168, 175, 271, Ilex coriacea, 64, 69 418, 427 338, 358, 383, 418 Ilex glabra, 64, 65, 69, 77, 83, 84, 90, Florida Ecological Network, 392 Graminoid, 75, 309 163 434 Index

Ilex vomitoria, 63, 64, 66, 69, 84 284, 285, 291, 339, 380, 404, 231–234, 238, 273, 277, 288, Imazapyr, 142, 152, 275, 280, 284, 405 289, 389, 405 285 Logistic factors, 353 Nature conservancy, 91, 124, 200, Imperata cylindrica, 310, 317 Longleaf alliance, 275 310, 320, 338, 391, 413, 415, Indiangrass, 71 Lopsided Indiangrass, 71 419, 421, 427 Indians, 14, 15, 18, 24, 25, 31 Louisiana, 22, 25, 26, 28, 30, 36, 51, Naval stores industry, 3, 18, 19, 21, Indicators, 5, 7, 8, 72, 74, 178, 422, 55, 75, 78, 79, 89, 101, 157, 43, 44, 46, 193, 339 426 169, 172, 187, 199, 232, 251, Nokuse plantation, 391, 415, 419 Inkberry, 69 257, 259, 261, 262, 328, 338, North Carolina, 3, 4, 9, 13–37, 43, Interactions, 5, 98, 135–151, 300, 359, 363, 364, 366, 377 51–53, 57, 58, 67, 80, 81, 85, 309, 336, 351, 367, 377–380, 88, 90, 101, 139, 163, 169, 171, 383–385, 388, 390, 414, 420 Magnolia grandiflora, 70, 84 173, 199, 200, 301, 311, 319, case studies, 142 Magnolia virginiana, 65, 70 343, 344, 359, 367, 391, 392, overstory, 137 Mammals, 158, 159, 167, 172, 174, 407 understory, 135 179–191, 197, 199, 280, 289, Northern Bobwhite Quail, 193, 199, Interference, 136 331, 336, 338, 358, 362, 363, 237, 246 International Paper, 415, 419, 427 365, 366, 405 Northwest Florida Water Invasion, 34, 97, 169, 195, 199, 279, characteristic mammals of longleaf Management District, 415, 418, 425 pine savannas, 180 419 Inventory, 13, 58, 76, 122, 124, 157, Marking rules, 230 Nurse crop, 285 223–226, 236, 237, 246, 275, Mastodon, 186 Nyssa sylvatica, 70, 81, 161 288, 289, 344, 423, 427 Mesic sites, 57, 74, 77, 110, 111, Ips spp., 153 284, 285, 307, 309 Oak, 15, 16, 18, 22, 34, 57, 59, 67, Metapopulation dynamics, 189, 197 70–73, 76, 83, 84, 90, 105, 139, Kalmia latifolia, 63, 69, 79 Metapopulation theory, 344, 345 141, 159, 161, 163, 165, 166, Keystone species, 5, 98, 135, 138, Microhabitat, 163, 165, 192, 350, 363 169–175, 181, 187, 193, 284, 175, 178, 198, 338–340 Midstory, 38, 108, 115, 136, 141, 290, 301, 307, 308, 327, 330, Kleptoparasites, 348–350, 352 178, 221, 237, 239, 243, 337, 364, 365, 380, 383, 384, Kudzu, 124 246–248, 284, 307, 308, 386, 391, 418, 426 326–328, 331, 332, 339, 350, Odocoileus virginianus, 179, 180, 185 Laddering of fuels, 287 404, 426 Old growth, 151, 253, 300, 389 Land conservation, 394 Military mission, 427, 428 OustTM, 283 funding for, 394 Mississippi, 15, 17, 25, 28, 36, 52, Overstory, 4, 6, 59, 98, 104, 105, Landowner Incentive Program (LIP), 55–58, 73, 75–81, 84, 89, 159, 107, 109, 111, 115–126, 406, 410 165, 172, 173, 193, 200, 254, 135–154, 220, 229, 231–236, Landowner programs, 392 257, 261, 262, 309, 328, 338, 246, 248, 271–292, 331, 332, Landscape, 6–9, 13–18, 29–38, 51, 359, 362–365, 392, 419 339, 380 54–56, 76–81, 84, 86, 89, 96, Mockernut hickory, 16, 302, 383 common names, 68–72 97, 115, 119, 120, 157, 160, Mohr’s bluestem, 68 scientific names, 68–72 163, 166, 175, 189, 195, 197, Monitoring, 7, 8, 108, 109, 200, 243, 200, 242, 254, 271, 278, 299, 278, 331, 336, 337, 346, 347, Panicum virgatum, 62–66, 70, 76, 77, 300, 312, 330–332, 337, 339, 352, 353, 357, 361, 362, 365, 80, 90 367, 377–397, 403, 413–428 368, 369, 388, 419, 420, 423, Partial restoration, 300 Landscape ecology, 6, 378, 379, 381, 424, 426 Pasture, 3, 11, 25, 26, 32, 33, 36, 388–390 Mosaic, 13, 15, 16, 97, 115, 120, 193, 274, 280, 283, 305, 316, conservation strategies, 389–391 122, 200, 316, 379, 386, 418 317, 405, 408, 409 Large gallberry, 69 Mountain laurel, 69, 79 Payne’s Prairie, 198 Liberation cutting, 228, 236 Mutualism, 138 Pensacola, 17, 418 Lightwood, 18–20, 27, 44 Mycosphaerella dearnessii, 99, 276 Peromyscus polionotus, 165, 193 Liquidambar styraciflua, 65, 70, 141, Myrtle oak, 71 Physiognomy, 51, 59, 337, 380, 384 163, 284, 380 Picloram, 275 Lizard, 159, 167–174, 190–194, National Forests in Alabama, 415, Picoides borealis, 176, 178, 217, 335, 364 419 338, 386, 404, 415 characteristic lizards of longleaf Native Americans, 13, 15 Pin oak, 71 pine savannas, 168 Natural Heritage Program, 58, 91, Pine straw, 34, 271, 301, 307, 312, Loblolly pine, 16, 18, 29, 34–37, 70, 200 316, 317, 319, 405, 406, 96, 99, 120, 123, 139, 140, 144, Natural regeneration, 95, 97, 101, 409–411 153, 166, 169, 170, 176, 179, 108, 109, 111, 115–120, 126, Pinus echinata, 16, 63, 70, 76, 80, 99, 194, 218, 228, 232, 247, 262, 140, 217, 218, 227, 228, 286, 339, 380 Index 435

Pinus elliottii, 64, 65, 70, 82, 87, 287, Post oak, 16, 71, 139, 284, 327, 380 246, 289, 306, 330, 335, 336, 339, 404 Postplanting management, 286 338–369, 386–391, 397, 404, Pinus serotina, 16, 64–66, 70, 81 Prairies, 15, 79, 83, 84, 90, 158, 173, 415, 418, 419, 423–426 Pinus sondereggeri,99 179, 183, 187, 303, 418 demographic, 340–344 Pinus taeda, 60, 62, 66, 70, 96, 139, Preparatory cut, 117–119, 289 diseases and parasites, 358, 359 284, 339, 380, 404 Prescribed burning, 110, 114, 118, dispersals, 344, 345 Pinus virginiana, 70, 81, 343 122, 125, 139, 179, 194, 200, geographic variation, 359 Pinus, 3, 15, 16, 43, 48, 60–67, 70, 233, 248, 306, 307, 326, 330, metapopulation theory, 344, 345 74, 76, 79, 80–87, 95, 96, 99, 331, 332, 339, 346, 348, 357, sociobiology, 343, 344 135, 138–140, 150, 157, 163, 369, 395, 418, 419, 425–427 translocation, 339–361 176, 177, 186, 217, 246, 251, negative impacts, 328 Reference communities, 5, 38 284, 286, 287, 337, 339, 343, precautions, 328 Reference information, 299 377, 379, 380, 383, 403, 404 role of, 326 Reference sites, 299 Pioneer species, 4, 285 schematic, 327 Reference systems, 6 Pitch kettles, 18 techniques, 326 Regeneration, 4, 29, 30, 32, 35, Plant diversity, 379, 383, 425 Prescribed fire, 102, 108, 109, 112, 95–127, 135, 140, 141, 153, Plantation, 5, 11, 13, 17, 27, 36, 119, 118, 120, 122, 124, 135, 141, 178, 187, 195, 217–248, 272, 135, 138, 139, 144, 154, 163, 142, 153, 154, 196, 199, 200, 273, 277–281, 284, 286–305, 165, 193, 196, 197, 237, 244, 217, 231–236, 243, 246, 272, 332, 389, 405 245, 247, 257, 260, 264, 300, 277, 278, 280, 281, 285, 289, artificial, 111, 135, 272, 279, 289 301, 303, 304, 309, 310, 391, 305–307, 310, 319, 320, 327, ecology, 95–127 411, 415, 419 350, 386–391, 395, 419, 420, environment, 95, 98 first pine, 36 424–426 natural, 95, 97, 101, 108, 109, Greenwood, 244, 245 dormant season, 286 111, 115, 116, 118, 120, 126, Nokuse, 391, 415, 419 growing season, 118 140, 217, 218, 227, 228, pine, 4, 11, 13, 36, 136, 139, 142, Pyrogenic, 141, 194, 337 231–234, 238, 273, 277, 288, 153, 154, 163, 165, 169, 191, 289, 389, 405 196, 197, 262, 263, 300, 304, Q factor, 123 problems, 95 305, 309, 310, 391, 411 Quercus, 16, 57–65, 67, 70–84, 89, requirements, 108 Red Hills, 248 139, 141, 159, 284, 285, 290, Regional conservation planning, 394 restoration, 300 307, 326, 380 Regulation, 5, 120–124, 222–227, Planting stock, 114, 233 Quercus chapmanii, 61, 64, 67, 70, 75, 230–238, 407, 411 Pleistocene (Rancholabrean) 84 area-based, 227 vertebrates, 184 Quercus coccinea, 63, 70, 81 stand structure, 225–227 Pocket gopher, 107, 171, 173, Quercus geminata, 57, 73, 84 volume, 223–225 180–182, 191, 193, 198–200, Quercus georgiana, 63, 70 Rehabilitation, 228, 390 383 Quercus hemisphaerica, 60, 61, 71, 73 of understocked conditions, 228 Policy, 6, 355, 390, 403, 405, 406, Quercus incana, 16, 59–62, 72–76 Reineke stand density index, 262, 409, 410 Quercus inopina,71 263 Policy options, 406 Quercus laevis, 16, 57, 59–62, 67, 71, Reintroduction, 6, 37, 38, 285, 287, Pollination, 95, 99, 100, 312 73–76, 159, 284, 290, 307, 326 328, 332, 335–369, 425 Pond cypress, 161 Quercus margarettiae, 60, 61, 71, 77 definition of, 336 Pond pine, 11, 16, 70 Quercus marilandica, 16, 61–63, 71, Release method, 337, 349, 355, 356, Population, 17, 31, 95, 117, 126, 78–80, 139, 284, 380 361 159, 167, 171, 172, 174, 182, Quercus minima, 61, 64, 71, 75, 83 Release treatments, 219, 220, 222, 188, 189, 193, 194, 197–200, Quercus myrtifolia, 60, 64, 71 276 230, 245, 246, 289, 305, 311, Quercus palustris, 63, 71 Removal cut, 118, 119, 235 312, 320, 336, 339–368, 396, Quercus prinus, 63, 71, 81 Reproduction method, 4, 5, 95, 115, 403, 407, 418, 419, 423–427 Quercus pumila, 61–65, 71, 73, 76, 82, 120, 124–127 donor population, 336, 342, 344, 83, 89 alternative strip, 115 347–349, 353, 357–359, 361, Quercus stellata, 16, 61, 62, 71, 77, clearcutting method, 115, 116 364 139 progressive strip, 118 expansion, 351, 352 seed-tree, 116 founder populations, 331, 332, Radicle, 102, 103 shelterwood, 116, 117 355 Recreation, 200, 395, 418–428 Reproductive biology, 95, 99, 231, metapopulation, 189, 197, Red Hills plantations, 248 320 344–346, 351, 361, 388, 389, Red-cockaded woodpecker, 5, 6, Reptiles, 158, 159, 167–169, 179, 396, 397, 416 176, 178, 185, 195–201, 217, 183, 189, 191–194, 197–199, scale, 346, 348, 360, 361 218, 231, 235, 237, 238, 245, 271, 337, 338, 358, 362–364 436 Index

Reptiles (cont.) Saw timber, 308 seed cut, 118, 119 characteristic reptiles resident in Scalping, 114, 274, 281, 283, 315 uniform, 117, 119, 120, 125 longleaf pine savannas, 168 Scansorial, 158 Shortleaf pine, 15, 16, 29, 70, 99, Resettlement Virgin longleaf pine, Scarlet oak, 70 120, 123, 218, 223–231, 238, 10, 24, 28, 193, 382 Scientiﬁc names, 68, 160 239, 254, 259, 286, 287, 290, Resilience, 6, 343 Sciurus niger, 179, 180, 338, 383, 339, 343, 380 Restoration, 3–7, 37, 38, 52, 58, 90, 391 Silty uplands, 56, 59, 62, 76–78 95, 97, 99, 124, 136, 139, 151, Scottish Whiskey stills, 46 Silviculture, 5, 95, 111, 115, 119, 153, 269–429 Scrub oaks, 16, 59, 75, 139, 174, 120, 124–127, 135, 139, 196, economics of, 404 179, 284, 285, 326, 327 197, 215–264, 381, 394 ground layer, 297–299, 306 Seed, 4, 22, 29, 32, 33, 95, 96–102, even-aged, 99 landscape approach, 6 108–113, 116–125, 141, uneven-aged, 5, 119–120, of overstory, 271–292 146–154, 179, 186, 219, 221, 217–239 of understory, 6, 153, 275, 284 222, 227, 228, 231–235, 238, Silvicultural system, 115, 125, partial, 300 252, 272, 273, 277–282, 218–221, 223 policy aspects of, 6 285–291, 298, 302, 304, Single tree selection, 201, 279, socioeconomics of, 6 311–321, 331, 332, 381, 405 281 strategies, 309 bed, 277, 280, 304 Site index, 255, 259–262 thresholds, 6 cleaning, 313, 314, 315 Site index curves, 260, 261 trajectories, 6, 7 collection, 311, 313, 314 Site preparation, 114, 115, 196, 197, Restoration ecology, 6, 378 dispersal, 95, 100, 108, 116, 186 219, 228, 233, 236, 251, 272, Reverse J, 223, 237, 239, 246, 278, germination, 315 273, 275, 279, 284–287, 279 handling, 313 302–304, 310, 311, 316, 380 Riparian strips, 384, 385 harvest, 313, 314, 315 3-in-1 plow, 273 Road networks, 395 production, 98, 101, 102, 108, bedding, 273 Rosin yards, 24 141, 146, 148, 149, 154, 232, bulldozing, 273 Rotation age, 219, 227, 405, 409, 410 298, 312 burning, 274 Ruderal species, 304, 305 storage, 313, 314, 315 chemical, 272, 273 Runner oak, 57, 71, 159 Seed cut, 117–119, 231, 234 disking, 273 Running oak, 71, 83 Seed tree, 101, 108, 153, 219, 234, drum chopping, 273 Rust, 99 235, 252, 277, 288–291 mechanical, 114, 115 Seedling, 111, 112, 135, 272, 280, scalping, 274 Sabal palmetto, 58, 65, 71, 87 290, 305, 310, 311, 315, 316, shear and pile, 273 Salamander, 5, 159–163, 166, 189, 318–320 sub soiling, 274 193–198, 338, 362, 363, 383, bare-root, 112–115 Site quality, 101, 154, 253, 259, 261, 387, 418, 419, 423–428 containerized, 113, 114 262 dwarf, 163, 195 development, 95, 102 Slash pine, 11, 18, 30, 35–37, 43–46, list of salamanders of longleaf pine growth, 104, 105, 107–109, 138, 70, 88, 96, 99, 163, 181, 191, savanna, 161 140, 233, 276, 279 196, 197, 262, 285, 287, 300, Mabee’s, 161, 163 mortality, 106, 107 301, 310, 339, 342, 362, 404, tiger, 161, 163, 189 plug, 114, 305, 319 405, 408, 411 Sand laurel oak, 71 Seedling exclusionary zone, 110 Snags, 98, 125, 174, 179, 181, 190, Sand live oak, 70 Seeps, 56, 59, 65, 84, 87, 89, 97 191 Sandhill oak, 71 Selection method, 120, 122–223, Snakes, 5, 158, 159, 167–174, Sandhills, 3, 15, 19, 53–68, 73–76, 227–233, 237–239 190–193, 198, 365, 387, 391, 79, 80, 86, 87, 97, 105, 107, group, 120 396 110, 157–175, 179, 181, 182, single tree, 120 characteristic snakes of longleaf 190, 198, 199, 298, 301, 304, Serenoa repens, 58, 60, 61, 64, 67, 71, pine savannas, 168 307–309, 313, 316, 317, 326, 73, 75, 81–84, 123, 307, 328, Society for Ecological Restoration, 5, 328, 343, 344, 357, 364, 365, 386 6 377, 379, 383, 386, 390–392, Shade-intolerant, 138, 221 Socioeconomic, 6, 336 418, 425 Shear and pile, 273 Soil orders, 56 Savanna, 6, 10, 11, 16, 18, 34, 56, Shelterwood, 95, 109, 116–120, 124, Soil productivity, 6, 273 79, 81, 84–90, 159–201, 316, 125, 218, 219, 231–238, Soils, 14, 16, 21, 25, 29, 52–59, 73, 330–332 277–279, 288–290, 405 76–90, 98, 103, 105, 107, 110, Savannah River Site, 136, 142, 330, irregular, 119, 120, 125 112, 142, 153, 157, 165, 167, 344 modiﬁed, 278, 279 169, 171, 173, 175, 181, 186, Saw palmetto, 58, 71, 122, 124, 166, preparatory cut, 117–119, 289 194, 195, 199, 252, 261, 274, 174, 307, 328, 386, 387 removal cut, 118, 119, 235 283, 290, 304, 308, 330–332, Index 437

343, 377, 379, 381, 383, 384, Sulfometuron methyl, 275 characteristic turtles of longleaf 396 Sweet gum, 163 pine savannas, 168 Solenopsis geminata, 199 Sweetbay, 70 Sonderegger invicta, 199 Switchgrass, 70 U.S. Fish and Wildlife Service, 178, Sonderegger pine,99 199, 200, 335, 339, 340, 346, Sorghastrum nutans, 62, 71, 75 Tar Heelers, 24 348, 349, 350, 353, 355, 357, Sorghastrum secundum, 60, 71–73, 75 Taxodium ascendans, 161 359–366, 369, 406, 426 South Carolina, 4, 13, 15, 21, 23, 25, Temporal factors, 354 Ultisol, 76 27, 28, 30, 33–37, 53, 54, 57, 58, Temporary ponds, 164–166, 175, Understory, 4, 6, 10, 13, 16, 37, 59, 67, 73, 74, 76, 81, 85–87, 108, 189, 196 73, 74, 75, 76, 79, 82–84, 86, 165, 189, 190, 193, 199, 200, Texas, 3, 4, 9, 13, 17, 18, 21, 22, 25, 98, 105, 108, 109, 114, 115, 297, 298, 307, 308, 313, 330, 28, 36, 51, 57, 68, 97, 99, 157, 122, 125, 127, 135–145, 331, 335, 343, 344, 359, 360, 158, 163, 172, 198, 200, 257, 151–154, 178, 197, 220, 221, 362, 365, 392, 407, 410, 426 259, 328, 339, 344, 359, 366, 229, 233, 237, 243–248, 262, Southern bald eagle, 363, 366 377 271, 275, 278, 283–292, 307, Southern magnolia, 70 The Nature Conservancy, 200, 310, 309, 326–328, 331, 337, Southern red oak, 16, 70, 380, 383 320, 338, 391, 413, 415, 419, 404–407, 425, 426 Sowing, 112, 113, 152, 272, 316–319 421, 427 common names, 68–72 Spatial ecology, 377, 379, 381, 383, Thinning, 38, 117, 119, 123, 135, scientific names, 68–72 385, 387–397 139, 142–147, 154, 201, Understory diversity, 327 Species richness, 5, 8, 56, 57, 78, 86, 219–221, 229, 234, 248, 252, Understory reinitiation stage, 138, 89, 159, 175, 176, 196, 200, 259, 260, 264, 271, 281, 287, 139 274, 302, 303, 305, 307–309, 288, 304, 305, 310, 331, 332, Uneven-aged silviculture, 5, 119, 317, 425 386, 405 120, 217–249 Species selection, 311 Threatened, 3, 5, 6, 171, 198–201, USDA Forest Service, 44–47, 95, Spodosol, 81 271, 298, 309, 335, 336, 338, 135, 217, 222, 223, 232–239, St. Augustine, 17 342, 356, 362, 364, 404, 408, 251, 254, 297, 326, 330, 382 Stands, 3, 4, 5, 11, 13, 15, 16, 22, 26, 423, 425 28, 34, 37, 38, 44, 52, 58, 63, Thresholds, 6, 7, 8, 260, Vegetation classification, 52, 58, 66, 97–99, 101, 104, 107, 108, 110, 264 90, 396 111, 115–123, 135, 138, 139, Titi, 69, 160, 418 national, 52, 58, 66, 90, 396 145–153, 163, 178, 179, 191, Toothache grass, 69 of longleaf pine communities, 195–198, 217– 239, 246–248, Trajectory, 5–7 58 251–263, 272, 275–280, Translocation, 6, 335–337, 339–369, Vegetation management, 135, 154 286–292, 303–308, 328, 330, 417, 418 Venus flytrap, 15, 69, 81 331, 337, 339, 343, 381–383, concerns, 357 Vertebrate diversity, 5 386–389, 404, 405, 409 demographic factors, 352 Virginia, 3, 9, 12, 13, 16, 18–21, stand density, 46, 101, 108, 110, factors affecting success, 348 25–37, 51–53, 67, 70, 72, 81, 118, 158, 222, 225, 243, 253, habitat conditions, 349 97, 157, 168, 180, 200, 339, 260, 262, 263, 281, 426 logistic factors, 353 343, 344, 359, 369, 377 Stand regulation, 120–124, 235 methodology, 356 Virginia pine, 70, 343 Stem exclusion stage, 138 need for, 339 Volume, 3–5, 18, 51, 52, 96, 97, 107, Stoddard–Neel approach, 5, 237, of Red–Cockaded woodpecker, 113, 116, 119–123, 135, 158, 239, 244 339, 343, 344, 353 195, 217–225, 229–231, 235, Stoddard–Neel System, 246 population scale, 360 236, 242–245, 248, 251–259, Stratification, 112, 146, 315 regional scale, 359 262, 264, 271, 283, 297, 298, Strip scalping, 114 strategies, 359 337, 379, 380, 403, 404 Striped newt, 161–163, 166, 167, success, 346 Volume-guiding diameter limit 189, 193, 198, 387 temporal factors, 354 (V-GDL), 120–124 Strobili, 99, 100 type, 354 Stumpholes, 165, 191, 193 Triclopyr, 142, 152, 275, 307, 309 Wade Tract, 78, 246 Stumpwood, 48 Turkey oak, 16, 71, 159, 163, 166, Warm-season grass, 158 Sub soiling, 274 169, 171–173, 175, 181, 284, Water mining, 393 Subxeric sandy uplands, 56, 57, 74, 290, 301, 307, 364, 365 Water quality, 273, 281, 418, 424, 75 Turpentine, 3, 4, 9, 18, 20–26, 29, 427, 428 Succession, 6, 34–36, 116, 119, 140, 34, 35, 43, 46–48, 193, 381 Wetland, 157, 173, 174, 189, 195, 176, 180, 181, 218, 331, 332 Turtles, 5, 159, 168, 169, 171, 172, 196, 331, 377, 383–387, 390, community, 6, 116 174, 175, 183, 184, 186, 189, 419 old field, 11 190 White fringed beetle, 283 438 Index

White-tailed deer, 179, 183 Wiregrass, 4, 15, 34, 38, 57, 59, 68, Xeric sand barrens, 56–60, 67 Wildland–urban interface, 395, 425, 73, 74, 76, 78, 80, 98, 135, 138, Xeric sites, 56, 67, 73, 111, 285, 298, 426 141, 165, 166, 173, 174, 195, 302, 316, 405 Wildlife Habitat Incentives Program 200, 217, 307–309, 313–320, (WHIP), 406, 410 330–332, 380, 381, 396 Yaupon, 69