Range Collapse, Genetic Differentiation

Clemson University TigerPrints

All Dissertations Dissertations

5-2014 RANGE COLLAPSE, GENETIC DIFFERENTIATION, AND CLIMATE CHANGE: AN ECOLOGICAL HISTORY OF THE DIANA FRITILLARY, SPEYERIA DIANA AND PROJECTIONS FOR ITS FUTURE Carrie Wells Clemson University, [email protected]

Follow this and additional works at: https://tigerprints.clemson.edu/all_dissertations Part of the Biology Commons

Recommended Citation Wells, Carrie, "RANGE COLLAPSE, GENETIC DIFFERENTIATION, AND CLIMATE CHANGE: AN ECOLOGICAL HISTORY OF THE DIANA FRITILLARY, SPEYERIA DIANA AND PROJECTIONS FOR ITS FUTURE" (2014). All Dissertations. 1382. https://tigerprints.clemson.edu/all_dissertations/1382

This Dissertation is brought to you for free and open access by the Dissertations at TigerPrints. It has been accepted for inclusion in All Dissertations by an authorized administrator of TigerPrints. For more information, please contact [email protected].

RANGE COLLAPSE, GENETIC DIFFERENTIATION, AND CLIMATE CHANGE: AN ECOLOGICAL HISTORY OF THE DIANA FRITILLARY, SPEYERIA DIANA AND PROJECTIONS FOR ITS FUTURE

A Dissertation Presented to the Graduate School of Clemson University

In Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy Biological Sciences

by Carrie N. Wells May 2014

Accepted by:

Dr. David Tonkyn, Committee Chair Dr. Peter Adler Dr. Peter Marko Dr. Saara DeWalt

ABSTRACT

The geographic ranges of most plant and animal species are tied closely to climatic factors, including temperature, precipitation, and soil moisture. For this reason, recent changes in the global climate due to human activities are predicted to have profound effects on natural populations, communities and ecosystems over a relatively short period of time. Combined effects from global warming and other anthropogenic activities such as land-use changes, pollution, and habitat loss/fragmentation, are altering species’ distributions faster than they can be documented. Recent climate change has also been shown to alter species’ breeding behaviors and alter the synchrony and timing of species’ phenologies.

The Diana Fritillary (Speyeria diana Cramer 1777) is a North American butterfly that appears to have declined over the past century. This butterfly species once ranged from coastal Virginia westward to Missouri and Arkansas, southward to the northern tips of Georgia and Alabama, and northward through the Ohio River Valley. It has since disappeared from large portions of its once semi-continuous range, and persists now primarily in two geographically distinct regions separated by an 850 km disjunction. The

North Carolina and Arkansas Heritage Programs currently list S. diana as an imperiled species of special concern (rank S2/S3) due to its rapid decline over the past two decades; it is also included on the Xerces Society Red List of Pollinator Insects. The conservation network, NatureServe, assigns S. diana a Global Status of G3/G4, which describes the species as very rare or local throughout its range, found locally in a restricted range (21 to

100 occurrences), and threatened throughout its range. Because of its rapid disappearance across portions of its former distribution, S. diana may soon become a candidate for

listing under the Endangered Species Act of the United States (Federal Register 1991,

Vol.56, no. 225, pp. 58, 831). Currently, S. diana is not protected through any special conservation status, despite its apparent decline.

The Diana fritillary is univoltine, producing one generation per year. Adult males emerge and take flight in late May, typically several weeks before females. Males patrol along the edge of forest habitat, and have an active and mobile lifestyle. While males begin to die off in late July, females persist somewhat cryptically into early October.

Females are believed to be longer lived than the males, and are often found resting quietly in the cover of forest for much of the day, nectaring or ovipositing on the forest floor. In general, S. diana inhabits moist cove forests and deep woodland areas near streams. Adult Diana fritillaries are often found in open areas feeding on tall, high-quality nectar sources such as milkweeds, butterfly bushes or large fall composites. Violets

(Viola spp., Violaceae) are the only larval host plants used by Speyeria. Each female

Diana fritillary can lay thousands of eggs singly on ground litter during the month of

September in the vicinity of violets. The hatched larvae immediately burrow deep into the leaf litter of the forest floor where they overwinter until the following spring.

This dissertation is a comprehensive study of this butterfly over time, which documents the species’ changing distribution over time, and explores the causes for its decline. I have documented the pattern and timing of the range collapse, having compiled over 2,300 records of occurrence from the literature, my own field work, and public and private collections in the US and Europe. I show that the species has disappeared from lowland sites across its range, and now persists only at higher elevations in the southern

Appalachian Mountains in the east and the Ozark and Ouachita Mountains in Arkansas,

iii

with no populations between. This first dissertation chapter, entitled Range Collapse in the Diana fritillary, was published in Insect Conservation and Diversity in 2013. Second,

I have documented patterns in genetic variation using mitochondrial DNA of the cytochrome oxidase II (COII) gene from historical (museum) and field-collected specimens. There are clear differences between eastern and western populations, with the earliest split between east and west occurring around 20K years ago, long before the recent range collapse. An historical comparison suggests that lowland populations have disappeared, taking a unique haplotype with them. This second chapter is presently under review in the journal Conservation Genetics. Finally, I use bioclimate envelope modeling to predict the future distribution of S. diana under several climate change scenarios. I also explore alternative explanations for the range collapse, including changes in land use or fire management, recovery of white-tailed deer, and aerial spraying for gypsy moth. The

Diana fritillary appears to be threatened by severe reduction and fragmentation of suitable habitat, especially in the southern Appalachian Mountains, by the year 2050.

Study species: The Diana fritillary (Speyeria diana Cramer 1777) male on top; female is on bottom.

DEDICATION

This dissertation is dedicated to my parents, Phil and Connie Wells, and my sister,

Carlee Jones. They have supported me through thick and thin, and I would not be where I am today without them. I am grateful for everything, and love you all so very much.

ACKNOWLEDGMENTS

I am especially grateful to the members of my PhD Committee: David Tonkyn,

Peter Adler, Peter Marko, and Saara DeWalt. I also would like to thank Barbara Speziale and the Clemson University Creative Inquiry Program for funding a majority of this research, and allowing me the opportunity to engage so many undergraduate students in butterfly ecology research projects. H. Nance, S. Emme, S. Marchant, and the rest of the

Marko Lab provided valuable technical support for DNA sequencing and project design.

Research funds were also awarded by the Sarah Bradley Tyson Memorial Fellowship, the

Blue Ridge Parkway Foundation, and a collection grant from the American Museum of

Natural History. I would also like to thank all of the curators who contributed time and expertise to this project, including Bob Robbins (Smithsonian National Museum of

Natural History), Blanca Huertas (British Natural History Museum), David Grimaldi

(American Museum of Natural History), James Boone (Field Museum), John Rawlins

(Carnegie Museum of Natural History), Tim Tomon (Carnegie Museum of Natural

History), George Austin (University of Florida Museum of Natural History & McGuire

Centre for Lepidoptera and Biodiversity), Behnaz van Bekkum-Ansari (Netherlands

Centre for Biodiversity Naturalis Collection), Jacques Pierre (Paris Museum national d’Histoire naturelle) and Willem Hogenes (Zoölogisch Museum Amsterdam). We thank

Irving Finkelstein, Cindy Wentworth, Jason Love, Jessica Salo, Lori Spencer, Don

Simons, Eric Smith, Bill Garth, Tom Payne, Bill Baltosser, Brent Kelley, Chelsea

Woodworth, Marielle Abalo, Connie Wells, and the dozens of Clemson University

Biological Sciences students who assisted with field work.

vii

TABLE OF CONTENTS

Page

TITLE PAGE ...... i

ABSTRACT ...... ii

DEDICATION ...... vi

ACKNOWLEDGMENTS ...... vii

LIST OF TABLES ...... x

LIST OF FIGURES ...... xii

CHAPTER

I. RANGE COLLAPSE IN THE DIANA FRITILLARY

Abstract ...... 15 Introduction ...... 16 Methods ...... 17 Results ...... 21 Discussion ...... 25 Literature Cited ...... 32

II. THE PHYLOGEOGRAPHIC HISTORY OF THE THREATENED DIANA FRITILLARY, SPEYERIA DIANA (LEPIDOPTERA: NYMPHALIDAE), WITH IMPLICATIONS FOR CONSERVATION

Abstract ...... 56 Introduction ...... 56 Methods ...... 58 Results ...... 62 Discussion ...... 66 Literature Cited ...... 70

III. PREDICTED EFFECTS OF CLIMATE CHANGE ON THE FUTURE GEOGRAPHIC DISTRIBUTION OF THE DIANA FRITILLARY IN SOUTHEASTERN US

viii

Table of Contents (continued) Page

Abstract ...... 91 Introduction ...... 91 Methods ...... 94 Results ...... 98 Discussion ...... 99 Literature Cited ...... 101

IV. CONCLUSIONS: WHAT DOES THE FUTURE HOLD FOR SPEYERIA DIANA? ...... 111

Literature Cited ...... 114

APPENDIX

Supplemental bibliography of all distributional literature sources, in order of appearance in Table 2.2 ...... 115

LIST OF TABLES

Table Page

1.1 Summary of Speyeria diana distributional data sources ...... 42

1.2 Summary of literature referencing the distribution of Speyeria diana

1.3 Field-sampled Speyeria diana (2006-2009). Records are provided to the level of county ...... 46

2.1 Mitochondrial DNA population diversity indices for S. diana populations from fresh and museum specimens. Haplotypes in bold are private to their respective population ...... 74

2.2 Mitochondrial DNA diversity indices for S. diana populations, including nucleotide (π) and haplotype (h) diversities. Neutrality statistics Tajima’s D, and Fu’s FS are bold when significant at α = 05. Extirpated populations are marked with † and western populations are highlighted in grey ...... 75

2.3 Population pairwise ΦST (Weir & Cockerman 1984) values among S. diana populations are above diagonal; those with significant differentiation after sequential Bonferroni correction are in bold print. Extirpated populations are marked with a †. East-west comparisons are highlighted in grey. Associated P-values are shown below the diagonal ...... 76

2.4 Speyeria diana population structure based on 549 bp mtDNA COII differentiation in spatial analysis of molecular variance (SAMOVA). Groupings K=2-6 above the line include three extirpated populations (†) in the analysis, while those below the line do not. Western populations are highlighted in grey ...... 77

2.5 Estimates of current (ΘE = Eastern and ΘW = Western) and ancestral (ΘA) theta, migration (m1 and m2), and divergence time (t) in years, from IMA2.0. The 95% highest posterior density (HPD) intervals are given for the pooled analysis only, where posteriors rose from and dropped down to zero probability. Extirpated populations are indicated with a †, and significance at α = 0.05 is indicated with a * ...... 78

List of Tables (continued)

Table Page

3.1 The 19 bioclimate variables from the WorldClim dataset (Hijmans et al., 2005) collapsed into groups of highly correlated variables (Pearson’s correlation coefficient, r ≥ +/- 0.70), and their corresponding contribution to the Maxent model. The ten variables kept in the final model are bold and highlighted in grey ...... 108

LIST OF FIGURES

Figure Page

1.1 All distributional records for Speyeria diana specimens documented from 1777 to 1960. This species once ranged widely across the southeastern United States, from the coastal plains of North Carolina and Virginia across the Appalachian Mountains and the Ohio River Valley, and west to Arkansas and Missouri ...... 50

1.2 All distributional records for Speyeria diana specimens documented from 1961 to 2010. This species has experienced a range collapse resulting in two geographically separated population groups: the southern Appalachian Mountains in the east, and the Ozark and Ouachita Mountains in the west. Populations have disappeared from the intervening lowlands and the eastern coastal plain ...... 51

1.3 (a) Latitude was negatively correlated with the year of S. diana capture, reflecting the recent expansion of S. diana into its southern range boundary (Latitude = 53-0.008*year; R2 = 0.04, n = 1680, P < 0.0001); (b) Longitude was also significantly and negatively correlated with year, reflecting westward expansion (Longitude = -35-0.02*year; R2 = 0.03, n = 1680, P < 0.0001) ...... 52

1.4 Speyeria diana are shifting to significantly higher elevations in recent times, and this shift varies slightly with latitude (Elevation (m) = 212 - 74.3*latitude + 0.4*year + 0.04*year*latitude; R2= 0.33, n = 1680, all effects significant at P < 0.0001). Butterflies were at slightly higher elevations and climbing slightly faster in the north than in the south ...... 53

1.5 Speyeria diana females are flying progressively earlier in recent times, and this varies slightly with latitudes (Flight day of year = 984 + 2.34*latitude -0.57*Year +0.004*Year*latitude; R2= 0.19, n = 840, all effects significant at P < 0.0001) ...... 54

1.6 Speyeria diana males have a statistically significant relationship between flight date, year and latitude (Flight day of year = 123 + 1.32*latitude - 0.34*year + 0.01*year*latitude; R2= 0.13, n = 840, all effects significant at P < 0.0001) ...... 55

xii

List of Figures (continued)

Figure Page

2.1 Distributional data for Speyeria diana specimens documented in Wells and Tonkyn (2013) showing the two geographically separated population groups: the southern Appalachian Mountains in the east, and the Ozark and Ouachita Mountains in the west. Open circles represent specimens collected from 1777 to 1960 (N= 881); Black circles represent S. diana specimens collected from 1961 to 2010 (N= 2,517) (See Wells and Tonkyn 2013 for a complete description of these data) ...... 84

2.2 Collection sites for Speyeria diana. Eastern samples were taken from Georgia (GA), South Carolina (SC), Tennessee-1 (TN-1), Tennessee-2 (TN-2), Tennessee-3 (TN-3), North Carolina (NC), and Virginia-1 (VA-1); Western samples were taken from Arkansas-1 (AR-1), Arkansas-2 (AR-2), Arkansas-3 (AR-3), and Oklahoma (OK); Museum samples were sampled to represent extirpated populations (indicated with †) from Virginia-2 (VA-2), Indiana (IN), and Ohio (OH). See Table 1 for sources of all museum samples. The size of each pie graph represents the sample size of each population. The size of each slice is proportional to the number of individuals represented by each haplotype ...... 85

2.3 Unrooted SplitsTree4 (V 4.13.1) (Huson & Bryant 2006) haplotype network for 549 bp of COII. The circle size in each network is proportional to the number of individuals of each haplotype. sample size, with the smallest circle representing one haplotype copy. See Figure 2.2 for haplotype legend ...... 86

2.4 Plot of ΦST values (based on COII sequence data) versus geographic distance among all collection sites indicating evidence for isolation by distance ...... 87

2.5 Posterior probability distributions for east-west divergence times (t) and for rates of migration (m) for pooled samples of S. diana ...... 88

2.6 Estimates of ΘE (eastern), ΘW (western), and ancestral (ΘA) theta from IMA2.0 for pooled samples of S. diana ...... 89

2.7 Geographic distribution of divergence time (t) in years for eastern, western and east vs. west comparisons ...... 90

xiii

List of Figures (continued)

Figure Page

3.1 The entire geographic distribution of the butterfly, Speyeria diana, with indices of habitat suitability as predicted by maximum entropy modelling (Maxent) under current climatic conditions ...... 109

a. (a) Habitat suitability indices for the projected future distribution of Speyeria diana under the CCMA and MIROC RCP 4.5 climate change scenarios; (b) Habitat suitability indices for the projected future distribution of Speyeria diana under the CCMA and MIROC RCP 8.5 climate change scenarios ...... 110

xiv

CHAPTER ONE

RANGE COLLAPSE IN THE DIANA FRITILLARY, SPEYERIA DIANA

Abstract

Global warming can affect the distributions, phenology and, ultimately, conservation status of species worldwide, but most published studies on its biological effects have focused on higher latitude species. We extended this work to the Diana fritillary, a butterfly which once ranged throughout the southeastern United States but now is severely restricted in range. We searched for all scientific records of this species, from publications, catalogued and uncatalogued specimens in public and private collections in the United States and Europe, online databases, contemporary field surveys by scientists and amateurs, and our own field surveys. We analyzed these records for shifts in latitude, longitude, elevation and phenology. We found that the Diana fritillary has disappeared entirely from the Atlantic coastal plain, where it was first described, and from interior lowland sites. It now persists in two disjunct parts of its former range, the southern Appalachian Mountains and the Interior Highlands of Oklahoma and Arkansas, and is shifting to higher elevations at about 18 m per decade. In addition, females are being collected 4.3 days earlier per decade though males, which emerge first, have not shifted their phenology. All these patterns are weakly dependent on latitude. These shifts in distribution and phenology are consistent with the predicted effects of global warming, but we review other large scale changes to the region which also might contribute singly or jointly to these patterns. We also comment on the implications for the conservation of this species.

Introduction

Butterflies of the North American fritillary genus Speyeria Scudder are of growing concern to conservation biologists, as a number have become threatened with extinction in the past 200 years (Hammond & McCorkle 1983; Cech & Tudor 2005). The

Diana fritillary, Speyeria diana (Cramer) is an eastern North American fritillary that has become increasingly rare across portions of its distribution (Rudolph et al. 2006;

Campbell et al. 2007; Ross, 2008). Early accounts describe S. diana as once ranging from coastal Virginia westward to Missouri and Arkansas, southward to the northern tips of

Georgia and Alabama, and northward throughout the Ohio River Valley from western

Pennsylvania to Illinois (Edwards 1864; Strecker 1878; French 1886; Hine 1887a, b;

Scudder 1889; Skinner 1889; Clark 1937; Clark & Williams 1937; Clark & Clark 1951).

This species has since disappeared from much of its range, including coastal Jamestown,

Virginia, where the type specimen was described by the Dutch explorer, Pieter Cramer

(Cramer & Stoll 1777). Populations across the northern portion of the range declined rapidly in the early 1800s and are thought to have disappeared by the early 1900s (Shuey et al. 1987; Opler & Malikul 1992).

The result of these declines appears to be a present-day range collapse for the species, resulting in a geographic disjunction between the two major remaining population centers in the Interior Highlands of Arkansas and Oklahoma and the southern

Appalachian Mountains (Carlton & Nobles 1996; Moran & Baldridge 2002; Baltosser

2006). It is important to understand the causes for this range collapse, in case they threaten the future of the species.

To study the changing distribution of S. diana over time, we conducted a

comprehensive search for all records of the butterfly since its description in 1777, representing essentially the entire accumulated scientific record of the species. These include published records, catalogued and uncatalogued specimens in public and private collections, and contemporary field surveys by scientists and amateurs. With these data, we addressed the following questions: (1) how has the distribution of the Diana fritillary changed over the past 200 years, (2) which populations of S. diana have disappeared or are in decline, and which are stable or expanding, (3) has the phenology of this species also changed over the past 200 years, and (4) do these changes point to global warming or other specific factors as the underlying causes?

Methods

Study species

Speyeria diana Cramer is the largest member of the greater fritillaries (Speyeria

Scudder, 1872 (Nymphalidae: Heliconiinae: Argynnini)), with a wingspan ranging from approximately 88 to 112 mm (Opler & Malikul 1992; Dunford 2009). The sexes are dimorphic in size and color. Males exhibit typical orange fritillary coloration and have a deep brown colored center and bright orange outer-wing margin. Females are typically twice the size of males and have a spectacular iridescent blue to blue-green coloration.

The female color pattern is thought to mimic that of the aposematic pipevine swallowtail

(Battus philenor) (Scudder 1889; Brower 1958; Poulton 1909; Brower & Brower 1962); however, this mimetic relationship remains speculative (Hovanitz 1963). The female

Diana is the only eastern fritillary to deviate from the typical orange and brown fritillary coloration.

Each female Diana will lay thousands of eggs singly on ground litter in the vicinity of violets (Viola spp., Violaceae) in late summer before expiring (Allen 1997;

Ross & Henk 2004; Wells et al. 2011). First-instar larvae hatch after several weeks and immediately burrow into the forest floor litter where they overwinter in diapause. Larvae emerge from the soil in early spring to feed on fresh Viola foliage. Violets are the only known host plants used by members of the genus Speyeria (Holland 1931; Wells et al.

2011). Caterpillars pupate in early spring, with males taking flight in late May, several weeks in advance of the females (Glassberg 2011). As adults, S. diana males and females can be found nectaring on milkweed (Asclepias spp.) (Opler & Krizek 1984), buttonbush

(Cephalanthus occidentalis), coneflower (Echinacea spp.), compassplant (Silphium laciniatum) (Moran & Baldridge, 2002), Joe-pye weed (Eupatorium maculatum), iron- weed (Vernonia glauca), and mint (Pycnanthemum incanum) (Scholtens 2004). Males begin to die off in late July, while females can persist well into October (Adams &

Finkelstein 2006).

Data collection

We collected distributional and phenological data for S. diana from all sources that we could identify (Table 1) and under all taxonomic designations ever assigned to this species. These include the names Papillio diana, Nymphalis diana, Phaleratus diana,

(Cramer & Stoll 1777), Argynnis diana (Holland 1889, 1931; Simonson et al. 2006),

Speyeria diana (dos Passos & Grey 1945, 1947), as well as the common names Diana fritillary, Great Smokies fritillary, and Ozark Diana fritillary (Dunford 2009). We additionally searched the subgenus name Semnopsyche Scudder (1875), as S. diana has historically been classified into this taxonomic group (dos Passos & Grey 1945; Klots

1951; Miller & Brown 1981; Grey 1989).

We began by contacting all natural history collections and insect repositories that we could identify as holding S. diana. The lead author traveled to 15 major museums (9 in North America and 6 in Europe) to inspect their S. diana holdings and catalog the label data. Eighteen additional museums provided their S. diana collection data via mail, email, and telephone correspondence (Table 1). We contacted all appropriate federal, state, and non-profit agencies in the southeastern US and requested all available survey records for

S. diana. The largest contribution to our dataset came from online repositories where butterfly records have been collected and stored for years or even decades by professional and amateur lepidopterists. These records came from the Global Biodiversity Information

Facility (GBIF), NatureServe, the North American Butterfly Association (NABA),

Butterflies and Moths of North America (BAMONA), the Carolina Butterfly Society

(CBS), The Lepidopterist’s Society, and the All Taxa Biodiversity Inventory (ATBI) conducted within Great Smoky Mountains National Park. We also received a large number of records from independent naturalists regarding the location, ecology, and habits of S. diana. We further obtained distribution records from several online entomology listserves including Carolina Leps, Insectnet.com, Oklahoma Leps, and the

Washington Area Butterfly Club (Table 1.1).

We next searched for all existing records in the scientific and popular literature mentioning S. diana (Table 2.1). Any data relating to the ecology or distribution of S. diana were recorded (e.g., collecting date and year, collector, locality, the butterfly’s sex). We manually searched all non-digitized volumes of Zoological Records, 1864-

1978, which are held in special collections at Cooper Library, Clemson University, to

locate the earliest indexed literature sources from the nineteenth century. We additionally used internet-based search engines, Web of Knowledge, Biosis, and Google Scholar, using the search terms above to locate published media referencing S. diana.

During summers 2006-2009, we visited a number of sites where S. diana has been reported historically (Table 3). Sites were surveyed from May through September, between 1000 and 1700 hrs EST, on days with temperatures ranging from 19° to 36° C.

We spent a total of 2,500 person-hours searching for S. diana, using a combination of slow driving along forest edges and hiking into interior habitats. Adults of S. diana were captured with a handheld net and given a unique number on the right forewing with a

Sharpie® marker, to alert us in case of recapture.

Data analysis

We combined all the distributional data for the Diana fritillary into a comprehensive relational database, which will be permanently housed with BAMONA (Table 1). We used ArcMap 10.0 (ESRI 2011) to create distribution maps for S. diana. To map distribution data, we geo-referenced each specimen with the latitude and longitude of its capture site. Where an exact location was provided, we also obtained the elevation using

ArcGIS; otherwise, we localized records to county and obtained the latitude, longitude and elevation of the county centroid, using 1:250 000 U.S. Geological Survey (USGS)

Digital Elevation Model (DEM) files obtained from the USGS internet site (U.S.

Geological Survey 1987, 1998). When multiple butterfly records were found from the same date and location of capture, we only included a single occurrence in our analyses to avoid introducing bias.

We used linear regression analyses with JMP10 (JMP 2012) to test for correlations between the specimens’ year of capture and the corresponding latitude and longitude, to look for shifts in geographical range. We then carried out a multiple linear regression of elevation on the year and latitude of capture and their possible interaction, to see whether the butterflies are shifting in elevation, and whether this response varies with latitude. To avoid sampling bias from the greater number of modern records, we sorted all observations by year and scored the number of unique records from each decade. We pooled data in the earliest decades (1777-1899), as there were few records containing location data and dates prior to 1900. We then subsampled the data by decade from 1901 to 2010, randomly selecting forty male and forty female records per decade for analysis. All these regressions were conducted on the sexes separately and together.

In addition, we used a t-test to compare the average elevation of S. diana records from those counties only having S. diana records prior to 1950, to those records from counties only having S. diana reported after 1951.

As an index of S. diana phenology, we converted all collection dates for adult butterflies into the ordinal day of the year (out of 365 days), accounting for leap years.

We then assessed changes in flight dates over time by conducting a multiple linear regression of capture dates with year and latitude of capture, and their possible interaction. Because S. diana males can emerge weeks prior to females (Glassberg

2011), this analysis was also conducted for males and females separately.

Results

Historical data collection

We extracted all distributional and phenological data mentioned from the sources summarized in Table 2. The full bibliography associated with Table 2 is included as

Supplemental information (Appendix S1). We obtained data from 1,323 pinned S. diana specimens from 38 natural history museum collections (Table 2). The oldest specimens were those housed in European museums, as the earliest specimens were brought back to

Europe from North America by explorers during the 19th century. The British Museum of

Natural History held 31 of the oldest (1777-1989) specimens from 17 US counties, including the type specimen from coastal Virginia. The McGuire Center at the University of Florida provided 409 S. diana specimens (1900-2007) from 43 US counties. The next largest collection of S. diana came from the Carnegie Museum of Natural History, with

142 specimens (1889-2000) across 26 counties. The Smithsonian National Museum of

Natural History, the American Museum, and Field Museum contributed data for 129,

104, and 98 museum specimens, respectively (Table 1.1).

Four hundred thirty five records (1938-2012) from 39 US counties were provided by the Butterfly and Moth Information Network and the participants who contribute to its

Butterflies and Moths of North America project (Table 1.1). Our literature survey yielded

153 additional records of S. diana (1818-2011) across 54 US counties (Table 1.2).

Field surveys

We documented S. diana populations in 25 counties from 7 states (Table 3). We collected 419 S. diana butterflies in our field surveys, with 288 males (68.7%) and 131 females (31.3%) (male:female ratio of 2.2:1). In Arkansas and Oklahoma, we sampled individuals from the Ozark Plateau, the Arkansas River Valley, and the Ouachita

Mountain ecoregions. We found the largest number of S. diana (N = 37) during our surveys of Mount Magazine State Park (840 m), the highest point in Arkansas located in

Logan County. In the Appalachian Mountains, we sampled S. diana populations from the

Blue Ridge Mountains, Great Smoky Mountains, and the Southern Appalachian

Mountains of Tennessee and Virginia. We also sampled a population in the Mountain

Bridge Wilderness Area, which borders the Blue Ridge Escarpment in the foothills of

South Carolina. The Great Smoky Mountains in eastern Tennessee harbored populations from which we sampled the greatest number of butterflies (N = 75). This is likely due to increased sampling intensity from several volunteers assisting with our surveys both inside and outside of Great Smoky Mountains National Park 2007-2009. All of our survey records were included in the following analyses.

Range collapse

The Diana fritillary was once more widely distributed across the southeastern US than at present (Fig. 1.1). The species has not been collected from its type locality on the

Virginia coast, or from the piedmont of Virginia, since the 1950s. Speyeria diana was extirpated from the northernmost portion of its range, across the Ohio River Valley, by the early 1900s. A record from Vermilion, Illinois, in 1960 was the last specimen reported near this region (Irwin & Downy, 1971). No S. diana were reported from Ohio,

Illinois, or Indiana during comprehensive NABA field surveys conducted 1999-2010.

Populations in the Blue Ridge, Great Smoky, and Appalachian Mountains are expanding, while lowland areas have steadily declined. Since the 1950s, this species has existed in geographically disjunct eastern and western population centers (Fig. 1.2). Populations in the Ozark and Ouachita Mountains of Arkansas have expanded rapidly since the 1980s.

Breeding populations have been confirmed in 30 Arkansas counties (L. Spencer, pers. comm. 2012). Oklahoma populations are expanding, with 19 counties adding S. diana records since the 1990s. No individuals have been reported from Missouri since 1994.

With sexes combined, we found a weak but statistically significant shift southward of 0.08o latitude, or 8.9 km, per decade (latitude (oN) = 52.7 – 0.008*year; R2

= 0.04, n = 1680, P<0.0001) (Fig. 1.3a). We also found a weak but statistically significant shift westward of about 0.2o longitude per decade (longitude (oW) = -34.9 –

0.02* year; R2 = 0.03, n = 1680, P<0.0001) (Fig 1.3b). This westward shift equals a shift rate of loss in eastward extent of about 18.3 km per decade, as measured at 35o north latitude, which is on the southern border of Tennessee and near the southern limit of S. diana’s current range. Similar results were obtained when males and females were analyzed separately. These changes probably reflect the recent expansion of S. diana into the mountains of Oklahoma and Arkansas and not a shift in the limits to the range.

Indeed, some of the southern-, northern- and eastern-most populations have disappeared in recent years.

With sexes combined, we found that S. diana were shifting to significantly higher elevations in recent times, and this shift varied slightly with latitude (Elevation (m) = 212

- 74.3*latitude + 0.4*year + 0.04*year*latitude; R2= 0.33, n = 1680, all effects significant at P < 0.0001) (Fig. 1.4). These results reveal that the butterflies were at slightly higher elevations and climbing slightly faster in the north than in the south. For example, at 35o north latitude, their mean elevation was 1,212 m in the year 2000 and rising by 18 m per decade, while at 40o N, the corresponding values were 1,240 m and 20 m per decade.

Again, we obtained similar results when the sexes were treated separately.

Mean elevation of counties reporting S. diana records only prior to 1950 was significantly lower (457 + 65 m, n=1365) than that of counties only reporting S. diana after 1951 (765 + 42 m, n=1995) (two sample t-test, P< 0.001).

Our phenological analysis indicated that S. diana females are flying progressively earlier in recent times, and this varies slightly with latitude (Flight day of year = 984 +

2.34*latitude -0.57*Year +0.004*Year*latitude; R2= 0.19, n = 840, all effects significant at P < 0.0001) (Fig. 5). Therefore, the females are emerging 4.3 days earlier per decade at

35o latitude, and 4.1 days earlier each decade at 40o latitude. With males, there is also a statistically significant relationship between flight date, year, and latitude (Flight day of year = 123 + 1.32*latitude - 0.34*year + 0.01*year*latitude; R2= 0.13, n = 840, all effects significant at P < 0.0001) (Fig. 6). However, the magnitude and signs of this relation are such that there is essentially no change in male phenology throughout the range.

Discussion

Climate change

In North America and Europe, at least 39 species of butterflies have expanded their ranges northward up to 240 km in the past century (Parmesan et al. 1999; Chen et al.

2011). Also in the past century, a number of species in Europe and some in the western

US have shifted to higher elevations. For example, in California, Edith’s Checkerspot butterfly has shifted its range 92 km northward and gained 124 m in elevation since the early 20th century (Parmesan 1996). Such recent geographical and elevational shifts have typically been attributed to climate change (Britten et al. 1994; Parmesan et al.

1999; Parmesan & Yohe 2003; Wilson et al. 2005).

Most studies of shifts in range and phenology in butterflies have been conducted

in Europe or other higher latitude regions (Mattila et al. 2011). There has been an increase in North American studies on this subject in the past decade (Forister & Shapiro

2003; Forister et al. 2010; Breed et al. 2012; Polgar et al. 2013), however, ours is the first study to examine distributional shifts on a southeastern US butterfly species. We found that Speyeria diana is not shifting its range northward, as are many other butterflies.

Instead, it has been expanding into new counties in the mountainous portions of its range, the Ozark and Ouachita Mountains of Arkansas and Oklahoma and the southern

Appalachians of Virginia, Tennessee, and North Carolina, and it has disappeared from the intervening lowlands and from the Atlantic coastal plain. This elevational shift is occurring at the rate of about 18 m per decade at 35o latitude and has led to the complete disjunction of the remaining population centers. This disjunction appeared by the 1950s.

Because these butterflies are at slightly higher elevations in the northern part of their range and appear to be shifting upwards faster there than in the south, there may be complex topography in the North that is moderating a climate response or there may be additional environmental changes at play.

This distributional shift poses a threat to the butterfly, as species restricted to high elevations are predicted to be particularly threatened by climate change in the 21st century

(Walther et al. 2002; Parmesan 2006; Randin et al. 2008). As the species expands upwards in elevation, it may eventually run out of suitable habitat, a phenomenon that threatens a number of taxa with extinction (Marris 2008; García-Camacho et al. 2012;

Fox 2013). It is not known what effects the distributional and phenological shifts in S. diana will have on its food and nectar plants, or its natural enemies, though it seems unlikely they will all shift in concert.

Climate change has been established as the main cause of significant changes in the phenology, or seasonal timing, of many terrestrial species (Crick et al. 1997; Sparks

& Menzel, 2002; Badeck et al. 2004; Schwartz et al. 2006; Morisette et al. 2009; Fox

2013). Many phenological events, including butterfly emergence and flight, are occurring earlier now than in the recent past (Sparks & Yates 1997; Parmesan & Yohe 2003; Root et al. 2003; Stefanescu et al. 2003; Dell et al. 2005). Roy and Sparks (2000) found that the emergence of most British butterflies has advanced by 2 to 10 days per decade, corresponding to warming of 1°C. Dell et al. (2005) reported an even stronger response to increasing spring temperatures in Switzerland, with males and females of Apatura iris

(Lepidoptera: Nymphalidae) emerging respectively 19 and 24 days earlier in 2002 than two decades previous.

We detected a significant shift toward an earlier flight date in females of about 4.3 days per decade at 35o north latitude, which is consistent with most results from northern species, and with the hypothesis of climate warming. In this species, males historically emerged much earlier than females, and this timing has not changed. A shift toward earlier spring flight with climate change could be a threat to future S. diana populations.

If the emergence of S. diana shifts at a different rate than its larval host plant, Viola spp., the result could be a phenological mismatch, decoupling larvae from their only food source (Stenseth & Mysterud 2002).

The shift by S. diana to higher elevations and earlier female flight times are both consistent with predicted effects of climate change, and either result alone has been taken as evidence of such effects. However, the range collapse, but not the phenological shift, also might be due to other regional environmental changes acting separately or jointly.

For example, the disappearance of S. diana from Indiana and Ohio by the late 1800s pre- dates contemporary climate change, leading us to believe that other factors contributed to the loss of those populations. Fox (2013) made a similar point in his review of moth declines in Great Britain. In the sections below, we summarize some of the other major environmental changes to this region that might have affected the butterflies.

Deforestation

Decreases in southern forests have resulted in the decline of a number of North American butterfly species in the United States (Hammond & McCorkle 1983). Agricultural development expanded rapidly in the eastern US by the mid-1800s, greatly increasing rates of deforestation (Healy 1985; Alig 1986). Forest reduction from the timber industry was further accelerated in the late 1800s, resulting in the eventual harvest of virtually all old-growth forests in the Southeast U.S. Since the 1920s, the rate of clearing forests for urban development accounts for the most significant form of land-use change in the southeastern US (Alig 1986; Alig & Wear 1992; Wear & Greis 2002).

Deer herbivory

White-tailed deer (Odocoileus virginianus Boddaert) have played an equally important role in shaping southeastern US forests in the 20th century. White-tailed deer were abundant and widespread at the time of European settlement; however, this species was nearly extirpated during the 19th century as a result of unrestricted hunting, timber harvest, and habitat loss associated with agricultural development (Blackard 1971;

Chollet & Martin 2012). Intense management efforts restored deer populations across

their entire former range (McShea et al. 1997). Deer are currently so abundant across

North America that they have been documented as having widespread negative impacts on vegetation, especially forest understory herbs (Waller & Alverson 1997; Russell et al.

2001; Horsley et al. 2003; Chollet & Martin 2012; Chollet et al. 2013). Chronic deer herbivory has been shown to reduce the abundance of Viola species, the obligate larval host plant for all Speyeria butterflies, in the Great Smoky Mountains (Webster et al.

2005), as well as in the Ozark Mountains of Missouri (Korschgen et al. 1980). Removal of understory vegetation by deer browsing can also significantly affect the microclimate of the forest floor, where Speyeria larvae overwinter, by altering soil temperature, moisture and humidity (Augustine & Frelich 1998). While too much deer browsing may be harmful, Scott Swengel (pers. comm.) suggested that some degree of deer herbivory may be important to keep the understory open for herbaceous plants and their insect specialists.

Pesticide use

Widespread pesticide applications may play a significant role in limiting insect distributions. Efforts to suppress Lymantria dispar, gypsy moth, have resulted in several million hectares of forest land in the southeast US being sprayed regularly with the chemical and biological insecticides Dimilin® and Bacillus thuriengensis (BT) for the past 20 years, despite having detrimental effects on non-target lepidopteran species, including S. diana (Elkinton & Leibhold 1990; Peacock et al. 1998). The grizzled skipper, Pyrgus centaureae wyandot, has disappeared from much of its historical range due to the spraying of BT insecticides and habitat loss (Butler & Kondo 1993). However, while the decline of the P. centaureae wyandot closely matches the historical path of the

US Department of Agriculture’s gypsy moth eradication program as it moved from north to south, (Butler et al. 1995), the decline of S. diana does not. The chemical pesticide dichlorodiphenyltrichloroethane (DDT) was also widely used for decades in the southeast

US to combat mosquito-borne malaria and to protect agricultural crops, such as cotton, peanut and soybean (Flippen 2012). Peak usage of DDT in the United States occurred in

1959 and declined gradually over the 1960s until the United States Environmental

Protection Agency banned its use domestically in 1972. Despite this ban, non-target effects of DDT use have had detrimental and long-term residual effects on a number of organisms and ecosystems that warrant its mention (Newsome 1967; Ware 1980).

Fire regimes

Decades of intense fire suppression, combined with the widespread removal of large herbivores through overhunting, have greatly altered vegetation structure in the southeastern US for the past 150 years (Matthiessen 1959). The elimination of prescribed and natural burning can cause shrub and forest vegetation to accumulate in areas where they normally would not. In an effort to restore pre-European ecological relationships and biodiversity, managers in Arkansas are now actively restoring shortleaf pine-bluestem communities, with forest thinning, and frequent prescribed fire. This has become especially important due to recent efforts to support the recovery of the endangered red- cockaded woodpecker (Picoides borealis Vieillot 1807) (USFWS 2013), which was recently discovered in southern Arkansas (Moran & Baldridge 2002; Thill et al. 2004;

Rudolph et al. 2006). A number of S. diana populations in Arkansas and North Carolina appear to be benefitting from the combination of forest cutting and prescribed burning,

which increases local abundance of high quality nectar plants (Rudolph et al. 2006;

Campbell et al., 2007). The combination of forest thinning and burning has the potential to increase availability of nectar plants used by S. diana, as well as the abundance of

Viola spp. However, it is important to point out that some degree of fire prevention may be essential to provide refugia for local populations, as suggested by Schweitzer et al.

(2011). That is, core areas may need to be thinned regularly but not burned, since a great majority of specialized USA butterflies studied for this effect benefit greatly from having no fires in their core areas (Swengel & Swengel 2007; Swengel et al. 2011)

In conclusion, there are several possible explanations for the widespread disappearance of the Diana fritillary from lowland sites across its range, and any or all of them could be contributing, perhaps in different geographic regions. However, the dramatic shift in phenology uncovered by our historical analysis seems most compatible with expected responses of an ectothermic species to a warming landscape.

We recognize that the possibility of bias in our data collection exists. For example, variation in sampling intensity across space and time could result in a skewed perception of flight time and occurrence. However, butterflies, especially the Diana fritillary, have been well documented in the US since the late 1700s (Cramer and Stoll

1777). The Diana fritillary is a large and spectacular species that has been highly sought after by collectors since its discovery and is unlikely to be missed. In addition, our dataset is the result of a comprehensive effort to collect every single record of S. diana occurrence available to us, which we consider to be highly representative of this species’ distribution. Any study that relies on incomplete information also risks sampling bias.

Our approach, of analyzing essentially the entire cumulative documented

information on a species, could be duplicated with other Lepidoptera. However, it utilizes not only published records, but also information gleaned from traditional natural history collections and modern citizen scientists, and we endorse calls for their increased use and value in science (e.g., Suarez & Tsutsui 2004; Devictor et al. 2010).

Literature Cited

Adams JK, Finkelstein I (2006) Late season observations on female Diana fritillary (Speyeria diana) aggregating behavior. News of the Lepidopterists’ Society 48:106-107.

Alig RJ (1986) Econometric analysis of the factors influencing forest acreage trends in the Southeast. Forest Science 32:119-134.

Alig RJ & Wear D (1992) Changes in private timberlands: Statistics and projections for 1952 to 2040. Journal of Forestry 90:31-37.

Allen J (1997) The butterflies of West Virginia and their caterpillars. University of Pittsburgh Press, Pittsburgh, PA.

Augustine DJ, Frelich LE (1998) Effects of white-tailed deer on populations of an understory forb in fragmented deciduous forests. Conservation Biology 12:995-1004.

Badeck F, Bondeau A, Böttcher K, Doktor D, Lucht W, Schaber J, Sitch S (2004) Responses of spring phenology to climate change. New Phytologist 162:295-309.

Baltosser W (2007) Flitting with disaster: humans and habitat are keys to our state butterfly’s future. Arkansas Wildlife 38:9-12.

Blackard JJ (1971) Restoration of the white-tailed deer to the southeastern United States. Thesis, Louisiana State University, Baton Rouge, LA.

Breed GA, Stichter S, Crone EE (2012) Climate-driven changes in northeastern US butterfly communities. Nature Climate Change 1663:1-5.

Britten HB, Brussard PF, Murphy DD (1994) The pending extinction of the Uncompahgre fritillary butterfly. Conservation Biology 8:86-94.

Brower JVZ (1958) Experimental studies of mimicry in some North American butterflies. Part II. Battus philenor and Papilio troilus. P. polyxenes and P. glaucus. Evolution 12:123-136.

Brower LP, Brower JVZ (1962) The relative abundance of model and mimic butterflies in natural populations of the Battus philenor mimicry complex. Ecology 43:154–158.

Butler L, Kondo V (1993) Impact of Dimilin on non-target Lepidopteria: results of an operational Gypsy Moth suppression program at Coopers Rock State Forest, West Virginia, Agricultural & Forestry Experimental Station Bulletin 710, West Virginia University, Morgantown.

Butler L, Zivkovich C, Sample BE (1995) Richness and abundance of arthropods in the oak canopy of West Virginia’s eastern ridge and valley section during a study of impact of Bacillus thuringiensis with emphasis on macrolepidoptera larvae; Agricultural and Forestry Exp. Station Bulletin 711, West Virginia University, Morgantown.

Campbell JW, Hanula JL, Waldrop TA (2007) Observations of Speyeria diana (Diana fritillary) utilizing forested areas in North Carolina that have been mechanically thinned and burned. Southeastern Naturalist 6:179-182.

Carlton CE, Nobles LS (1996) Distribution of Speyeria diana in the highlands of Arkansas, Missouri and Oklahoma with comments on conservation. Entomological News 107:213-219.

Cech R, Tudor G (2005) Butterflies of the East Coast: an observer’s guide. Princeton University Press, Princeton and Oxford.

Chen IC, Hill JK, Ohlemüller R, Roy DB, Thomas CD (2011) Rapid range shifts of species associated with high levels of climate warming. Science 333:1024-1026.

Chollet S, Martin JL (2012) Declining woodland birds in North America: should we blame Bambi? Diversity and Distributions 19:481-483.

Chollet S, Baltzinger C, Ostermann L, Saint-André F, Martin JL (2013) Importance for forest plant communities of refuges protecting from deer browsing. Forest Ecology and Managment 289:470-477.

Clark AH (1937) Surveying the butterflies of Virginia. The Scientific Monthly 45:256- 265.

Clark AH, Williams CM (1937) Records of Argynnis diana and of some other butterflies from Virginia. Journal of the Washington Academy of Science 207:209-213.

Clark AH, Clark LF (1951) The Butterflies of Virginia, Smithsonian Miscellaneous Collections 116:237.

Cramer P, Stoll C (1777) De uitlandsche Kapellen voorkomende in de drie Waereld- Deelen Asia, Africa en America, II. Deel. De plaataen XCVII-CXCII. (Plate 98, Pages 3- 4).

Crick HP, Dudley DE, Glue DE, Thomoson DL (1997) UK birds are laying eggs earlier. Nature 388:526.

Dell D, Sparks TH, Dennis RLH (2005) Climate change and the effect of increasing spring temperatures on emergence dates of the butterfly Apatura iris (Lepidoptera: Nymphalidae). European Journal of Entomology 102:161-167.

Devictor V, Whittaker RJ, Beltrame C (2010) Beyond scarcity: citizen science programmes as useful tools for conservation biogeography. Diversity and Distributions 16:354-362. dos Passos CF, Grey LP (1945) A genitalic survey of Argynninae (Lepidoptera, Nymphalidae). American Museum Novitees 1296:1-29. dos Passos CF, Grey LP (1947) Systematic catalogue of Speyeria (Lepidoptera, Nymphalidae) with designations of types and fixations of the type localities. American Museum Novitees 1370:1-30.

Dunford JC (2009) Taxonomic overview of the greater fritillary genus Speyeria Scudder and the atlantis-hesperis species complexes, with species accounts, type images, and relevant literature (Lepidoptera: Nymphalidae). Insecta Mundi, 0090:1-74.

Elkinton JS, Liebhold AM (1990) Population dynamics of gypsey moth in North America. Annual Review of Entomology 35:571-596.

Edwards WH (1864) Description of the female Argynnis diana. Proceedings of the Entomolgical Society of Philadelphia 3:431.

ESRI (2011) ArcGIS Desktop: Release 10. Redlands, CA: Environmental Systems Research Institute.

Flippen JB (2012) DDT and the American century: Global health, environmental politics, and the pesticide that changed the world (review). Journal of Interdisciplinary History 43:335-336.

Forister ML, Shapiro AM (2003) Climatic trends and advancing spring flight of butterflies in lowland California. Global Change Biology 9:1130-1135.

Forister ML, McCall AC, Sanders NJ, Fordyce JA, Thorne JH, O’Brien J, Waetjen DP, Shapiro AM (2010) Compounded effects of climate change and habitat shift patterns of butterfly diversity. Proceedings of the National Academy of Science 107:2088–2092.

Fox R (2013) The decline of moths in Great Britain: a review of possible causes. Insect Conservation and Diversity 6:5-19.

French GH (1886) The butterflies of the eastern United States: For the use of classes in Zoology and private students. J.B. Lippincott Company, Philadelphia and London.

García-Camacho R, Albert MJ, Escudero A. (2012) Small-scale demographic compensation in a high-mountain endemic: the low edge stands still. Plant Ecology & Diversity 5:37-44.

Glassberg J (2011) Butterflies of North America. Sterling Publishing, New York, NY.

Global Biodiversity Information Facility (GBIF) Biodiversity occurrence accessed through GBIF Data Portal, www.gbif.net (2012-12-14).

Grey LP (1989) Sundry Argynninae concepts revisited (Nymphalidae). Journal of the Lepidopterists’ Society 43:1-10.

Hammond PC, McCorkle DV (1983) The decline and extinction of Speyeria populaions resulting from human environmental disturbances (Nymphaldae: Argynninae). The Journal of Research on the Lepidoptera 22:217-224.

Healy RG (1985) Competition for land in the American South. Conservation Foundation, Washington, D.C.

Hine JS (1887a) List of butterflies known to have been taken in Ohio. Ohio State Academy of Science 6th Annual Report: 22-27.

Hine JS (1887b) Ohio butterflies. Journal of the Proceedings of the Columbus Horticultural Society 12:77-89.

Holland WJ (1898) The butterfly book. Doubleday and McClure Co.; New York, New York.

Holland WJ (1931) The butterfly book. New and thoroughly revised edition. Doubleday, Doran and Co., Garden City, New York.

Horsley SB, Stout SL, DeCalesta DS (2003) White-tailed deer impact on the vegetation dynamics of a northern hardwood forest. Ecological Applications 13:98-118.

Hovanitz W (1963) Geographical distribution and variation of the genus Argynnis. III. Argynnis diana. Journal of Research on the Lepidoptera 1:209-221.

Irwin RR, Downey JC (1973) Biological notes no. 81, Annotated Checklist of the Butterflies of Illinois. State of Illinois Department of Registration and Education, Urbana, IL.

JMP, Version 10. SAS Institute Inc., Cary, NC, 2012.

Klots AB (1951) A field guide to the butterflies of North America, east of the Great Plains. Houghton Mifflin Company; Boston, Massachusetts.

Korschgen LJ, Porath WR, Torgerson O (1980) Spring and summer foods of deer in the Missouri Ozarks. The Journal of Wildlife Management 44:89-97.

LeGrand HE, Howard TE (2012) Butterflies of North Carolina 19th Approximation, http://ncparks.gov/19th.pdf

Marris E (2008) The escalator effect. Nature Reports Climate Change 1:94-96.

Matthiessen P (1959) Wildlife in America. Viking Press, New York.

Mattila N, Kaitala V, Komonen A, Päivinen J, Kotiaho JS (2011) Ecological correlates of distribution change and range shift in butterflies. Insect Conservation and Diversity 4:239-246.

McShea WJ, Underwood HB, Rappole JH (1997) The Science of Overabundance: Deer Ecology and Population Management. Smithsonian Institution Press, Washington D. C.

Moran M, Baldridge C (2002) Distribution of the Diana Fritillary, Speyeria diana (Nymphalidae) in Arkansas, with notes on nectar plant and habitat preference. Journal of the Lepidopterists’ Society 56:162-165.

Morisette JT, Richardson AD, Knapp AK, Fisher JI, Graham EA, Abatzoglou J, Wilson BE, Breshears DD, Henebry GM, Hanes JM, Liang L (2009) Tracking the rhythm of the seasons in the face of global change: phenological research in the 21st century. Frontiers in Ecology and the Environment 7:253-260.

Miller LD, Brown FM (1981) A catalogue/checklist of the butterflies of America North of Mexico. Memoirs No. 2. Florida: Lepidopterists’ Society vii.

Natureserve (2013) NatureServe Explorer: An online encyclopedia of life [web application]. Version 7.0. NatureServe, Arlington, VA. USA Available http://www.natureserve.org/explorer. (Accessed: 1/1/2013).

National Oceanic and Atmospheric Administration (NOAA) (2012) U.S. Climate Extremes Index. www.ncdc.noaa.gov/oa/climate/research/cei/cei.html. (Accessed: 1/1/2013).

Newsom LD (1967) Consequences of insecticide use on nontarget organisms. Annual Review of Entomology 12:257-286.

North American Butterfly Association (2000) NABA butterfly counts 1989-2010: North American Butterfly Association. Morristown, NJ.

Opler P, Malikul V (1992) A field guide to eastern butterflies. Houghton Mifflin Company, NY, NY.

Opler P, Krizek G (1984) Butterflies east of the Great Plains. Johns Hopkins University Press, Baltimore, MD.

Opler P, Lotts K, Naberhaus T (2012) Butterflies and Moths of North America. http://www.butterfliesandmoths.org/. (Accessed 2/07/2012).

Parmesan C (1996) Climate and species’ range. Nature 382:765-766.

Parmesan C (2006) Ecological and evolutionary responses to recent climate change. Annual Review of Ecology, Evolution, and Systematics 37:637–69.

Parmesan C, Ryrholm N, Stefanescu C, Hill J, Thomas C, Descimon H, Huntley B, Kaila L, Kullberg J, Tammaru T, Tennent, T, Thomas J, Warren M (1999) Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399:579-583.

Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37-42.

Peacock JD, Schweitzer JC, Dubois N (1998) Laboratory assessment of the effects of Bacillus thuringiensis on native Lepidoptera. Environmental Entomology 27:450-457.

Polgar CA, Primack RB, Williams EH, Stichter S, Hitchcock C (2013) Climate effects on the flight period of Lycaenid butterflies in Massachusetts. Biological Conservation 160:25-31.

Poulton EB (1909) Mimicry in butterflies of North America. Annals of the Entomological Society of America 2:203–242.

Randin CF, Engler R, Normand S, Zappa M, Zimmermann NE, Pearman PB, Vittoz, P, Thuiller W, Guisan A (2008) Climate change and plant distribution: local models predict high‐elevation persistence. Global Change Biology 15:1557-1569.

Ross GN (2008) Diana’s Mountain Retreat. Natural History 72:24-33.

Ross GN, Henk MC (2004) Notes on eggs and first instar larvae of three species of Speyeria (Nymphalidae). News of the Lepidopterists’ Society 46:53-63.

Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA (2003) Fingerprints of global warming on wild animals and plants. Nature 421:57–60.

Roy DB, Sparks TH (2000) Phenology of British butterflies and climatic change. Global Change Biology 6:407-416.

Rudolph DC, Ely AC, Schefer RR, Williamson JH, Thill RE (2006) The Diana fritillary (Speyeria diana) and Great Spangled fritillary (S. cybele): Dependence on fire in the Ouachita Mountains of Arkansas. Journal of the Lepidopterists’ Society 60:218-226.

Russell FL, Zippin DB, Fowler NL (2001) Effects of White-Tailed Deer (Odocoileus virginianus) on Plants, Plant Populations and Communities: A Review. American Midland Naturalist 146:1-26.

Scholtens B (2004) Survey for Speyeria diana in Sumter National Forest (Oconee Co., SC). Report to National Forest Service. Online: http://www.dnr.sc.gov/wcp/pdf/DianaFritillary.pdf

Scudder SH (1889) The butterflies of the Eastern United States and Canada with special reference to New England. 3 Vols. Published by author, Cambridge, Mass.

Schwartz MD, Ahas R, Aasa A (2006) Onset of spring starting earlier across the northern hemisphere. Global Change Biology 12:343-351.

Schweitzer DF, MC Minno, Wagner DL 2011. Rare, declining, and poorly known butterflies and moths (Lepidoptera) of forests and woodlands in the eastern United States. US Forest Service, Forest Health Technology Enterprise Team, FHTET-2011-01. Morgantown, WV.

Shuey JA, Calhoun JV, Iftner DC (1987) Butterflies that are endangered, threatened, and of special concern in Ohio. The Ohio Journal of Science 87:98-106.

Simonsen TJ, Wahlberg N, Brower AZ, de Jong R (2006) Morphology, molecules and fritillaries: Approaching a stable phylogeny for Argynnini (Lepidoptera: Nymphalidae). Insect Systematics and Evolution 37:405-418.

Skinner H (1889) A list of the butterflies of Philadelphia, PA. The Canadian Entomologist 11:126-131.

Sparks TH, Menzel A (2002) Observed changes in seasons, an overview. International Journal of Climatology 22:1715-1726.

Sparks TH, Yates TJ (1997) The effect of spring temperature on the appearance dates of British butterflies 1883–1993. Ecography 20:368–374.

Stefanescu C, Penuelaslla J, Filella I (2003) Effects of climatic change on the phenology of butterflies in the north west Mediterranean Basin. Global Change Biology 9:1494- 1506.

Stenseth NC, Mysterud A (2002) Climate, changing phenology, and other life history traits: Nonlinearity and match mismatch to the environment. Proceedings of the National Academy of Sciences 99:13379-13381.

Strecker H (1878) Butterflies and moths of North America. Press of B.F. Owen. Reading, PA.

Suarez AV, Tsutsui ND (2004) The value of museum collections for research and society. Bioscience 54:66-74.

Swengel AB, Swengel SR (2007) Benefit of permanent non-fire refugia for Lepidoptera conservation in fire managed sites. Journal of Insect Conservation 11:263–279.

Swengel SR, Schlicht D, Olsen F, Swengel AB (2011) Declines of prairie butterflies in the midwestern USA. Journal of Insect Conservation 15:327-339.

Thill RE, Rudolph DC, Koerth NE (2004) Shortleaf pine-bluestem restoration for red- cockaded woodpeckers in the Ouachita Mountains: implications for other taxa. In: R. Costa, S.J. Daniels (eds.). Red-cockaded woodpecker: Road to recovery. Blaine, WA: Hancock House Publishers: 657-671.

United States Fish & Wildlife Service. (2003) Recovery plan for the red-cockaded woodpecker (Picoides borealis): second revision. U.S. Fish and Wildlife Service, Atlanta, Georgia.

United States Geological Survey (1987) Digital Elevation Models: Data Users Guide 5. Department of the Interior, USGS, Reston, Virginia.

United States Geological Survey (1998) USGS Geospatial Data Clearinghouse, National Mapping and Remotely Sensed Data: Digital Elevation Models (DEMs). Available http://edcwww.cr.usgs.gov/nsdi/gendem.htm (Accessed: 1/1/2013).

Waller DM, Alverson WS (1997) The white-tailed deer: a keystone herbivore. Wildlife Society Bulletin 25:217-226.

Walther GR, Post E, Convey P, Menzel A, Parmesan C, Beebee T, Fromentin JM, Hoegh-Guldberg O, Bairlein F (2002) Ecological responses to recent climate change. Nature 416:389-395.

Ware GW (1980) Effects of pesticides on nontarget organisms. In Residue reviews (pp. 173-201). Springer New York.

Wear DN, Greis JG (2002) Southern Forest Resource Assessment. General Technical Report, SRS-53. U.S. Department of Agriculture, Forest Service, Southern Research Station. Asheville, NC.

Webster CR, Jenkins MA, Rock JH (2005) Long-term response of spring flora to chronic herbivory and deer exclusion in Great Smoky Mountains National Park, USA. Biological Conservation 125:297–307.

Wells CN, Edwards L, Hawkins R, Smith L, Tonkyn D (2011) A rearing method for Argynnis (Speyeria) diana (Lepidoptera: Nymphalidae) that avoids larval diapause. Psyche, doi:10.1155/2011/940280

Wilson RJ, Gutiérrez D, Gutiérrez J, Martínez D, Agudo R, Monserrat VJ (2005) Changes to the elevational limits and extent of species ranges associated with climate change. Ecology Letters 8:1138-1146.

Table 1.1 Summary of Speyeria diana distributional data sources

National Museums (N. American) Location No. of S. diana Range of specimen dates No. of Counties Carnegie Museum of Natural History Pittsburgh, Pennsylvania 142 1889-2000 26 National Museum of Natural History Washington, D.C. 129 1907-2002 26 American Museum of Natural History New York, New York 104 1921-1985 28 The Field Museum Chicago, Illinois 98 1889-1995 23 California Academy of Sciences San Francisco, California 88 1886-2000 12 Georgia Museum of Natural History‡ Athens, Georgia 15 1935-1987 8 Cleveland Museum of Natural History‡ Cleveland, Ohio 6 1921-1965 6 Denver Museum of Nature and Science Denver, Colorado 4 1939-1973 3 Mount Magazine State Park Paris, Arkansas 4 1997 1

National History Museums (European) British Natural History Museum London, UK 31 1777-1989 17 Paris Muséum national d'Histoire naturelle Paris, France 8 1890 1 Oxford Museum of Natural History‡ Oxford, UK 4 1937-1971 4 Zoölogisch Museum Amsterdam Amsterdam, Netherlands 4 1884-1921 3 NMC Naturalis Leiden, Netherlands 4 ------Royal Ontario Museum‡ Ontario, Canada 3 1933-1968 3

University Collections University of Florida Gainesville, Florida 409 1900-2007 43 University of Michigan‡ East Lansing, Michigan 66 1909-1985 13 Clemson University Clemson, South Carolina 43 1926-1978 5 Peabody, Yale University‡ New Haven, Connecticut 29 1904-1961 8 University of Missouri‡ Columbia, Missouri 29 1886-1980 8 University of Wyoming‡ Laramie, Wyoming 13 1955-1979 4 University of Arkansas, Little Rock Little Rock, Arkansas 12 2005-2007 5

University of California, Berkley‡ Berkley, California 12 1926-1981 6 University of Nebraska‡ Lincoln, Nebraska 14 1954-2003 7

North Carolina State University‡ Raleigh, North Carolina 10 1904-1964 9 University of Arkansas, Fayetteville‡ Fayetteville, Arkansas 10 1977-1994 5 Virginia Polytechnic Inst Blacksburg, Virginia 8 1911-1977 1 Louisiana State University‡ Baton Rouge, Louisiana 7 1984-1988 1 University of Wisconsin‡ Madison, WI 5 1926-1951 2 College of Charleston‡ Charleston, South Carolina 4 2008 2 West Virginia University‡ Morgontown, West Virginia 3 1977-1995 2 Furman University‡ Greenville, South Carolina 3 1929-1990 3 Dalton State College‡ Dalton, Georgia 2 2001 1

State Agencies, online databases, listserves, individuals, and organizations Field Surveys 469 1995-2012 46 Butterflies and Moths of America (BAMONA) 435 1938-2012 39 North Carolina 19th Approximation (http://149.168.1.196/nbnc/) 276 1938-2011 31 West Virginia Divisions of Natural Resources (wvdnr.gov) 204 1978-1999 11 Literature survey 153 1818-2011 54 Kentucky Dept. of Fish and Wildlife Resources (fw.ky.gov) 146 1936-2006 21 NABA annual count data (naba.org) 103 1999-2010 27 Georgia Dept. of Natural Resources (gadnr.org) 77 1994-2001 15 Global Biodiversity Information Facility (GBIF) 75 1974-2004 49 North Carolina Natural Heritage Program (nchp.org) 69 1989-2003 21 The Lepidopterists’ Society (lepsoc.org) 50 1973-2008 25 All Taxa Biodiversity Inventory (ATBI) (dlia.org/atbi) 46 1936-2007 4 Carolina Butterfly Society (CBS) 44 2001-2009 5 Carolinaleps 41 2007-2009 9 Washington Area Butterfly Club 29 2007 1 Oklahoma Leps 21 2005-2009 5 Insect.net 21 2007-2009 9

Table 1.2 Summary of literature referencing the distribution of Speyeria diana Reference Location Date of record(s) Description Cramer & Stoll 1775 Jamestown, Virginia 1775 holotype; male described by Pieter Cramer Blatchley 1859 Vanderburgh County, Indiana 1850s first record from Indiana, most northern record Edwards 1864 Kanawha, West Virginia 20-31 Aug, 1864 first description of female, took over 30 specimens Edwards 1874 Coalburgh, West Virginia Aug, Sept 1873 description of rearing Argynnis larvae Aaron 1877 Tennessee/ North Carolina 1877 populations are ample along Blue Ridge Kentucky 1877 locally abundant populations Strecker 1878 1878 West Virginia, Georgia, Kentucky, Tennessee, Arkansas Thomas 1878 Kentucky, Arkansas, southern Illinois 1878 common in Kentucky & Arkansas Fisher 1881 Illinois 1880 present in southern Illinois Holland 1883 Salem, North Carolina 1858-1861 described as “first pinned female specimen” Edwards 1884 southern Ohio 1880s first description in Ohio Hulst 1885 Waynesville, North Carolina 1882 locally abundant populations Warren Springs, North Carolina 1882 very common along the French Broad River Blatchley 1886 Evansville, Indiana early 1900s locally abundant populations French 1886 eastern United States 1886 W. Virginia to Georgia, Southern Ohio to Illinois, Kentucky, Tennessee, Arkansas Hine 1887a, b Medina County, Ohio 8/9/1887 single worn male, northernmost record in OH Kingsley 1888 Virginia 1887 Argynnis diana is described as the handsomest insect found in the United States Scudder 1889 southeast United States 1880s Semnopsyche diana; an inhabitant of hilly country of the south,38th parallel of latitude, taken as far west as Missouri & “Arkansaw” Skinner & Aaron 1889 Pennsylvania 1880s stray individual found in Pennsylvania Dixey 1890 eastern United States 1889 description of Argynnis diana wing spot pattern Blatchley 1891 Illinois 1890s female specimen from northern Danville, IL

Skinner 1896 southern Illinois 1890s Diana specimens from southern Illinois are larger than those further east Holland 1898 southern United States 1890s in two Virginias and Carolinas, northern Georgia, Tennessee, Kentucky, occasionally in s. Ohio and Indiana, and in Missouri and Arkansas; the most magnificent & splendid species of the genus Snyder 1900 Clay County, Illinois 1900 northern limit of S. diana in Illinois Strecker 1900 Missouri 1853 pair captured in copula, very early female Maynard 1901 habitat is West Virginia to Georgia, southern Ohio to Illinois, Tennessee & Arkansas Sell 1916 Greene County, Missouri 8/22/1900 southeast of Springfield Smyth 1916 southeast United States 1880-1916 Asheville, Brevard, North Carolina, Caesar’s Head, South Carolina, Montgomery, Washington & Giles Counties, Virginia Wood 1916 Camp Craig, Virginia August 1914 describes female color variation Murrill 1919 Virginia 1919 Poverty Valley Holland 1931 1930s The Virginias and Carolinas, northern GA Tennessee, Kentucky, occasionally in s. OH, Indiana, and in Missouri and Arkansas Knobel 1931 Hope, Arkansas 1930 from Mrs. Louise Knobel Kite 1934 Taney County, Missouri 7/31/1925 male and female reported Clark 1937 Virginia 1930s ranges from Bath County, Virginia to FL east almost to tidewater, and west to Illinois & Arkansas Clark & Williams 1937 Virginia late 1800s-1935 Bath, Alleghany, Giles, Bland, Dickenson, Smyth, Patrick, Montgomery & Washington Counties Allen 1941 West Virginia 1940 Pocahontas County, west to Kanawha and Lincoln Counties; abundant in Jefferson NF (Monroe County), Babcock State Park (Fayette County), and Fork Creek Wildlife Management Area (Boone County)

Chermock 1942 Conestee Falls, North Carolina summer 1941 s. Ohio and West Virginia, through the Appalachian mountains into Georgia and South Carolina, most abundant in mountains s. of Great Smoky Mountains National Park Bock 1949 Cincinnati, Ohio 1947 author collects hundreds of specimens from North Carolina mountains; gone from Indiana and Ohio Clark & Clark 1951 Southern Illinois early 1900s Chesterfield County, Virginia 1930 last known county record Northampton County, Virginia 1930 last known county record Klots 1951 Brevard, North Carolina 1950 in large numbers along roadsides; Chiefly in mountains and piedmont, W. Virginia s. to Georgia, w. to southern Ohio, Indiana, Missouri & Arkansas Mather & Mather 1958 Madison Parish, Louisiana 1958 record is a stray individual Evans 1959 Smoky Mountains of Tennessee September 1957 identification of an unknown S. diana larva Curtis & Boscoe 1962 Buncombe County, North Carolina 6/27/1962 collecting record near Asheville Hovanitz 1963 Salem, Roanoke County, Virginia 6/13/1937 comprehensive distribution data Ross & Lambremont 1963 Louisiana 1950s stray record from Mather & Mather 1958 Masters 1968 Newton County, Missouri 1960s locally very common Masters & Masters 1969 Perry County, Indiana 7/15/1962 last record known from Indiana Shull & Badger 1971 Indiana 1971 no longer resident in Indiana Harris 1972 Georgia 1972 summarizes historic reports from White, Union, Fannin, Habersham, Rabun Counties Irwin & Downey 1973 Vermilion County, Illinois 8/20/1960 female, last known Illinois record Southern Illinois 1880 Illinois natural history survey Howe 1975 1950s extirpated from type locality, Jamestown Kentucky, West Virginia 1970s species is scarce in Kentucky & West Virginia Georgia 1970s not uncommon in northern Georgia Ceasar’s Head, South Carolina 1970s stable populations, not uncommon Nelson 1979 Ozark plateau of Oklahoma 1969 only found in eastern counties

Schowalter & Drees 1980 Poverty Hollow, Virginia 1973, 1978 field-captured and lab-reared S. diana gynandromorphs described in detail Pyle 1981 eastern United States 1980s has decreased its range due to forest loss, common in the Great Smoky Mountains Hammond & McCorkle 1983 Virginia & Tennessee 1975-1978 Appalachian populations are expanding Opler 1983 eastern United States 1980s some populations under decline Opler & Krizek 1984 1950s extirpated from Virginia Piedmont and coast 1800s extirpated from Ohio River valley Shuey et al. 1987 Cincinnati, Ohio 1900s-1930 eliminated by deforestation by early 1900s Shull 1987 Indiana late 1800s occurs in mountains and piedmont of West Virginia south to Georgia, west to southern Ohio, Indiana, Missouri, & Arkansas Watson & Hyatt 1988 Tennessee 1980s resident species of northeastern Tennessee Kohen 1989 Cumberland, Kentucky July 1984 aberrant male on milkweed Cohen & Cohen 1991 Bath County, Virginia 1990 George Washington National Forest Montgomery County, Virginia 1990 photograph of pair in copula Krizek 1991 western Virginia 7/11/1991 males preferred nectar over horse manure Adams 1992 Fannin County, Georgia 8/28/1992 female netted by Irving Finkelstein Opler & Malikul 1992 eastern United States 1992 central Appalachians west to Ozarks, formerly Atlantic coastal plain of Va. & NC & Ohio River Valley, rich forested valleys Skillman & Heppner 1992 Coopers Creek WMA Georgia 6/10/1988 Gynandromorph specimen found in n. GA Carlton & Nobles 1996 Arkansas, Missouri, Oklahoma 1819-1995 survey of Interior Highlands Allen 1997 West Virginia 1997 ranges from Virginia & W. Virginia south to northern Georgia and Alabama. A small population persists in Ozark Mountains of Arkansas and Missouri Ross 1997 Coweeta Forest, North Carolina 1990, 1996 classified as uncommon, 2-5 indivs. sighted Ross 1998 Mount Magazine, Arkansas 6/30/1993 photograph of male, locally abundant Mount Magazine, Arkansas 8/20/1992 photograph of female, locally abundant Glassberg 1999 eastern United States 1999 formerly throughout Ohio River Valley and southeastern Virginia & northwest N.C. Moran & Baldridge 2002 Arkansas, Missouri, Oklahoma 1997-1999 22 counties inhabited, Arkansas expanding

Scholtens 2004 Oconee County, South Carolina 2002 present in Sumter National Forest Cech & Tudor 2005 2000s locally common in mountain colonies, s. W. Virginia to n. GA; also e. AL/KY, Ozarks Vaughan & Shepherd 2005 Red List species profile 2005 core of species distribution is in the southern Appalachians from central Virgina and W. VA through the mountains to northern Georgia & Alabama. Also in Ozarks of Missouri, Arkansas, and eastern Oklahoma Adams & Finkelstein 2006 Fannin County, Georgia 10/12/2006 lots of aggregating females flying late Rudolph et al. 2006 Ouachita Mountains, Arkansas 1999-2005 feeding records by month sites Spencer 2006 Arkansas 2006 uncommon to locally common in colonies Scattered throughout the Interior Highlands Coastal Plain Campbell et al. 2007 North Carolina 6/17/2004 at least 4 males visiting flowering sourwood Ross 2008 Mount Magazine, Arkansas 2008 description of Mount Magazine State Park Wells et al. 2010 Mount Magazine, Arkansas 2009 copulating pair photographed Wells et al. 2011 Georgia, North Carolina, Tennessee 2009 females collected for rearing trial

Table 1.3 Field-sampled Speyeria diana (2006-2009). Records are provided to the level of county

State County Ecoregion # S. diana (m/f) Survey dates Arkansas Benton Ozark Plateau 7 (7/1) 12-14 June 2007, 22-23 June 2009 Carroll Ozark Plateau 9 (7/2) 15-16 June 2007, 23-24 June 2009 Boone Ozark Plateau 2 (2/0) 16 June 2007 Faulkner Arkansas River Valley 5 (5/0) 18-20 June 2006, 20 June 2007, 16 June 2008, 3-6 Aug 2009 Conway Arkansas River Valley 15 (11/4) 22 June 2007, 26 June 2008, 5 Aug 2009 Pulaski Arkansas River Valley 4 (2/2) 28 Aug 2009 Logan Arkansas River Valley 37 (29/8) 20-24 June 2006, 21-24 June 2007, 1-3 Aug 2009 Montgomery Ouachita Mountains 12 (7/5) 31 July 2008, 1-3 Sept 2009 Polk Ouachita Mountains 5 (1/4) 1-3 Sept 2009 Saline Ouachita Mountains 8 (7/1) 14 June 2008, 18 June 2009 Oklahoma Leflore Ouachita Mountains 3 (0/3) 30 Aug 2009 Georgia Fannin Blue Ridge Mountains 26 (17/9) 12-13 July & 1 Aug 2006, 12 July 2007, 22 June & 20 July 2008 Rabun Blue Ridge Mountains 8 (2/6) 7 Sept 2008, 29 Aug 2009 Union Blue Ridge Mountains 14 (6/8) 29 July 2007, 15 June & 5-7 Aug 2008,

North Carolina Ashe Blue Ridge Mountains 4 (4/0) 22-23 June 2007 Buncombe Blue Ridge Mountains 13 (8/5) 27 July 2006, 30 July 2007, 9 Aug 2008 McDowell Blue Ridge Mountains 15 (10/5) 9 Sept 2007, 24 June 2008, 30 June, 11 Sept 2009 Transylvania Blue Ridge Mountains 24 (19/5) 5 June 2006, 16 July & 5 Sept 2007, 14 June 2008, 26 June 2009 Watauga Blue Ridge Mountains 7 (5/2) 30 May & 9 June 2006, 25 July 2008, 19 Sept 2009

South Carolina Greenville Blue Ridge Escarpment 12 (7/5) 31 June 2006, 27-29 July 2007, 1 Sept 2008, 8-13 Sept 2009

Tennessee Blount Great Smoky Mountains 42 (33/9) 1-26 June 2007, 1-28 June & 20-29 Aug 2008, 1-15 Sept 2009

Sevier Great Smoky Mountains 33 (25/8) 1-26 June 2007, 26-29 June 2008, 5 June-26 Sept 2009 Carter Appalachian Mountains 57 (35/22) 5-9 June & 5-11 July 2006, 30-31 May 2007, 29-30 Aug 2008 Sullivan Appalachian Mountains 36 (25/11) 13-16 July 2006, 20-22 July 2007, 5 Aug, 18-20 Sept 2009 Virginia Montgomery Appalachian Mountains 21 (14/7) 3-7 July 2007, 2-4 July 20

Figure 1.1 All distributional records for Speyeria diana specimens documented from 1777 to 1960. This species once ranged widely across the southeastern United States, from the coastal plains of North Carolina and Virginia across the Appalachian Mountains and the Ohio River Valley, and west to Arkansas and Missouri

Figure 1.2 All distributional records for Speyeria diana specimens documented from 1961 to 2010. This species has experienced a range collapse resulting in two geographically separated population groups: the southern Appalachian Mountains in the east, and the Ozark and Ouachita Mountains in the west. Populations have disappeared from the intervening lowlands and the eastern coastal plain.

(a)

(b)

Figure 1.3 (a) Latitude was negatively correlated with the year of S. diana capture, reflecting the recent expansion of S. diana into its southern range boundary (Latitude = 53-0.008*year; R2 = 0.04, n = 1680, P < 0.0001); (b) Longitude was also significantly and negatively correlated with year, reflecting westward expansion (Longitude = -35-0.02*year; R2 = 0.03, n = 1680, P < 0.0001)

Figure 1.4 Speyeria diana are shifting to significantly higher elevations in recent times, and this shift varies slightly with latitude (Elevation (m) = 212 - 74.3*latitude + 0.4*year + 0.04*year*latitude; R2= 0.33, n = 1680, all effects significant at P < 0.0001). Butterflies were at slightly higher elevations and climbing slightly faster in the north than in the south

Figure 1.5 Speyeria diana females are flying progressively earlier in recent times, and this varies slightly with latitudes (Flight day of year = 984 + 2.34*latitude -0.57*Year +0.004*Year*latitude; R2= 0.19, n = 840, all effects significant at P < 0.0001).

Figure 1.6 Speyeria diana males have a statistically significant relationship between flight date, year and latitude (Flight day of year = 123 + 1.32*latitude - 0.34*year + 0.01*year*latitude; R2= 0.13, n = 840, all effects significant at P < 0.0001).

CHAPTER TWO

THE PHYLOGEOGRAPHIC HISTORY OF THE THREATENED DIANA FRITILLARY, SPEYERIA DIANA (LEPIDOPTERA: NYMPHALIDAE), WITH IMPLICATIONS FOR CONSERVATION

Abstract

The Diana fritillary, Speyeria diana (Cramer 1777) (Lepidoptera: Nymphalidae), is a

North American endemic butterfly that disappeared from low-elevation sites throughout its range in the 20th century. It now persists in two geographically isolated mountainous regions, with an

800 km disjunction. Using mitochondrial cytochrome oxidase II (COII) DNA sequences from museum and field-sampled specimens, we found greater mtDNA diversity and more widespread differentiation among eastern populations than western ones. Using coalescent-based population divergence models, we dated the earliest splitting of eastern and western populations at approximately 20,000 years ago, during the Last Glacial Maximum. The recent range collapse across the center of the historical species distribution may have exacerbated an ancient genetic differentiation between eastern and western populations. Finally, the loss of lowland haplotypes and the relatively large variation among local populations suggests that dispersal is low and lowland populations did not move to higher elevations, but, rather, appear to have vanished. Our results highlight the value of incorporating genetic data from preserved specimens when investigating the phylogeographic history and conservation status of a threatened species.

Introduction

Habitat connectivity is important for maintaining genetic variation in natural populations.

Restrictions in range and fragmentation of habitat can lead to an overall decrease in available habitat and reductions in population size for the species. The ecological and evolutionary

consequences of habitat fragmentation are typically most pronounced in small populations, which are expected to hold less genetic variation than the system as a whole (Broquet et al.

2010). The impacts of habitat fragmentation will also depend on how much migration is maintained among remaining patches. When populations are both small and isolated, they become even more vulnerable to extinction from demographic and genetic processes, including a reduced ability to respond evolutionarily to random and directional changes in the environment

(Frankham et al. 2002).

Habitat loss and population fragmentation have caused several North American butterfly species from the fritillary genus Speyeria to become threatened with extinction over the past 200 years (Hammond and McCorkle 1983; Williams 2002; Cech and Tudor 2005). The Diana fritillary, Speyeria diana (Cramer), is an example, having disappeared from large portions of its former distribution. Historically, this species was distributed across the southeastern US, ranging from coastal Virginia, up to the Ohio River Valley, and west to Arkansas and Missouri (Opler and Krizek 1984; Moran and Baldridge 2002). Over the past century, S. diana populations have disappeared entirely from the Ohio River Valley and coastal habitat in North Carolina and

Virginia (Cech and Tudor 2005). The result is a geographic disjunction of close to 800 km between remaining eastern populations in the southern Appalachian Mountains from Georgia to

Virginia and western populations in the Ozark and Ouachita Mountains of Arkansas and

Oklahoma (Wells and Tonkyn 2013) (Figure 2.1).

Nothing is known about the genetic and demographic consequences of recent population extinctions across most of the range of S. diana, or the phylogeographic background against which these extinctions have occurred. To address this, we characterized the population genetic structure of S. diana throughout its range by sequencing a 549-bp segment of the cytochrome 57

oxidase II (COII) mitochondrial gene from field-sampled and museum S. diana specimens collected over the past century. The purpose of our study is to describe patterns of genetic variation and population structure in the threatened Diana fritillary and to use those patterns to reconstruct the demographic history of this species, using coalescent methods. Given that S. diana was continuously distributed across the southeastern U.S. a little over a century ago, we ask if there is measurable genetic differentiation across the present range and, if so, did it arise concurrent with the recent range collapse (Wells and Tonkyn 2013).

Studies based solely on mtDNA may not accurately represent the entire phylogeographic history of populations and species, but they remain a useful first step in phylogeographic analyses, particularly for studies involving ancient DNA or poorly preserved museum specimens

(Avise 2000). Our study involved extracting DNA from a number of old, often degraded, museum specimens, and because the successful amplification of full-length DNA sequences from dried and pinned insect specimens preserved over ten years is generally lower than that of fresh samples (Hajibabaei et al. 2006), we relied on mtDNA data to produce reliable sequence data for our analysis. While the addition of additional unlinked markers would greatly improve the ability to infer the population history, we could not amplify other loci described for the related S. idalia (Williams et al. 2003), and consider a mitochondrial locus to be an appropriate marker for our phylogeographic study.

Methods

Field sampling

During the summers of 2006-2009, we sampled 11 S. diana populations across the species’ entire current range (Figure 2.2). These include seven populations from the southern

Appalachian Mountains, from Georgia to Virginia, representing the eastern range and four populations, including the Ozark and Ouachita Mountains of Arkansas and Oklahoma, and

Mount Magazine, the highest point in Arkansas (839 m), representing the western range. Adult butterflies were captured with a handheld net and non-lethally sampled by removal of a single posterior tarsus. Tarsi were preserved in 95% ethanol and stored at -20°C.

Museum specimens

Samples from natural history museums and other collections were obtained to represent extirpated portions of the species’ former distribution, including coastal Virginia (1900-1920s),

Indiana (1928-1934), and Ohio (1911-1930), all populations that were extirpated by the 1950s

(Table 1). A single rear tarsus was removed from dried and pinned museum specimens for DNA extraction.

DNA extraction

The DNeasy kit (Qiagen, Inc., Valencia, CA) was used to extract DNA from fresh samples of S. diana, from approximately half of each tarsal segment. Each sample was ground in liquid nitrogen with a mini-pestle and incubated overnight with 5µL of proteinase K in a heated

(56°C) shaking block (900 rpm). Purified DNA was eluted in 50 µL of 50% buffer EB and 50% sterile water. For the museum samples, we used the forensic DNA extraction protocol from the

QIAamp DNA Investigator kit (Qiagen, Inc., Valencia, CA). Whole tarsi were incubated overnight with 20 µL of proteinase K in a heated (56°C) shaking block (900 rpm). DNA was eluted in 50 µL of sterile water, dried via vacuum centrifuge, and stored at -20°C.

PCR amplification and haplotype sequencing

Species-specific primers for the second mitochondrial subunit of COII (SdFCOII: 5’-

TGG CAG ATT ATA TGT AAT GGA TT-3’; SdRCOII: 5’-TAA TCG TCC AGG RTT AGC

RTC A-3’) were designed to amplify 549 bp of DNA from the closely related great spangled fritillary, Speyeria cybele (Genbank Accession No. AF492412). We performed PCR reactions in

50 µL volumes, consisting of 10µL of 5X Green buffer, 3µL (1.25mM) dNTPs, 2µL (35.2µM) primer SdFCOII, 2µL (23.2.2µM) primer SdRCOII, 0.7 U GoTaq polymerase (Promega),

30.75µL sterile Sigma® water, and 2 µL of genomic DNA. The amplification cycle consisted of an initial denaturing step at 94°C for 30 s, followed by 40 cycles of 30 s at 40°C, and 60 s at

68°C. PCR products were visualized on a high-melt 0.8% agarose gel, excised, and purified using a QIAquick Gel Extraction Kit (Qiagen Inc., Valencia, CA).

All of the field-sampled specimens amplified easily with the first round of PCR. Out of

72 museum specimens tested, only 30 specimens yielded enough amplification product for sequencing. For these individuals, gel-purified products (5 µL of a gel-isolated plug melted in

100 uL water) were used as a template in an additional PCR with a second set of (nested) primers designed to amplify a 470 bp fragment of COII (SFCOIIb: 5’-TGT AAT GGA TTT AAA CCC

CA -3’; SRCOIIb: 5’- GTT AGC TCA ACT TTT ACT CCA -3’). We used the following amplification cycle for the nested PCR: 96°C for 1 minute, 25 cycles of 50°C for 5 sec, followed by slow ramp (1°C /sec) to 60°C for 4 min. Gel-purified PCR products were sequenced in both directions using 3µL of DNA and 2µM of primers, SdFCOII and SdRCOII. All sequencing reactions were visualized on an ABI 3130 automated sequencer, chromatograms were edited using Sequencher v.4.2.2 (Gene Codes Corp., Ann Arbor, MI, USA), and sequences were aligned with Clustal-X v.2.1 (Larkin et al. 2007). All of the sequences were deposited in

GenBank.

Population Genetic Diversity and Structure

We used MODEL TEST 3.8 (Posada 2006) to identify the best substitution model for

COII, HKY+I, under the Bayesian Information Criteria (BIC). We used ARLEQUIN (Excoffier et al. 2010, version 3.5) to estimate haplotype diversity (h) and nucleotide diversity (π), and to generate a matrix of pairwise ΦST values based on pairwise differences between haplotypes. We evaluated statistical significance based on 1000 permutations of the data and applied sequential

Bonferroni corrections to pairwise comparisons (Rice 1989). We also estimated Tajima’s D

(Tajima 1989) and Fu’s FS (Fu 1997) statistics to identify genetic signatures characteristic of either recent demographic change or selective sweeps.

We conducted a spatial analysis of molecular variance (SAMOVA) to define population groupings that were geographically homogeneous and maximally differentiated (Dupanloup et al. 2002). We ran SAMOVA with K= 2 to 6 groups to identify the number at which ΦCT was maximized. We tested for significant isolation by distance (IBD) with Mantel tests performed in

ARLEQUIN. A geographical distance matrix for all populations was created using The Geographic

Distance Matrix Generator (Ersts 2013). We used the SplitsTree4 phylogenetic program (Huson

& Bryant 2006, version 4.13.1) to build an unrooted haploptype network from our molecular sequence data (Figure 2.3).

Isolation with migration analysis

We estimated migration rates (m1: west to east, and m2: east to west), population divergence time (t), and genetic diversities (ΘE, ΘW, and ancestral ΘA) among eastern and western populations using IMa2 (Hey and Nielsen 2007; Hey 2010). This software uses an

isolation-with-migration model and Markov chain Monte Carlo (MCMC) sampling of gene genealogies to estimate the posterior densities of the divergence time, population size, and migration. From these values, we converted migration parameters m1 and m2 into the number of migrants per generation, Nm = (Θ m)/4. To convert estimates of model parameters into demographically meaningful units, we applied a divergence rate of 2.3% per million years

(corresponding to a substitution rate of 5.49 X 10-6 substitutions/site per year) to the mtDNA sequences, as estimated for arthropod mtDNA (Brower 1994). Because S. diana is univoltine, we used a generation time of one year. All IMa2 runs were submitted to the Palmetto computer cluster at Clemson University (http://citi.clemson.edu/palmetto/). Results are provided for the fully parameterized isolation with migration model, as removal of the non-significant migration parameters yielded almost identical results. We conducted the IMa2 analysis for all pairwise comparisons of populations and with eastern and western samples pooled.

Results

MtDNA Diversity

We identified 26 COII haplotypes among populations of S. diana (Table 2.1, Figs. 2.2 &

2.3). The most common haplotype, H2, was found in all populations, with the exception of OK and the extirpated IN population, which were fixed for haplotypes H1 and H7, respectively

(Table 2.1, Fig. 2.3). The H7 haplotype was common only in the three extirpated populations,

VA-2, IN, and OH. The population from SC was the only extant population to contain the H7 haplotype (9%).

Overall, haplotype diversity (h) was greatest in SC (h = 0.78, N=11), followed by AR-2

(h = 0.66, N=18), AR-1 (h = 0.64, N=10), and TN-1 (h = 0.63, N=28) (Table 2.2). The two

remaining western populations, AR-3 (h = 0.15, N=13) and OK (h = 0.0, N=5), displayed lower diversity along with two other monomorphic populations, IN (h = 0.0, N=5) and TN-3 (h = 0.0,

N=4) (Table 2). Both Fu’s FS and Tajima’s D were significantly negative in TN-1 (FS = -7.78, D

= -1.92), TN-2 (FS = -1.49, D = -1.75), and VA-1 (FS = -3.22, D = -1.44) (Table 2). The extirpated eastern VA-2 (FS = -0.88, D = -0.71) and western AR-2 (FS = -1.82, D = -0.93) populations also had significant negative values of Fu’s FS, and negative, although not significant, values for Tajima’s D (Table 2.2).

Genetic structure

Pairwise ΦST values between eastern and western populations were generally large and significant, ranging from 0.12 to 1.0 with a mean of 0.54 (Table 2.3). Eighty-eight percent of east-west comparisons were significant at P < 0.05; 70% remained significant with a Bonferroni correction (Table 2.3). Within each region, pairwise ΦST estimates were smaller, with an average of 0.27 in the east and 0.08 in the west. However, a small number of pairwise ΦST values between geographically close eastern and western populations were as small, or smaller, than estimates among eastern and western comparisons (Table 2.3).

Among the extirpated eastern populations, none of the pairwise ΦST comparisons with

VA-2 were significant at P < 0.05; all pairwise comparisons with IN were significant, but only the IN-TN3 comparison was significant following Bonferroni correction. Similarly, 78% of pairwise ΦST comparisons involving OH were significant, whereas 56% remained significant following Bonferroni correction.

Overlap in values of ΦST between east-west and within-region comparisons was apparent in the SAMOVA analysis, which did not group extant eastern and western samples into two

separate groups (Table 2.3). When no extinct populations were included in the analysis, ΦCT was maximized at K=2, with three of four western samples in one group and with one western sample

(AR1) grouped with all of the eastern samples (Table 2.4). Inclusion of extinct populations in the SAMOVA emphasized their distinctiveness; ΦCT was also maximized at K = 2 by separation of the extinct IN population from all other extinct and extant populations. Larger values of K (2-

4) split the remaining extinct from extant populations (Table 2.5).

Isolation by distance was significant for all pairwise comparisons (R2 =0.4480, P =

0.002) (Figure 2.4), but the relationship between geographical distance and genetic differentiation was driven primarily by differences among eastern and western groups (Figure

2.4). Geographical distance and genetic differentiation were not significantly correlated within either the east (R2 = 0.0002, P = 0.734) or the west (R2 = 0.0012, P = 0.639).

Isolation with Migration Analysis

Pooled East vs. West

Pooling all extant samples into one eastern and one western group yielded an IMa2-based east-west splitting time of 76,685 years (95% HPD range: 22,404-359,016 years) (Table 2.5,

Figure 2.5). The value of ΘA was 0.01 (95% HPD range: 0-13.7), several orders of magnitude smaller than estimates of both ΘW = 66.72 (95% HPD range: 27.04-205.6 individuals) and eastern ΘE = 2.91 (0.45-9.25) (Table 2.5). Migration rate estimates showed a large asymmetry after Bonferroni correction, with significant migration occurring east to west forward in time

(Table 2.5, Figure 2.6). We repeated this analysis with extinct populations included, and the results were qualitatively the same; therefore, they have been omitted here.

Pairwise comparisons, East vs. West

Most pairwise comparisons of extant eastern versus western populations followed the same pattern as the pooled comparison, with estimates of ΘE being much larger than ΘW. East to west migration (forward in time) was significant after Bonferroni correction for some pairwise comparisons involving westernmost eastern samples: TN1-AR1, TN1-AR3, and TN1-OK (Table

2.5, Figure 2.6). Pairwise divergence times between eastern and western samples ranged widely, from only 20,674 years between the geographically closest east-west population comparison

(TN3-AR1), to the divergence between OK-NC, dated at 133,000 years ago.

The pairwise comparisons involving extirpated populations from IN and OH resulted in the largest pairwise values of t, ranging from 102,095 years (OH-AR2) to 181,694 years (IN-

AR2). A divergence time of 103,370 years was estimated between VA2-OK, but the remaining east-west divergence times involving VA-2 were younger, ranging from 51,275 years (VA2-

AR2) to 86,794 years (VA2-AR3).

Pairwise comparisons, Eastern and Western

Geographically closer populations, among both eastern and western populations, yielded younger estimates of t than did pairwise comparisons between east and west (Table 2.5, Figure

2.7). Pairwise estimates of population divergence times among eastern comparisons ranged from

1,366 years (GA-TN3) to 88,069 years (NC-TN1), while western comparisons ranged from

13,388 years (AR1-AR2) to 40,710 years (AR2-OK) (Table 2.5). There was a large amount of overlap between the divergence time estimates for east and west, and within-region pairwise comparisons, with some east-west comparisons having estimates of t as small as some within- region estimates (Table 2.5).

The pairwise comparisons of divergence time involving the extirpated populations, IN and OH, were significantly older than extant eastern and western comparisons (Table 6) ranging from 108,834 years (TN1-OH) to 174,954 (GA-IN) years. The population comparisons involving the extirpated coastal VA population (VA-2) had a younger divergence time relative to those with IN and OH, ranging from 41,439 years (GA-VA2, NC-VA2) and 74,408 years (TN1-VA2).

Estimates of migration were not significant in either direction for any pairwise comparisons of eastern or western populations.

Discussion

Our phylogeographic study is the first to assess the population genetic structure of the threatened fritillary butterfly, S. diana, which has experienced a significant range collapse over the past century. We found highly significant genetic structure both within and among eastern and western regions, but without consistently strong east-west differentiation. Although much of the species’ range has collapsed in only the last 100 years (Wells and Tonkyn, 2013), pairwise

IMa2 comparisons showed that the geographically proximate eastern and western populations last shared a common ancestor much earlier, approximately 20,000 years ago. Without confidence intervals on this estimated divergence date, we cannot say it is significantly older than 100 years; however, none of the point estimates (peaks in posterior distributions) from the pairwise comparisons were within two orders of magnitude of 100 years, so we are confident the divergence date is much older than 100 years. In combination with the distribution of east-west pairwise ΦST values and their corresponding geography, our results suggest that gene flow between the remaining eastern and western populations was reduced long before the recent range collapse.

Effect of habitat fragmentation

Our analysis also revealed patterns of greater mtDNA diversity and more widespread population differentiation across the eastern distribution than the western distribution. Although we sampled fewer populations in the west, our sampling efforts spanned the entire latitudinal range of S. diana in both regions (Wells and Tonkyn 2013). Eastern populations displayed greater haplotype diversity within populations and greater population differentiation between populations than did populations of S. diana in the west.

The closely related, regal fritillary, S. idalia, has been characterized as a high gene-flow species, only revealing patterns of genetic differentiation over hundreds of kilometers of distance

(Williams 2001a, 2001b, 2002). This is not the case with S. diana, which displayed clear differentiation between populations at geographic distances less than 100 km within both the

East and the West. Recent observations of resource use by S. diana in the southern Appalachian

Mountains suggest that S. diana may be highly philopatric, which would limit dispersal out of habitat patches when high quality nectar resources are available (Wells and Smith 2013).

Consistent with these observations, all pairwise eastern and western estimates of 2Nm were < 1.

Pooled east-west migration rate estimates showed a large and significant asymmetry, with the equivalent of 6 migrants per generation (2Nm = 11.6) moving from east to west since the time of separation. The estimate of Nm ~ 6 indicates enough gene flow to preserve the potential for adaptive connectivity (Lowe and Allendorf 2010) between east and west but not enough (i.e.,

Nm >10) to result in drift connectivity, let alone demographically significant connectivity (Lowe and Allendorf 2010). Likewise, values of Nm in pairwise comparisons all fell well below this threshold for adaptive connectivity as well.

Range collapse

This range collapse is consistent with the predicted effects of climate warming, as S. diana has disappeared from lowland sites (IN, OH, VA-2) along the Atlantic Coastal Plain and interior

Midwest (Wells and Tonkyn 2013). If so, this would be one of the first documented cases of a species’ range collapse due to climate change in the southeastern US. Regardless of the cause of this collapse, these three lowland populations contained high frequencies of the H7 haplotype

(IN, 100%; OH, 50%; VA-2, 20%), which was never detected in the extant, higher elevation populations. Perhaps the lowland populations did not move upslope, but rather, have disappeared completely. This scenario is consistent with the species’ inferred low dispersal abilities and suggests that it is particularly vulnerable.

The loss of the Ohio River Valley populations appears to be a significant event in establishing the geographic separation of eastern populations from those in the west. The

SAMOVA did not cleanly split east and west into two distinct groups, however, but first clearly separated out the IN and OH populations, which were extirpated from the Ohio River Valley by the early 1900s (Shuey et al. 1987, Opler and Malikul 1992). This result was consistent with ΦST values, which were significant and large between almost all pairwise comparisons involving the

IN and OH populations.

We found significant evidence of recent demographic expansion in the TN-1, TN-3, and

VA-1 populations, which were significantly negative for Tajima’s D and Fu’s Fs. These three populations represent collection sites with the highest elevations in the eastern distribution (mean elevation = 1166 m). The only western population to show similar signs of recent demographic expansion was AR-2, a population located atop Mount Magazine, the highest point in Arkansas

(839 m). These results are consistent with the distributional change evident in museum

collections, that S. diana is shifting to higher elevations at the rapid rate of 18 m per decade

(Wells and Tonkyn 2013). The high genetic diversity in the southern Appalachian populations further supports the expansion of this species into predominately high-elevation habitat.

Whether this is the result of spatial or demographic expansion remains unknown. For example, the highest elevation populations might have been present for a long time, with individuals being too rare to be collected, but have only recently undergone a demographic expansion. This demographic expansion possibly occurred at the end of the Pleistocene, however, without further analysis these events are difficult to interpret.

Conservation implications

Potential management strategies to counteract the effects of habitat reduction and fragmentation include the establishment of wildlife corridors or stepping-stones to facilitate gene flow, and large reserves to increase effective population size (Haddad 2000; Frankham et al.

2002). Both elements will be important for preserving species with low dispersal abilities, like S. diana. Within these corridors and reserves, it will be important to promote the production of both larval food plants (Viola spp.), as well as abundant nectar plants for adult butterflies. Regular fire, and thinning, regimes will likely help maintain suitable habitat for the butterfly and other species in southeastern US forests, including the federally endangered red-cockaded woodpecker

(Picoides borealis Vieillot 1807) (Thill et al. 2004; Rudolph et al. 2006; Campbell et al. 2007).

The use of historical specimens from extinct populations, combined with extant populations, can provide insights into the relationship between past and present population structure, contributing to the development of appropriate long-term management strategies

(Keyghobadi et al. 2012; Münzbergová et al. 2013). The inclusion of genetic data from S. diana

populations that were extirpated in the early twentieth century revealed that most pairwise comparisons of divergence time involving the extirpated populations were significantly older than extant eastern and western comparisons. The mean age of separation for extinct populations was also significantly greater than extant populations. Although our sample size of extinct populations is small, we suggest that the ancient isolation of these S. diana populations may have rendered them more vulnerable to extinction and that population isolation time may be a demographic factor to consider in management plans, along with population size and migration rates. Our results highlight the value of incorporating genetic data from preserved specimens when investigating the phylogeographic history and conservation status of a threatened species.

Literature Cited

Avise JC (2000) Phylogeography: The History and Formation of Species. Harvard University Press, Cambridge, Massachusetts.

Broquet T, Angelone S, Jaquiery J, Joly P, Lena JP, Lengagne T, Plenet S, Luquet E, Perrin N (2010) Genetic bottlenecks driven by population disconnection. Conservation Biology 24:1596- 1605.

Brower AVZ (1994) Rapid morphological radiation and convergence among races of the butterfly Heliconius erato inferred from patterns of mitochondrial DNA evolution. Proceedings of the National Academy of Sciences USA 91:6491–6495.

Cech R, Tudor G (2005) Butterflies of the east coast. Princeton University Press, Princeton, NJ. Dupanloup I, Schneider S, Excoffier L (2002). A simulated annealing approach to define the genetic structure of populations. Molecular Ecology 11:2571-2581.

Ersts PJ [Internet] Geographic Distance Matrix Generator (version 1.2.3). American Museum of Natural History, Center for Biodiversity and Conservation. Available from http://biodiversityinformatics.amnh.org/open_source/gdmg. Accessed on 2013-10-15.

Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: A new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology. Research 10:564- 567.

Frankham R, Ballou JD, Briscoe DA (2002) Introduction to conservation genetics. Cambridge University Press, Cambridge, UK.

Fu Y-X (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics 147:915-925.

Haddad, N. (2000) Corridor length and patch colonization by a butterfly, Junonia coenia. Conservation Biology 14:738-745.

Hajibabaei M, Smith M, Janzen DH, Rodriguez JJ, Whitfield JB, Hebert PD (2006) A minimalist barcode can identify a specimen whose DNA is degraded. Molecular Ecology Notes 6:959-964.

Hammond PC, McCorkle DV (1983) The decline and extinction of Speyeria populations resulting from human environmental disturbances (Nymphaldae: Argynninae). The Journal of Research on the Lepidoptera 22:217-224.

Hey J (2010) Isolation with migration models for more than two populations. Molecular Biology Evolution 27: 905–920.

Hey J, Nielsen R (2007) Integration within the Felsenstein equation for improved Markov chain Monte Carlo methods in population genetics. Proceedings of the National Academy of Science 104:2785-2790.

Huson DH, Bryant D (2006). Application of Phylogenetic Networks in Evolutionary Studies. Molecular Biology Evolution 23:254-267.

Keyghobadi N, Koscinski D, Weintraub JD, Fonseca DM (2012) Historical specimens reveal past relationships and current conservation status of populations in a declining species: the regal fritillary butterfly. Insect Conservation and Diversity 6:234-242.

Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA, McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, Thompson JD, Gibson TJ, Higgins DG (2007) ClustalW and ClustalX version 2. Bioinformatics 23:2947-2948.

Lowe WH, Allendorf W (2010) What can genetics tell us about population connectivity? Molecular Ecology 19:3038-3051.

Moran M, Baldridge C (2002) Distribution of the Diana fritillary, Speyeria diana (Nymphalidae) in Arkansas, with notes on nectar plant and habitat preference. Journal of the Lepidopists’ Society 56:162-165.

Münzbergová Z, Cousins SAO, Herben T, Plačková I, Mildén M, Ehrlén J (2013) Historical habitat connectivity affects current genetic structure in a grassland species. Plant Biology 15:195–202.

Opler P, Krizek G (1984) Butterflies east of the Great Plains. Johns Hopkins University Press, Baltimore, MD.

Opler P, Malikul V (1992) A Field Guide to Eastern Butterflies. Houghton Mifflin Company, New York City, New York.

Posada D (2006) ModelTest Server: a web-based tool for the statistical selection of models of nucleotide substitution online. Nucleic Acids Research 34:W700-W703.

Rice WR (1989). Analyzing tables of statistical tests. Evolution 43:223-225.

Shuey JA, Calhoun JV, Iftner DC (1987) Butterflies that are endangered, threatened, and of special concern in Ohio. The Ohio Journal of Science 87:98–106.

Tajima F (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics 125:585-593.

Thill RE, Rudolph DC, Koerth NE (2004) Shortleaf pine-bluestem restoration for Red-cockaded woodpeckers in the Ouachita Mountains: Implications for other taxa. Pp. 657-671 In R. Costa & S. J. Daniels (eds.) Red-cockaded Woodpecker: road to recovery. Hancock House Publishers, Blaine, Washington. 744 pp.

Wells CN, Smith EA (2013) Observations of resource use by the threatened Diana Fritillary butterfly (Speyeria diana) in the Southern Appalachian Mountains, USA. Journal of Insects doi:10.1155/2013/130694

Wells CN, Tonkyn DW (2013) Range collapse in the Diana fritillary, Speyeria diana (Nymphalidae), Insect Conservation and Diversity DOI: 10.1111/icad.12059

Williams BE (1999) Regal fritillaries in a tailspin: a story of east and west, DNA, and the urgent need for conservation of a flagship species. American Butterflies 7:16–25

Williams BE (2001a) Patterns of morphological variation in Speyeria idalia (Lepidoptera: Nymphalidae) with implications for taxonomy and conservation. Annual Entomological Society of America 94:239-243

Williams BE (2001b) Recognition of western populations of Speyeria idalia (Nymphalidae) as a new subspecies. Journal of the Lepidopterists’ Society 55:144 -149

Williams BE (2002) Conservation genetics, extinction, and taxonomic status: a case history of the regal fritillary. Conservation Biology 16:148-157

Williams BE, Brawn JD, Page KN (2003) Landscape scale genetic effects of habitat fragmentation on a high gene flow species: Speyeria idalia (Nymphalidae). Molecular Ecology 12:11-2

Table 2.1 Mitochondrial DNA population diversity indices for S. diana populations from fresh and museum specimens. Haplotypes in bold are private to their respective population. N = number of individuals sequenced, P = number of haplotypes, † = extirpated pops. Population County/Region N P COII haplotypes (count) Source of mtDNA

Georgia (GA) Fannin/ Blue Ridge Mountains 13 2 H2 (10), H6 (3) Fresh 2006-2008 North Carolina (NC) Transylvania/ Blue Ridge Mountains 18 5 H2 (12), H9, H10 (3), H11 (2) Fresh 2006-2009 H12 (2) South Carolina (SC) Greenville/ Blue Ridge Escarpment 11 3 H2 (5), H6 (2), H7, H11 (3) Fresh 2006-2009 Tennessee (TN-1) Carter/Appalachian Mountains 28 10 H1, H2 (17), H12, H15, H16, H17, Fresh 2006-2009 H18 (2), H19, H20, H21, H22 Tennessee (TN-2) Sevier/ Great Smoky Mountains 12 4 H2 (9), H13, H14, H15 Fresh 2007-2009 Tennessee (TN-3) Sullivan/ Appalachian Mountains 4 1 H2 (4) Fresh 2007 Virginia (VA-1) Montgomery/ Appalachian Mountains 21 6 H2 (15), H3 (3), H4, H23, H25, H26 Fresh 2007-2008 Virginia (VA-2)† James City/ Coastal lowlands 5 2 H2 (2), H7, H22, H24 BNHM 1900, AMNH 1920 Indiana (IN) † Vanderburgh/ Ohio River Valley 5 1 H7 (5) USNMNH 1934, Field 1928 Ohio (OH) † Hamilton/ Ohio River Valley 4 2 H2, H7 (2), H8 CNHM 1930, ZMA 1911 Arkansas (AR-1) Benton, Carroll/ Ozark Plateau 10 2 H1 (4), H2 (5), H6 Fresh 2008, AMNH 1945, UALR 2006-2008 Arkansas (AR-2) Logan/ Mt. Magazine 18 5 H1 (10), H2 (6), H3, H5 Fresh 2006-2008 Arkansas (AR-3) Polk/ Ouachita Mountains 13 2 H1 (12), H2 Fresh 2008-2009, UALR 2006-2008 Oklahoma (OK) Latimer/ Ouachita Mountains 5 1 H1 (5) Fresh 2009 Museum specimens noted in italics: BNHM= British Natural History Museum, AMNH = American Museum of Natural History, USNMNH= United States National Museum of Natural History (Smithsonian), Field= Field Museum, CMNM= Carnegie Museum of Natural History, ZMA= Zoölogisch Museum Amsterdam, UALR= University of Arkansas, Little Rock

Table 2.2 Mitochondrial DNA diversity indices for S. diana populations, including nucleotide (π) and haplotype (h) diversities. Neutrality statistics Tajima’s D, and Fu’s FS are bold when significant at α = 0.05. Extirpated populations are marked with † and western populations are highlighted in grey.

Population π h D Fs GA 0.0007 0.3846 0.4256 0.6891 NC 0.0026 0.5490 -0.5909 -0.3507 SC 0.0015 0.7091 0.8505 -0.3227 TN-1 0.0016 0.6349 -1.9186 -7.7870 TN-2 0.0012 0.4545 -1.7468 -1.4886 TN-3 0.0000 0.0000 ------VA-1 0.0013 0.4952 -1.4443 -3.2156 VA-2† 0.0006 0.3333 -0.7099 -0.8873 IN† 0.0000 0.0000 ------OH† 0.0018 0.500 -0.7099 1.0986 AR-1 0.0013 0.6444 0.2217 -0.0465 AR-2 0.0014 0.6667 -0.9259 -1.8224 AR-3 0.0002 0.1538 -1.1491 -0.5371 OK 0.0000 0.0000 ------

Table 2.3 Population pairwise ΦST (Weir & Cockerman 1984) values among S. diana populations are above diagonal; Associated P- values are shown below diagonal. Extirpated populations are marked with a †. East-west comparisons are highlighted in grey. GA NC SC TN-1 TN-2 TN-3 VA-1 VA-2† IN† OH† AR-1 AR-2 AR-3 OK GA * 0.126 0.039 0.048 0.070 0.009 0.080 0.158 0.875* 0.683 0.327 0.424 0.766 0.778

NC 0.063 * 0.131 0.101 0.086 -0.023 0.117 0.051 0.634 0.440 0.213 0.329 0.512 0.465

SC 0.234 0.018 * 0.081 0.125 0.048 0.145 0.007 0.768 0.554 0.288 0.395 0.665 0.629

TN-1 0.045 0.009 0.009 * 0.005 -0.123 0.021 0.022 0.710 0.527 0.180 0.307 0.521 0.521

TN-2 0.072 0.036 0.009 0.324 * -0.128 0.024 0.039 0.800 0.582 0.239 0.364 0.679 0.662

TN-3 0.486 0.432 0.396 0.991 0.991 * 0.103 0.000 1.000 0.667 0.221 0.338 0.871 1.000

VA-1 0.063 0.000 0.000 0.090 0.189 0.991 * 0.031 0.765 0.580 0.119 0.247 0.537 0.528

VA-2† 0.189 0.207 0.162 0.252 0.450 0.991 0.324 * 0.826 0.500 0.351 0.229 0.741 0.706

IN† 0.000 0.000 0.000 0.000 0.000 0.036 0.000 0.000 * 0.063 0.785 0.807 0.961 1.000

OH† 0.000 0.045 0.000 0.000 0.009 0.162 0.000 0.162 0.000 * 0.635 0.603 0.857 0.826

AR-1 0.009 0.018 0.000 0.000 0.009 0.207 0.063 0.000 0.207 0.000 * -0.018 0.270 0.257

AR-2 0.000 0.000 0.000 0.000 0.000 0.063 0.009 0.081 0.000 0.000 0.556 * 0.068 0.054

AR-3 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.054 0.180 * -0.095

OK 0.009 0.000 0.000 0.000 0.000 0.018 0.000 0.000 0.000 0.018 0.099 0.234 0.991 *

Table 2.4 Speyeria diana population structure based on 549 bp mtDNA COII differentiation in spatial analysis of molecular variance (SAMOVA). Groupings K=2-6 above the line include three extirpated populations (†) in the analysis, while those below the line do not. Western populations are highlighted in grey.

K ΦCT ΦSC Grouping

2 0.616 0.736 [IN†] [OH†, VA2†, AR1, AR2, AR3, OK, VA1, SC, GA, NC, TN1, TN2, TN3] 3 0.579 0.696 [IN†] [OH†] [VA2†, AR1, AR2, AR3, OK, VA1, SC, GA, NC, TN1, TN2, TN3] 4 0.259 0.644 [IN†] [OH†] [OK] [AR1, AR2, AR3, VA1, VA2†, SC, GA, NC, TN1, TN2, TN3] 5 0.474 0.532 [IN†] [OH†] [OK] [AR1, AR2], [AR3, VA1, VA2†, SC, GA, NC, TN1, TN2, TN3] 6 0.457 0.499 [IN†] [OH†] [OK] [AR1, AR2] [AR3] [VA1, VA2†, SC, GA, NC, TN1, TN2, TN3]

2 0.378 0.441 [AR1, AR2, OK] [AR3, SC, NC, GA, TN1, TN2, TN3, VA1] 3 0.362 0.403 [AR1, AR2] [AR3, OK] [SC, NC, GA, TN1, TN2, TN3, VA1] 4 0.346 0.396 [AR1, AR2] [AR3] [OK] [SC, NC, GA, TN1, TN2, TN3, VA1] 5 0.339 0.338 [AR3, OK], [AR1, AR2] [SC] [NC] [GA, TN1, TN2, TN3, VA1] 6 0.333 0.332 [AR3, OK], [AR1] [AR2] [SC] [NC] [GA, TN1, TN2, TN3, VA1]

Table 2.5 Estimates of current (ΘE = Eastern and ΘW = Western) and ancestral (ΘA) theta, migration (m1 and m2), and divergence time (t) in years, from IMA2.0. The 95% highest posterior density (HPD) intervals are given for the pooled analysis only, where posteriors rose from and dropped down to zero probability. Extirpated populations = †, and significance at α = 0.05 is indicated with a *.

Comparison ΘE ΘW ΘA m1 Pm1 LLRm1 m2 Pm2 LLRm2 t East vs. West, Pooled 2.91 66.72 0.01 0 0.79 0 0.34 0.95 3.78* 76,685 (0.45-9.25) (27.04-205.6) (0-13.7) (22,404-359,016) East vs. West, Pairwise GA-AR1 0.71 2.65 0.89 0 1.39 0 0 0.93 0 56,557

GA-AR2 0.57 9.41 1.17 0 1.57 0 0 0.74 0 52,914

GA-AR3 0.67 0.55 0.11 0 1.67 0 0.26 0.71 0.44 124,863

GA-OK 0.73 0.05 0.43 0 1.81 0 0 1.39 0 107,923

NC-AR1 9.69 1.89 4.17 0 1.32 0 0 1.31 0 68,397

NC-AR2 8.23 4.21 3.91 0 2.08 0 0 1.72 0 133,242

NC-AR3 8.85 0.35 4.57 0 1.98 0 0 0.87 0 74,772

NC-OK 8.07 0.01 4.93 0 2.37 0 0 1.33 0 119,035 , SC-AR1 2.01 2.57 0.29 0 1.39 0 0 1.26 0 84,973

SC-AR2 1.35 5.79 1.17 0 1.70 0 0 1.32 0 84,608

SC-AR3 1.93 0.55 0.03 0 1.76 0 0.09 0.82 0.22 103,734

SC-OK 19.99 2.83 0.09 0 2.17 0 0 1.35 0 83,880

TN1-AR1 19.99 0.83 0.09 0 1.35 0 0.44 0.60 3.18* 86,612

TN1-AR2 19.99 4.61 0.03 0 1.77 0 0.26 0.96 2.38 89,526

TN1-AR3 19.99 0.55 0.07 0 0.93 0 0.24 0.96 3.06* 88,251

TN1-OK 19.99 0.01 0.11 0 1.34 0.11 0.21 1.13 3.51* 77,505

TN2-AR1 19.99 2.83 0.09 0 1.28 0 0 1.26 0 63,115

TN2-AR2 19.99 4.81 0.23 0 1.84 0 0 1.51 0 83,880

TN2-AR3 19.99 0.47 0.01 0 1.88 0 0 0.86 0.10 69,672

TN2-OK 19.99 0.01 0.03 0 2.18 0 0 1.35 0 83,880

TN3-AR1 0.01 2.91 1.29 0 1.10 0 0 0.87 0 20,674

TN3-AR2 0.05 19.99 1.45 0 1.19 0 0 0.63 0 25,956

TN3-AR3 0.03 0.57 0.69 0 1.24 0 0.29 0.66 0.50 90,437

TN3-OK 0.03 0.01 0.01 0 1.30 0 0 1.33 0 113,570

VA1-AR1 17.11 4.05 1.57 0 0.85 0 0 1.02 0 32,878

VA1-AR2 6.33 6.95 1.89 0 0.78 0 0 0.69 0 37,067

VA1-AR3 8.25 0.77 1.17 0.16 0.87 0.09 0.17 0.74 0.12 48,725

VA1-OK 7.17 0.09 0.01 0.17 0.93 0.16 0 1.37 0 65,118

VA2†-AR1 19.99 2.65 0.03 0 1.09 0 0 1.24 0 67,851

VA2†-AR2 19.99 5.87 1.37 0 1.28 0 0 1.31 0 51,275

VA2†-AR3 19.99 0.55 0.05 0 1.35 0 0 0.86 0 86,794

VA2†-OK 19.99 0.01 0.01 0 1.52 0 0 1.36 0 103,370

IN†-AR1 0.03 1.91 3.45 0 1.53 0 0 2.32 0 175,501

IN†-AR2 0.01 4.59 3.05 0 1.48 0 0.46 3.11 0 181,694

IN†-AR3 0.03 0.69 0.01 0 1.59 0 0 1.90 0 173,315

IN†-OK 0.03 0.01 6.37 0 1.12 0 0 0.76 0 123,588

OH†-AR1 1.97 1.91 3.19 0 0.92 0 0 1.58 0 136,885

OH†-AR2 1.63 4.79 2.63 0 1.13 0 0 1.85 0 102,095

OH†-AR3 2.37 0.49 3.81 0 1.06 0 0 1.09 0 146,903

OH†-OK 2.92 0.03 4.45 0 1.29 0 0 1.36 0 152,914

Eastern, Pairwise GA-NC 0.81 12.73 4.07 0 1.31 0 0 0.81 0 51,093

GA-SC 0.74 11.73 5.64 0 0.49 0 0 1.19 0 32,514

GA-TN1 1.39 19.99 0.01 0 1.26 0 0 0.99 0 72,404

GA-TN2 1.11 19.99 0.03 0 1.18 0 0 0.87 0 44,353

GA-TN3 1.17 0.01 0.01 0 0.98 0 0 0.95 0 1,366

GA-VA1 1.13 19.99 0.03 0 1.20 0 0 0.98 0 47,814

GA-VA2† 2.69 19.99 0.01 0 1.12 0 0 1.04 0 41,439

GA-IN† 0.73 0.01 3.17 0 2.06 0 0 1.57 0 174,954

GA-OH† 0.71 1.91 2.83 0 1.57 0 0.15 0.72 0.07 137,250

NC-SC 8.93 1.85 4.77 0.41 0.50 0.62 0 1.19 0 32,514

NC-TN1 12.49 19.99 0.01 0 0.84 0 0.04 0.69 0.01 88,069

NC-TN2 13.39 19.99 3.07 0 1.04 0 0 1.16 0 50,364

NC-TN3 16.29 0.03 4.15 0 0.73 0 0 1.07 0 21,038

NC-VA1 13.65 13.71 0.01 0 1.11 0 0 1.41 0 58,561

NC-VA2† 12.07 19.99 3.93 0 1.12 0 0 1.04 0 41,439

NC-IN† 7.23 0.01 6.07 0 3.27 0 0 1.43 0 154,189

NC-OH† 7.95 1.71 5.43 0 1.74 0 0 1.01 0 123,588

SC-TN1 3.07 19.99 0.01 0 1.21 0 0.43 0.57 1.53 74,954

SC-TN2 2.67 19.99 0.01 0 1.23 0 0 1.18 0 56,740

SC-TN3 2.57 0.01 0.07 0 0.98 0 0 1.05 0 35,064

SC-VA1 2.63 19.99 0.03 0 1.25 0 0 1.42 0 63,479

SC-VA2† 2.69 19.99 0.01 0 1.23 0 0 1.03 0 69,854

SC-IN† 1.85 0.01 3.21 0 2.36 0 0 1.54 0 180,055

SC-OH† 1.85 2.17 0.01 0 1.63 0 0 0.89 0 139,253

TN1-TN2 19.99 19.99 0.05 0 0.98 0 0 1.33 0 62,387

TN1-TN3 19.99 16.22 0.01 0 0.94 0 0 1.13 0 60,383

TN1-VA1 19.99 19.99 0.01 0.54 0.49 1.07 0 1.14 0 61,840

TN1-VA2† 19.99 19.99 0.01 0 0.87 0 0 1.04 0 74,408

TN1-IN† 19.99 0.03 0.01 0 4.77 0 0 1.44 0 113,570

TN1-OH† 19.99 2.51 0.19 0 1.74 0 0 1.12 0 108,834

TN2-TN3 19.99 0.01 0.09 0 0.77 0 0 1.05 0 33,424

TN2-VA1 19.99 19.99 0.01 0 1.05 0 0 1.15 0 54,736

TN2-VA2† 19.99 19.99 0.09 0 0.99 0 0 0.98 0 58,015

TN2-IN† 7.81 0.03 0.01 0 2.88 0 0 1.46 0 146,175

TN2-OH† 9.13 2.03 0.01 0 1.56 0 0 0.99 0 79,508

VA1-VA2† 16.33 15.33 0.07 0.26 0.56 0.24 0 0.80 0 63,115

VA1-IN† 9.13 0.03 0.03 0 3.77 0 0 1.47 0 158,197

VA2†-IN† 4.73 0.01 2.95 0 1.83 0 0 1.43 0 176,047

VA2†-OH† 19.99 2.07 2.59 0 1.22 0 0 0.96 0 98,270

IN†-OH† 0.03 17.43 0.01 0 1.11 0 0 0.76 0 23,588 Western, Pairwise

AR1-AR2 4.61 19.92 1.51 0 0.94 0 0 0.82 0 16,849

AR1-AR3 4.63 19.99 1.47 0 0.87 0 0 0.77 0 13,388

AR1-OK 2.71 0.01 0.89 0 0.92 0 0 1.09 0 24,135

AR2-AR3 19.99 1.35 1.09 0 0.87 0 0 0.68 0 13,388

AR2-OK 9.67 0.03 0.05 0 0.93 0 0 1.10 0 40,710

AR3-OK 19.99 1.61 1.55 0 0.99 0 0 1.39 0 40,710

Figure 2.1 Distributional data for Speyeria diana specimens documented in Wells and Tonkyn (2013) showing the two geographically separated population groups: the southern Appalachian Mountains in the east, and the Ozark and Ouachita Mountains in the west. Open circles represent specimens collected from 1777 to 1960 (N = 881); Black circles represent S. diana specimens collected from 1961 to 2010 (N = 2,517) (See Wells and Tonkyn 2013 for a complete description of these data).

Figure 2.2 Collection sites for Speyeria diana. Eastern samples were taken from Georgia (GA), South Carolina (SC), Tennessee-1 (TN-1), Tennessee-2 (TN-2), Tennessee-3 (TN-3), North Carolina (NC), and Virginia-1 (VA-1); Western samples were taken from Arkansas-1 (AR-1), Arkansas-2 (AR-2), Arkansas-3 (AR-3), and Oklahoma (OK); Museum samples were collected to represent extirpated populations (indicated with †) from Virginia-2 (VA-2), Indiana (IN), and Ohio (OH). See Table 1 for sources of all museum samples. The size of each pie graph represents the sample size of each population. The size of each slice is proportional to the number of individuals represented by each haplotype.

Figure 2.3 Unrooted SplitsTree4 (V 4.13.1) (Huson & Bryant 2006) haplotype network for 549 bp of COII. The circle size in each network is proportional to the number of individuals of each haplotype sample size, with the smallest circle representing one haplotype copy. See Figure 2.2 for the haplotype legend.

Figure 2.4 Plot of ΦST values (based on COII sequence data) versus geographic distance among all collection sites indicating evidence for isolation by distance.

Figure 2.5 Posterior probability distributions for east-west divergence times (t) and for rates of migration (m) for pooled samples of S. diana.

Figure 2.6 Estimates of ΘE (eastern), ΘW (western), and ancestral (ΘA) theta from IMA2.0 for pooled samples of S. diana.

Figure 2.7 Geographic distribution of divergence time (t) in years for eastern, western and east vs. west comparison

CHAPTER THREE

EFFECTS OF CLIMATE CHANGE ON THE FUTURE GEOGRAPHIC DISTRIBUTION OF THE DIANA FRITILLARY IN SOUTHEASTERN US

Abstract

Climate change is predicted to alter the geographic distribution of a wide variety of taxa, including butterfly species. Research has focused primarily on high latitude species in North America, with no known studies examining responses of taxa in the southeastern US. To evaluate how climate change might influence the geographic distribution of the Diana fritillary (Speyeria diana), we developed species distribution models using Maxent. We used two global circulation models, CCSM and MIROC, under two emissions scenarios to predict the future distribution of S. diana. The Maxent models were evaluated using Receiver Operating Characteristics Area Under Curve test

(AUC ranged from 0.86-0.96), which revealed that the models produced were significantly better than random (0.5). All four climate scenarios resulted in an average loss of 91% of suitable habitat by 2050. Habitat in the southern Appalachian Mountains is predicted to suffer the most severe fragmentation and reduction in suitability. These results suggest that the Diana fritillary is under threat of decline, and is most likely going to be negatively affected by future global warming; however, other factors will likely contribute to changes in the range of S. diana besides climate alone.

Introduction

In 2013, the average surface temperature for the contiguous United States was

0.6°C warmer than in any other year since national records were first kept (Blunden &

Arndt 2012). Hundreds of species in the US and elsewhere have responded to the warming climate by shifting to higher latitudes or elevations (Parmesan 1996; Parmesan

& Yohe 2003; Thomas et al. 2004; Crozier & Dwyer 2006). Such range shifts have been documented in a number of disparate taxa (Walther et al. 2002; Root et al. 2003;

Parmesan 2006), including alpine plants (Walther et al. 2005), marine invertebrates

(Cheung et al. 2009), marine fishes (Perry et al. 2005), mosquitoes (Epstein et al. 1998), birds (Thomas & Lennon 1999; Hitch & Leberg 2007), and butterflies (Parmesan 1996;

Parmesan et al. 1999; Wilson et al. 2005; Wilson et al. 2007; Asher et al. 2011; Wilson &

Maclean 2011). Species that do not make such shifts, because of a lack of dispersal ability or routes or other reasons, will need to adapt to changing climate in situ or may disappear (Wilson et al. 2005, 2007).

Understanding how species distributions might shift with changing climate is a critical component of managing and protecting future biodiversity. In recent years, a number of models have been developed to predict the impacts of climate change on species distributions. Species distribution models (SDM) can take many forms, but the so-called bioclimate envelope models (BCM) identify the climatic bounds within which a species currently occurs, and projects where those climatic bounds will be found in the future, under various climate projections (Pearson & Dawson 2003; Peterson 2003; IUCN

2007a, b; Settele et al. 2008, Elith & Leathwick 2009; Fordham et al. 2012).

There are two general methods used for modeling species distributions: ones that use presence–absence data, and ones that use only presence data (Pearson & Dawson

2003; Peterson et al. 2004). When we can document a species’ presence and absence from various sites, we can directly compare the environmental conditions between the

two groups of sites to infer the conditions needed for occurrence. These cases are limited, however, especially for highly mobile species (Dettmers & Bart 1999). More commonly, we only have records of species’ occurrences to work with, and then we must use indirect methods to infer their climatic requirements (Thuiller et al. 2005b; Walther et al. 2005;

Willis et al. 2007). Maxent is among the best performing of this second type of model, as determined by direct competitions, and it performs well even with low sample sizes typical of rare species (Elith et al. 2011; Weber 2011; Garcia et al. 2013). Maxent compares data from occurrence sites with those from a random sample of sites from the larger landscape, perhaps set by the overall species range, and minimizes the relative entropy of statistical models fit to each data set. Species distribution models have been criticized as overly simplistic, as they do not incorporate external factors such as inter- species interactions (Pearson & Dawson 2003; Araújo & Luoto 2007; Elith et al. 2011).

However, they are still able to project with reasonable accuracy whether species ranges will increase or decrease under a changing climate (Araújo et al. 2005a; Green et al.

2008), which is the primary objective of this study.

Here, we predict the effects of climate change on the distribution of a threatened butterfly species endemic to the southeastern US, the Diana fritillary (Speyeria diana).

This species has undergone a recent range collapse, resulting in an 800 km geographic disjunction across the center of its distribution between western populations in the

Ouachita and Ozark Mountains and eastern ones in the southern Appalachian Mountains.

It has disappeared from lowland sites in between as well as along the Atlantic coastal plain (Wells and Tonkyn 2013). We used the maximum entropy approach to model the distribution of S. diana under several future climatic scenarios, to forecast how it might

shift under future conditions. Forecasts of large range reductions, or small overlap between current and future ranges, would suggest high vulnerability to climate change.

Methods

Study species

The Diana fritillary, Speyeria diana, is a large and sexually dimorphic nymphalid butterfly, endemic to the southeastern US. Adult males emerge in late May to early June, with females flying several weeks to a month later (Opler & Krizek 1984). Once mated, each female can lay thousands of eggs singly on ground litter during the months of

August and September in the vicinity of Viola spp., the larval host plant for all Speyeria

(Allen 1997). After hatching, first instar larvae immediately burrow deep into the leaf litter layer of the forest floor, where they overwinter (Cech & Tudor 2005). In spring, larvae feed on the foliage of freshly emerging violets. Adult Diana butterflies are often found along forest edges or dirt roads containing tall, conspicuous nectar sources such as milkweeds, butterfly bushes or other large summer and fall composites (Ross 2003; Cech

& Tudor 2005; Baltosser 2007; Ross 2008). While males begin to die off in late July, females persist in large numbers, although somewhat cryptically, through October

(Adams & Finkelstein 2006).

Distribution data

We searched for all known records of S. diana, from publications, catalogued and uncatalogued specimens in public and private collections in the United States and

Europe, online databases, contemporary field surveys by scientists and amateurs, and our

own field surveys (see Wells and Tonkyn 2013 for details). We obtained distributional data from 1,323 pinned S. diana specimens from 38 natural history museum collections in the United States and Europe. Four hundred thirty-five additional records (1938–2012) were provided by the Butterfly and Moth Information Network and the participants who contribute to its BAMONA project. Our literature survey produced 153 records (1818–

2011) across 54 US counties. We also collected 419 S. diana butterflies in our own field surveys. This is an essentially complete dataset of all records for the species and may be as comprehensive as for any taxon in the region. For this reason, it should be especially informative in creating an accurate bioclimate envelope for the species.

Species distributional modeling

We developed ecological niche models using maximum entropy, or Maxent

(Phillips et al. 2006), a machine-learning algorithm for ecological modeling using presence-only records. Maxent works by finding the maximum entropy present within a spatial dataset in relation to a set of background environmental variables. Locality data and randomly sampled background points are combined with climatic data to predict the probability of the species’ occurrence within each raster grid cell (Elith et al. 2006,

2010). We used environmental climate data from WorldClim (Hijmans et al. 2005) at 30 arc-second resolution or approximately 1 km2 grid cells. Bioclimate variables and elevation layers were each clipped to the extent of North America using ESRI ArcMap

10.0, and data extracted to S. diana sample localities. Additionally, we collected locality data for three other species of North American butterflies (Speyeria cybele, Speyeria idalia, Battus philenor), which served as 5,628 random background points for our

models.

Climatic variables included 19 derived bioclimatic variables that describe annual and seasonal variation in temperature and precipitation averaged for 1950-2000, as well as elevation (Table 3.1). One concern when modeling species distributions is the strong correlation that occurs between multiple climate variables, which can significantly influence model predictions of species distributions (Beaumont et al. 2005). To test for collinearity, we performed spatial autocorrelation statistics between all pairs of the 19 bioclimate variables using ESRI ArcMap 10.0. We then selected the biologically most meaningful variable for each group of two or more variables with Pearson correlation coefficients higher than 0.7 (Table 3.1). This allowed us to reduce the number of bioclimate variables to just 9: Minimum Temperature of Coldest Month, Mean

Temperature of Driest Quarter, Precipitation of Wettest Month, Precipitation of Driest

Month, Precipitation of Driest Quarter, Isothermality, Mean Diurnal Range (Mean of monthly (maximum temperature - minimum temperature)), Temperature Annual Range,

Annual Precipitation, and elevation (Table 3.1). These variables are similar to those typically considered to be important determinants of the distributions of other butterfly species, such as mean temperature of the coldest month (related to the overwintering survival of first instar larvae), growing degree days above 5°C (regarded as a surrogate for the developmental threshold of the larvae), water balance (which corresponds to the moisture availability for the larval host and adult nectar plants), and the mean temperature of late summer (ensures proper adult emergence and mating) (Hill et al.

2003; Peterson et al. 2004; Mitikka et al. 2008; Filz et al. 2013; Zinetti et al. 2013).

Another concern when modeling species distributions is whether the occurrence

records are spatially biased with respect to site accessibility (e.g., towns, roads, trails)

(Kadmon et al. 2004; Louiselle et al. 2007). To address this concern, we applied a spatial filter to remove all sampling points that were within 5 km of each other using ESRI

ArcMap 10.0. The spatial filter resulted in 254 unique presence points for S. diana that were used in the final model. We first modeled the distribution of these 254 occurrences in present-day climate, and then projected the fitted species distribution under two future climate scenarios for the period 2040-2069 (hereafter referred to as 2050).

Our future climate scenarios were taken from two global circulation models

(GCMs) obtained from www.worldclim.org: the Community Climate System Model

(CCSM) (Gent et al. 2011) and the Model for Interdisciplinary Research on Climate

(MIROC) (Hasumi and Emori 2004, Nozawa et al. 2005). These GCMs differ in the reconstruction of several climatic variables and are well-known to produce different outcomes for butterfly species (Otto-Bliesner et al. 2007; Zinetti et al. 2013). For example, in hind-casting Mediterranean butterflies, the CCSM model projects narrower distributions at the last glacial maximum than does MIROC (Habel et al. 2010; Habel et al. 2013). For each of these two GCMs, we considered two different representative concentration pathways (RCPs) (Moss et al. 2010), which are cumulative measures of human emissions of greenhouse gases from all sources expressed in Watts per square meter (van Vuuren et al. 2011). These pathways were developed for the Fifth Assessment

Report of the Intergovernmental Panel on Climate Change and correspond to a total anthropogenic radiative forcing of RCP = 4.5 W/m2 (low) and RCP = 8.5 W/m2 (high)

(Moss et al. 2008; Riahi et al. 2011; Thomson et al. 2011).

We used Maxent default parameters (Phillips et al. 2006, 2009; Phillips & Dudík

2008) and a ten-fold cross-validation approach to further reduce bias with respect to locality data. This method divides presence data into ten equal partitions, with nine used to train the model, and the tenth used to test it. These partitions generate ten maps (one map per run), with each raster grid cell containing a value representing the probability of occurrence (Dubey et al. 2013). These values are used to designate habitat suitability ranging from 0 (unsuitable habitat) to 1 (highly suitable habitat). We averaged the resulting maps for the current climate, and for the two GCMs under RCP = 4.5 and RCP

= 8.5. This method resulted in the production of a “low” and “high” average prediction for S. diana species distribution in 2050, represented with habitat suitability maps. We measured the goodness of fit for the models using the area under the curve (AUC) of a receiver–operating characteristic (ROC) plot (Fielding & Bell, 1997). We used criteria of

Swetts (1998) and considered AUC values higher than 0.7 representative of model predictions significantly better than random values of 0.5 or less (Elith 2002, Phillips et al. 2006; Elith et al. 2011). The relative importance of each variable’s contribution was assessed by sequential variable removal by Jackknife (Phillips et al. 2006).

Results

The Maxent modelling resulted in “excellent” model fits for Speyeria diana

(Swets 1998), with an average AUC score of 0.91 for RCP = 4.5 and 0.87 for RCP = 8.5

(Table 3.1). The distribution of S. diana was largely determined by annual precipitation under both RCPs (17.9%, RCP = 4.5; 19.4%, RCP = 8.5). Among the remaining bioclimatic variables, mean temperature of driest quarter had the next highest average percent contribution (10.3%, RCP = 4.5; 25.0%, RCP = 8.5), followed by minimum

temperature of coldest month (20.1%, RCP = 4.5; 10.4%, RCP = 8.5), isothermality

(7.3%, RCP = 4.5; 7.6%, RCP = 8.5), precipitation of wettest month (3.5%, RCP = 4.5;

3.9%, RCP = 8.5), precipitation of driest month (1.4%, RCP = 4.5; 5.4%, RCP = 8.5), precipitation of driest quarter (3.3%, RCP = 4.5; 2.4%, RCP = 8.5), Elev (1.5%, RCP =

4.5; 3.5%, RCP = 8.5), mean diurnal range (1.8%, RCP = 4.5; 2.8%, RCP = 8.5), and temperature annual range (1.6%, RCP = 4.5; 1.3%, RCP = 8.5) (Table 3.1).

Modelling with Maxent under the selected climate-change scenarios predicted that habitat suitability will decrease for S. diana by 2050 (two-tailed paired t-tests comparing current Maxent values with those of 2050; all P<0.01). The MIROC model resulted in more loss of suitable habitat than CCSM under both RCP scenarios (88.2% versus 92.4% of suitable habitat retained for RCP 4.5, and 90.2% versus 94.3% of suitable habitat retained for RCP 8.5 in CCSM and MIROC, respectively). Both climate models indicate that the loss of core distributional area is modest, with an average of 91.3% of present distributional areas retained. The most drastic reduction in habitat is apparent across the southern Appalachian Mountains (Fig. 3.2).

Discussion

The Diana fritillary is considered a threatened species, as it has undergone a dramatic range collapse across the center of its distribution over the past century (Wells and Tonkyn 2013). Modelling the present-day distribution of S. diana suggests that while suitable habitat occurs across the entire landscape where this species is currently found, it becomes patchy and slightly less favorable toward the center of the range. Both CCSM and MIROC climate-change models predict severe habitat loss on the eastern edge of the

species’ range by 2050, which is of serious concern. According to a recent population genetics study of S. diana (Chapter 2), the southern Appalachian Mountains are an important source of genetic diversity for the species.

The vulnerability of a species to climate change can be assessed by how much suitable habitat will remain in the future and whether it will be contiguous or fragmented, and near or far away (Bakkenes et al. 2002, Settele et al. 2008). Both climate models predict severe fragmentation of suitable habitat in the southern Appalachian Mountains where the species presently occurs, and expansion in higher latitude areas on the eastern coast of the US (Fig. 3.2). The southern edge of highly suitable habitat in S. diana’s western distribution is predicted to recede by 2050, with habitat only persisting in the higher elevations of the Ozark and Ouachita mountains of Arkansas. While the models did not predict a severe decrease in overall habitat by 2050, the fragmentation and loss of suitable habitat in the southern Appalachians suggests that S. diana will likely decline in that region.

In mountainous ecosystems, the high elevation species are generally predicted to be affected the most by warming climates (Grabherr et al., 1994; Beniston, 2000; Holten,

2001; Skov & Svenning, 2004, Ruiz-Labourdette et al. 2012). This is a concern for the

Diana fritillary, which has already disappeared from lowland sites and is shifting to higher elevations at a rate of 18m per decade, suggestive of response to global warming

(Wells & Tonkyn 2013). However, Wells & Tonkyn (2013) describe in detail other key factors besides climate that are also likely to influence the geographic distribution of this species, such as deforestation, deer herbivory, pesticide use, and fire regimes.

These results should be interpreted with some caution, as we acknowledge that

100

our model is limited in a number of ways. First, future climate predictions for SDM are most often derived from general circulation models; however, the applicability of SDMs at regional scales has been questioned due to their coarse resolution (Beaumont et al.

2008). Therefore, we are presently developing a random forest prediction of habitat suitability for S. diana using R 2.3.1 statistical environment (R Core Development Team,

2005) for comparison with our Maxent SDM. Random forest is a modelling technique based on classification (and regression) tree analysis (Breiman 2001) that is capable of capturing non-linear relationships in predictor variables having complex interactions

(Schrag et al. 2007). In addition, we are testing a variety of GCM predictions rather than basing final conclusions on CCSM and MIROC models alone.

Literature Cited

Adams JK, Finkelstein I (2006) Late season observations on female Diana fritillary (Speyeria diana) aggregating behavior. News of the Lepidopterists’ Society, 48:106-107.

Allen J (1997) The butterflies of West Virginia and their caterpillars. University of Pittsburgh Press, Pittsburgh, PA.

Araújo MB, Luoto M (2007) The importance of biotic interactions for modeling species distributions under climate change. Global Ecology and Biogeograpy 16:743-753.

Araújo MB, Pearson RG, Thuiller W, Erhard M (2005a) Validation of species-climate impact models under climate change. Global Change Biology 11:1504-1513.

Araújo MB, Whittaker RJ, Ladle RJ, Erhard M (2005b) Reducing uncertainty in projections of extinction risk from climate change. Global Ecology and Biogeography 14:529-538.

Asher J, Fox R, Warren MS (2011) British butterfly distributions and the 2010 target. Journal of Insect Conservation 15:291-299.

Bakkenes M, Alkemade JRM, Ihle F, Leemans R, Latour JB (2002) Assessing effects of forecasted climate change on the diversity and distribution of European higher plants for 2050. Global Change Biology 8:390-407.

101

Baltosser W (2007) Flitting with disaster: humans and habitat are keys to our state butterfly’s future. Arkansas Wildlife 38:6-11.

Beaumont L, Hughes L, Poulsen M (2005) Predicting species distributions: use of climatic parameters in BIOCLIM and its impact on predictions of species’ current and future distributions. Ecological Modelling 186:250-269.

Beaumont LJ, Hughes L, Pitman AJ (2008). Why is the choice of future climate scenarios for species distribution modelling important? Ecology letters 11:1135-1146.

Blunden J, Arndt DS (2012) State of the Climate in 2011. Bulletin of the American Meteorology Society 93:S1-S264.

Breiman L (2001) Random forests. Machine learning 45:5-32.

Cech R, Tudor G (2005) Butterflies of the east coast. Princeton University Press, Prinecton, NJ.

Cheung WWL, Lanm WYV, Sarmiento JL, Kearney K, Watson R, Pauly D (2009) Projecting global marine biodiversity impacts under climate change scenarios. Fish and Fisheries 10:235-251.

Crozier L, Dwyer G (2006) Combining population-dynamic and ecophysiological models to predict climate-induced insect range shifts. The American Naturalist 167:853-866.

Dettmers R, Bart J (1999). A GIS modeling method applied to predicting forest songbird habitat. Ecological Applications 9:152-163.

Dubey S, Pike DA, Shine R (2013) Predicting the impacts of climate change on genetic diversity in an endangered lizard species. Climatic Change 117:319-327.

Elith J (2002) Quantitative methods for modeling species habitat: comparative performance and an application to Australian plants. In: Ferson S, Burgman M, editors. Quantitative methods for conservation biology. New York: Springer-Verlag. 39– 58.

Elith J, Graham CH, NCEAS Species Distribution Modelling Group (2006) Novel methods improve prediction of species’ distributions from occurrence data. Ecography 29:129-151.

Elith, J, Leathwick JR (2009) Species distribution models: ecological explanation and prediction across space and time. Annual Review of Ecology, Evolution, and Systematics 40:677–697.

102

Elith J, Kearney M, Phillips S. (2010) The art of modeling rang-shifting species. Methods Ecol Evol, 1: 330-342.

Elith J, Phillips SJ, Hastie T, Dudík M, Chee YE, Yates CJ (2011) A statistical explanation of MaxEnt for ecologists. Diversity and Distributions 17:43-57.

Epstein PR, Diaz H, Elias FS, Grabherr G, Graham NE, Martens WJM, Mosley- Thompson E, Susskind EJ (1998) Biological and physical signs of climate change: focus on mosquito-borne disease. Bulletin of the American Meteorological Society 78:409– 417.

Fielding AH, Bell JF (1997) A review of methods for the assessment of prediction errors in conservation presence/absence models. Environmental Conservation, 24:38–49.

Filz KJ, Schmitt T, Engler JO. (2013) How fine is fine-scale? Questioning the use of fine- scale bioclimatic data in species distribution models lused for forecasting abundance patterns in butterflies. European Journal of Entomology 110:311-317.

Fordham DA, Resit Akçakaya H, Araújo MB, Elith J, Keith DA, Pearson R, Auld TD, Mellin C, Morgan JW, Regan TJ, Tozer M, Watts MJ, White M, Wintle BA, Yates C, Brook BW (2012) Plant extinction risk under climate change: are forecast range shifts alone a good indicator of species vulnerability to global warming? Global Change Biology 18:1357–1371.

Garcia K, Lasco R, Ines A, Lyon B, Pulhin F (2013) Predicting geographic distribution and habitat suitability due to climate change of selected threatened forest tree species in the Philippines. Applied Geography 44:12-22.

Gent PR, Danabasoglu G, Donner LJ, Holland MM, Hunke EC, Jayne SR, Lawrence DM, Neale RB, Rasch PJ, Vertenstein M, Worley PH, Yang ZL, Zhang M (2011) The Community Climate Systme Model Version 4. Journal of Climate 24:4973-4991.

Green RE, Collingham YC, Willis SG, Gregory RD, Smith KW, Huntley B (2008) Performance of climate envelope models in retrodicting recent changes in bird population size from observed climatic change. Biology Letters 4:599-602.

Habel JC, Rödder D, Scalercio S, Meyer M, Schmitt T (2010) Strong genetic cohesiveness between Italy and North Africa in four butterfly species. Biological Journal of the Linnean Society 99:818–830.

Habel JC, Husemann M, Schmitt T, Dapporto L, Vandewoestijne S (2013) A forest butterfly in Sahara desert oases: isolation does not matter. Journal of Heredity 104:234–247.

Hasumi H, Emori S. (2004) K-1 coupled GCM (MIROC) description, K-1 Tech. Rep. 1, 34pp., Center for Climate Systems Research, Univ. of Tokyo, Tokyo.

103

Hijmans RJ, Cameron SE, Parra JL, Jones PG, Jarvis A (2005) Very high resolution interpolated climate surfaces for global land areas. International Journal of Climatology 25:1965-1978.

Hill JK, Thomas CD, Fox R, Telfer MG, Willis SG, Asher J, Huntley B (2002) Responses of butterflies to twentieth century climate warming: implications for future ranges. Proceedings of the Royal Society of London, Series B-Biological Sciences 269:2163–2171.

Hill JK, Thomas CD, Huntley B (2003). Modelling present and potential future ranges of European butterflies using climate response surfaces. In: Butterflies: Ecology and Evolution Taking Flight (editors Boggs, C.L., Watt, W.B. & Ehrlich, P.R.). University of Chicago Press, Chicago, pp. 149–167.

Hitch AT, Leberg PL (2007) Breeding distributions of North American bird species moving north as a result of climate change. Conservation Biology 21:534–539.

IPCC (2007a) Climate Change 2007- The Physical Science Basis. Cambridge University Press, Cambridge.

IPCC (2007b) Climate Change 2007- Impacts, Adaptation and Vulnerability. Cambridge University Press, Cambridge.

Kadmon R, Farbr O, Danin A (2004) Effect of roadside bias on the accuracy of predictive maps produced by bioclimatic models. Ecological Applications 14:401-413.

Loiselle BA, Jørgensen PM, Consiglio T, Jiménez I, Blake JG, Lohmann LG, Montiel OM (2007) Predicting species distributions from herbarium collections: does climate bias in collection sampling influence model outcomes? Journal of Biogeography. 35:105-116

Mitikka V, Heikkinen RK, Luoto M, Araújo MB, Saarinen K, Pöyry J, Fronzek S (2008) Predicting range expansion of the map butterfly in Northern Europe using bioclimatic models. Biodiversity and Conservation 17:623-641.

Moss RH, Babiker M, Brinkman S, Calvo E, Carter T, Edmonds J, Elgizouli I, Emori S, Erda L, Hibbard K, Jones R, Kainuma M, Kelleher J, Lamarque JF, Manning M, Matthews B, Meehl J, Meyer L, Mitchell J, Nakicenovic N, O’Neill B, Pichs R, Riahi K, Rose S, Runci P, Stouffer R, van Vuuren D, Weyant J, Wilbanks T, van Ypersele JP, Zurek M (2008) Towards New Scenarios for Analysis of Emissions, Climate Change, Impacts, and Response Strategies. Intergovernmental Panel on Climate Change, Geneva, 132 pp.

Moss RH, Edmonds JA, Hibbard KA, Manning MR, Rose SK, van Vuuren DP, Carter TR, Emori S, Kainuma M, Kram T, Meehl GA, 2, Mitchell JFB, Nakicenovic N Riahi K, Smith SJ, Stouffer RJ, Thomson AM, Weyant JP, Wilbanks TJ

104

(2010) The next generation of scenarios for climate change research and assessment. Nature 463:747–756.

Nozawa T, Nagashima T, Shiogama H, Crooks S (2005) Detecting natural influence on surface air temperature change in the early twentieth century. Geophysical Research Letters 32: doi:10.1029/2005GL023540

Opler PA, Krizek G (1984) Butterflies east of the Great Plains. Johns Hopkins University Press, Baltimore, MD.

Otto‐Bliesner BL, Hewitt CD, Marchitto TM, Brady E, Ouchi A, Crucifix M, Weber SL (2007) Last Glacial Maximum ocean thermohaline circulation: PMIP2 model intercomparisons and data constraints. Geophysical Research Letters 34:L12706.

Parmesan C (1996) Climate and species’ ranges. Nature 382:765-766

Parmesan C (2006) Ecological and evolutionary responses to recent climate change. Annual Review of Ecology, Evolution, and Systematics 37:637–69

Parmesan C, Ryrholm N, Stefanescu C, Hill J, Thomas C, Descimon H, Huntley B, Kaila L, Kullberg J, Tammaru T, Tennent, Thomas J, Warren M (1999) Poleward shifts in geographical ranges of butterfly species associated with regional warming. Nature 399:579-583.

Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across natural systems. Nature 421:37-42.

Pearson RG, Dawson TP (2003) Predicting the impacts of climate change on the distribution of species: are bioclimate envelope models useful? Global Ecology and Biogeography 12:361-371.

Peterson AT, Martínez-Meyer E, González-Salazar C, Hall PW (2004) Modeled climate change effects on distributions of Canadian butterfly species. Canadian Journal of Zoology 82:851–858.

Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate change and distribution shifts in marine fishes. Science 308:1912-1915.

Peterson AT (2003) Projected climate change effects on Rocky Mountain and Great Plain birds: generalities on biodiversity consequences. Global Change Biology 9:647–55.

Phillips SJ, Anderson R, Schapire RE (2006) Maximum entropy modeling of species geographic distributions. Ecological Modelling 190:231–259.

105

Phillips SJ, Dudík M, Elith J, Graham CH, Lehmann A, Leathwick J, Ferrier S (2009) Sample selection bias and presence-only distribution models: implications for background and pseudo-absence data. Ecological Applications 19:181-197.

Phillips SJ, Dudík M (2008) Modeling of species distributions with Maxent: new extensions and a comprehensive evaluation. Ecography 31:161-175.

Riahi K, Rao S, Krey V, Cho C, Chirkov V, Fischer G, Kindermann G, Nakicenovic N, Rafaj P (2011) RCP 8.5—A scenario of comparatively high greenhouse gas emissions. Climatic Change 109:33-57.

Root TL, Price JT, Hall KR, Schneider SH, Rosenzweig C, Pounds JA (2003) Fingerprints of global warming on wild animals and plants. Nature 421:57–60

Ross GN (2003) What’s for dinner? A new look at the role of phytochemicals in butterfly diets. News of the Lepidopterists’ Society 45:83-89.

Ross GN (2008) Diana’s Mountain Retreat. Natural History 72:24-28.

Ruiz-Labourdette D, Nogués-Bravo D, Ollero HS, Schmitz MF, Pineda FD (2012) Forest composition in Mediterranean mountains is projected to shift along the entire elevational gradient under climate change. Journal of Biogeography 39:162-176.

Schrag A, Bun A, Graumlich L (2007) Influence of bioclimatic variables on treeline conifer distribution in the Greater Yellowstone Ecosystem: implications for species of conservation concern. Journal of Biogeography doi.10.1111/j.1365-2699.2007.018015.x

Settele J, Kudrna O, Harpke A, Kühn I, van Swaay C, Verovnik R, Warren M, Wiemers M, Hanspach J, Hickler T, Kühn E, van Halder I, Veling K, Vliegenthart A, Wynhoff I, Schweiger O (2008) Climatic risk atlas of European butterflies. BioRisk 1:1–710.

Swets JA (1988) Measureing the accuracy of diagnostic systems. Science 240: 1285- 1293.

Thomas CD, Lennon JJ (1999) Birds extend their ranges northwards. Nature 399:213.

Thomas CD, Franco AM, Hill JK (2004) Range retractions and extinction in the face of climate warming. TRENDS in Ecology and Evolution 21:415-416.

Thomson A, Calvin K, Smith S, Kyle P, Volke A, Patel P, Delgado-Arias S, Bond- Lamberty B, Wise M, Clarke L, Edmonds J (2011) RCP4.5: a pathway for stabilization of radiative forcing by 2100. Climatic Change 109:77-94.

Thuiller W, Lavorel S, Araújo MB (2005a) Niche properties and geographical extent as predictors of species sensitivity to climate change. Global Ecology and Biogeography, 14, 347–357.

106

Thuiller W, Lavorel S, Araújo MB, Sykes MT, Prentice IC (2005b) Climate change threats to plant diversity in Europe. Proceedings of the National Academy of Sciences 102:8245–8250. van Vuuren DP, Edmonds J, Kainuma M, Riahi K, Thomson A, Hibbard K, Hurtt GC, Kram T, Krey V, Lamarque JF, Masui T, Meinshausen M, Nakicenovic N, Smith S, Rose S (2011) The Representative Concentration Pathways: An Overview. Climatic Change 109:5-31.

Walther GR, Post E, Convey P, Menzel A, Parmesan C, Beebee T, Fromentin JM, Hoegh-Guldberg O, Bairlein F (2002) Ecological Responses to recent climate change. Nature 416:389-395.

Walther GR, Beibner S, Conradin A (2005) Trends in the upward shift of alpine plants. Journal of Vegetation Science 16:541-548.

Weber TC (2011) Maximum entropy modeling of mature hardwood forest distribution in four US states. Forest Ecology and Management 261:779-788.

Wells CN, Tonkyn DW (2013). Range collapse in the Diana fritillary, Speyeria diana (Nymphalidae) Insect Conservation and Diversity doi: 10.1111/icad.12059

Willis KJ, Araújo MB, Bennett KD, Figueroa-Range B, Froyd CA, Myers N (2007) How can a knowledge of the past help to conserve the future? Biodiversity conservation and the relevance of long-term ecological studies. Philosophical Transactions of the Royal Society of Britain 362:175–186.

Wilson RJ, Gutiérrez D, Gutiérrez J, Martinez D, Aguado R, Monserrat VJ (2005) Changes to the elevational limits and extent of species ranges associated with climate change. Ecology Letters 8:1138-1146.

Wilson RJ, Gutiérrez D, Gutiérrez J, Monserrat V (2007) An elevational shift in butterfly species richness and composition accompanying recent climate change. Global Change Biology 13:1873- 1887.

Wilson RJ, Maclean IMD (2011) Recent evidence for the climate threat to Lepidoptera and other insects. Journal of Insect Conservation 15:259–268.

Zinetti F, Dapporto L, Vovlas A, Chelazzi G, Bonelli S, Balletto E, Ciofi C (2013) When the rule becomes the exception. No evidence of gene flow between two Zerynthia cryptic gutterflies suggests the emergence of a new model group. PLOS ONE 8, e65746-e657

107

Table 3.1. The 19 bioclimate variables from the WorldClim dataset (Hijmans et al., 2005) collapsed into groups of highly correlated variables (Pearson’s correlation coefficient, r ≥ ± 0.70), and their corresponding contribution to the Maxent model. The ten variables kept in the final model are bold and highlighted in grey. CCCM and MIROC global circulation models are shown under representative concentration pathways (RCPs) 4.5 and 8.5, as predicted by the IPCC 5th report on climate. % Contribution Bioclimate variables Abbreviation CCCM-45 MIROC-45 AVG CCCM-85 MIROC-85 AVG Annual Mean Temperature Bio 1 4.4 0.7 2.5 0.5 1.4 0.96 Max Temperature of Warmest Month Bio 5 0.6 1.7 1.2 1.4 0.8 1.1 Min Temperature of Coldest Month Bio 6 3.9 36.3 20.1 2.6 3.3 10.4 Mean Temperature of Wettest Quarter Bio 8 14.1 10.2 12.2 4.0 16.8 2.6 Mean Temperature of Driest Quarter Bio 9 15.5 5.1 10.3 30.2 19.8 25.0 Mean Temperature of Warmest Quarter Bio 10 0.5 0.8 0.7 0.1 0.3 0.2 Mean Temperature of Coldest Quarter Bio 11 0.8 12.5 11.9 3.3 1.5 2.4 Precipitation of Wettest Month Bio 13 3.7 0.2 3.5 2.0 5.8 3.9 Precipitation Seasonality Bio 15 6.0 3.7 4.9 8.7 2.7 5.6 Precipitation of Wettest Quarter Bio 16 0.8 0.6 0.7 0.2 0.9 0.6 Precipitation of Warmest Quarter Bio 18 1.1 0.3 1.0 1.9 1.0 1.5 Precipitation of Driest Month Bio 14 0.9 1.6 1.4 2.7 8.0 5.4 Precipitation of Driest Quarter Bio 17 4.2 2.3 3.3 2.2 2.6 2.4 Precipitation of Coldest Quarter Bio 19 0.1 0.2 0.2 0.2 1.7 0.9 Altitude Elev 2.0 1.0 1.5 4.9 2.0 3.5 Isothermality (BIO2/BIO7) (* 100) Bio 3 11.0 3.5 7.3 8.5 6.6 7.6 Temperature Seasonality Bio 4 6.4 1.0 3.7 0.0 4.2 2.1 Mean Diurnal Range (Mean of monthly Bio 2 0.6 3.0 1.8 2.0 3.6 2.8 (max temp - min temp)) Temperature Annual Range (BIO5-BIO6) Bio 7 1.2 1.9 1.6 1.5 1.0 1.3 Annual Precipitation Bio 12 22.3 13.4 17.9 22.9 15.9 19.4 AUC 0.86 0.96 0.91 0.87 0.86 0.87

108

Figure 3.1 The present-day geographic distribution of Speyeria diana, with indices of habitat suitability as predicted by maximum entropy modelling (Maxent) under current climatic conditions (1950-2010).

109

Figure 3.2 (a) Habitat suitability indices for the projected future distribution of Speyeria diana under the CCMA and MIROC RCP 4.5 climate change scenarios; (b) Habitat suitability indices for the projected future distribution of Speyeria diana under the CCMA and MIROC RCP 8.5 climate change scenarios.

110

CHAPTER FOUR

CONCLUSIONS: WHAT DOES THE FUTURE HOLD FOR SPEYERIA DIANA?

Understanding the boundaries of species' ranges and the variations in population dynamics across a species' distribution is important, especially when the species is of conservation concern. While it is useful to document current distributional patterns, few researchers have considered historical data of past distributions for comparison with present ones. This is unfortunate considering the vast amount of historical information available on species occurrences in early literature, museum and herbarium collections.

Comparing historical species distributions to the present day species’ ranges may help shed light on what factors limit the distribution, allowing better estimates of future distributions.

The Diana fritillary, Speyeria diana, appears to be threatened by severe loss and fragmentation of suitable habitat, and impending climate change. Our comprehensive collection of locality data, assembled from numerous publications, museum specimens, online databases, and field surveys, comprises what we consider to be the entire distribution history of this butterfly. We analyzed these records for significant shifts in latitude, longitude, elevation, and phenology and found that the Diana fritillary has disappeared entirely from the Atlantic coastal plain, where the holotype was first described, as well as from interior lowland sites across the Ohio River Valley. The species now persists in two geographically separated parts of its former range with an 850 km disjunction separating them. We found that this species is shifting to higher elevations at the rapid rate of 18 m per decade and females are being collected 4.3 days earlier per decade, which is consistent with predicted responses to climate change. This is the first

111

study to show such responses in a southeastern US species.

Our population genetic study dates the earliest splitting of eastern and western populations at approximately 20,000 years ago, during the Last Glacial Maximum. This suggests that gene flow between east and west was severed long before the range collapse documented in this species over the past century. We detected the loss of a unique lowland haplotype, which was found to be common only in extirpated populations represented by museum specimens. The large amount of genetic variation and strong differentiation between geographically proximate populations in the southern

Appalachians suggests that dispersal in S. diana is somewhat low. It appears that the lowland populations did not just move to higher elevations, but rather, have disappeared entirely. These results highlight the value of incorporating genetic data from preserved specimens when investigating the phylogeographic history and conservation status of a threatened species. Similar information has been used to classify a disjunct population of the closely related Regal fritillary, S. idalia, as a distinct evolutionary lineage (Williams

2002). Mitochondrial DNA sequence divergence and our extensive natural-history dataset suggest that the disjunct population of S. diana in its western distribution may similarly represent a distinct evolutionary lineage as well. While subspecies designation is complicated and controversial, our data indicate that the western population grouping of

S. diana is clearly differentiated from those in the east.

Finally, in modeling the future distribution of S. diana, under several scenarios of climate change, we have found that the amount of suitable habitat available to S. diana is predicted to decline and become increasingly fragmented by 2050. This is especially evident in the southern Appalachian Mountains, where a majority of the species’ genetic

112

diversity is found. While this prediction is based only on one method of species distributional modeling (Maxent), both global circulation models (CCCM and MIROC) were closely aligned in their outcomes. Species distribution modeling is currently being repeated with our dataset using a random forest method to provide a model comparison.

The predicted loss of suitable habitat for this species combined with its rapid shift to higher elevation habitat suggest that it will become increasingly vulnerable to rapid decline as climate change proceeds.

In closing, this dissertation showcases the response of a rare southern

Appalachian butterfly species to long-term changes in climate and landscape. These results are especially important, as other similar studies have focused extensively on species at higher latitudes, or in Europe, leaving southeastern US species largely untested. Combining our comprehensive dataset for S. diana with a phylogeographic study has allowed us to better understand patterns of historical and contemporary movement of this species throughout its range. We have also clearly identified climatic changes that are likely to reduce suitable habitat for other southern Appalachian butterfly species with similar life histories with S. diana. Our predictions suggest that by the year

2050, widespread changes to the North American climate and landscape will reduce suitable habitat for the eastern populations of S. diana. The loss of these populations would likely result in the loss of the unique genetic variation that they harbor. We suggest that southern Appalachian populations be targeted for conservation of larval host-plants

(Viola spp.) and high quality adult nectar plants to maintain suitable habitat. Collecting of

S. diana, especially in the eastern US, should be predominately limited to scientific studies focused on preserving the species into the future.

113

Literature Cited

Williams BL (2002) Conservation genetics, extinction, and taxonomic status: a case history of the regal fritillary. Conservation Biology 16:148-157.

114

APPENDIX

Supplemental bibliography of all distributional literature sources, in order of appearance in Table 1.2.

Cramer, P., & Stoll, C. (1777) De uitlandsche Kapellen voorkomende in de drie Waereld-Deelen Asia, Africa en America, II. Deel. De plaataen XCVII-CXCII. (Plate 98, Pages 3-4).

Blatchley (1859) A catalogue of the butterflies known to occur in Indiana. Report of State Geologist.

Edwards, W. H. (1864) Description of the female Argynnis diana. Proceedings of the Entomological Society of Philadelphia, 3, 431.

Hulst, G. D. (1872) Argynnis Diana, Cram. Entomologica Americana. Vol I. April 1885- March 1886.

Edwards, W. H. (1874) Notes on the larvae of Argynnis cybele, aphrodite, and diana. The Canadian Entomologist , 6, 121-125.

Aaron, E. M. (1877) Annual meeting of the Entomological Society of Ontario. The Canadian Entomologist, 9, 199-200.

Strecker, H. (1878) Butterflies and moths of North America. Press of B. F. Owen. Reading, PA.

Thomas, C. T. (1878) Seventh report of the state entomologist on the noxious and beneficial insects of the state of Illinois. D. W. Lurk State Printer and Binder, Springfield, IL.

Fisher, S. D. (1881) Transactions of the Department of Agriculture State of Illinois with reports from county agricultural boards, for the year 1880. H.W. Rokker State Printer and Binder, Springfield, IL.

Holland, W. J. (1883) A mystery and its solution. The Canadian Entomologist. 15: 41-42.

Blatchley, W. S. (1886) Some southem Indiana butterflies. Hoosier Naturalist, Nov. & Dec.

French, G.H. (1886) The butterflies of the eastern United States. For the use of classes in zoology for private students. J. B. Lippincott Company, Philadelphia, PA.

Hine, J. S. (1887a) List of butterflies known to have been taken in Ohio. Ohio State Academy of Science 6th Annual Report: 22-27.

115

Hine, J. S. (1887b) Ohio butterflies. Journal of the Proceedings of the Columbus Horticultural Society, 12, 77-89.

Kingsley, J. S. (1888) The riverside natural history. Volume II. Crustaceans and insects. Kean Paul, Trench & Co., Paternoster Square, London, UK.

Scudder, S. (1889) The butterflies of the eastern United States and Canada, with special reference to New England. Published by the author, Cambridge. MA.

Skinner, H. & Aaron, E. M. (1889) A list of the butterflies of Philidelphia, PA. The Canadian Entomologist, 21, 126-131.

Dixey, F. A. (1890) On the phylogenetic significance of the wing markings in certain genera of the Nymphalidae In: The Entomologist, ed. R. South. West, Newman & Co., London, UK.

Blatchley, W. S. (1891) Catalogue of the butterflies known to occur in Indiana. Annual Report of Indiana State Geologists, 17, 365-408.

Skinner, H. (1896) Argynnis diana. Entomological News, 6, 318.

Holland, W. J. (1898) The butterfly book. Doubleday and McClure Co., New York, NY.

Snyder (1900) The argynnids of North America. Occasional Memoirs of the Chicago Entomological Society, 1, 27-38.

Strecker, H. (1900) Lepidoptera, rhopaloceres and heteroceres, indigenous and exotic. Supplement No. 3, Reading, Pennsylvania.

Maynard, C. J. (1901) A manual of North American butterflies. De Wolfe, Fiske & Co. Boston, MA.

Sell, R. A. (1916) Hunting butterflies in the Ozarks (Lepidoptera) Entomological News, 27, 55-56.

Smyth, E. A. (1916) Entomological News and Proceedings of the Entomological Section of the Academy of Natural Sciences of Philadelphia. Volume XXVII. The Academy of Natural Sciences, Philadelphia, PA.

Wood, W.M. (1916) Argynnis diana (Lepidoptera). Entmological News, 27, 35.

Murrill, W. A. (1919) The natural history of Staunton, Virginia. Bronxwood Park, New York, NY.

Holland, W. J. (1931) The butterfly book. New and thoroughly revised edition. Doubleday, Doran and Co., Garden City, New York.

116

Knobel, L. (1931) Argynnis diana Cr. As Observed about Hope, Arkansas. Bulletin of the Brooklyn Entomological Society , 26, 184.

Kite, V. (1934) A calendar of Ozark butterflies, Lake Taneycomo region, Missouri. Entomological News, 45, 36-39.

Clark, A.H. (1937) Surveying the butterflies of Virginia. The Scientific Monthly, 45, 256- 265.

Clark, A.H., & Williams, C.M. (1937) Records of Argynnis diana and of some other butterflies from Virginia. Journal of the Washington Academy of Sciences, 207, 209-213.

Allen, J. (1941) The butterflies of West Virginia and their caterpillars. University of Pittsburgh Press, Pittsburgh, PA.

Chermock, R. L. (1942) Notes on collecting Argynnis diana, Proceedings of the Pennsylvania Academy of Science, 16, 59-61.

Bock, T. (1949) Notes on collecting Speyeria diana. The Lepidopterists’ News, 3, 1-2.

Clark, A.H., & Clark, L.F. (1951) The Butterflies of Virginia, Smithsonian Miscellaneous Collections, 116, 237.

Klots, A.B. (1951) A field guide to the butterflies of North America, East of the Great Plains. Houghton Mifflin Co., Boston, MA.

Mather, B. & Mather, K (1958) The butterflies of Mississippi. Tulane Studies in Zoology, 6, 63-109.

Evans, W. H. (1959) The saga of an orphan Speyeria diana larva. Journal of the Lepidopterists’ Society, 13, 93-95.

Curtis, C. C., & Boscoe, W. F. (1962) Butterflies at light in North Carolina and Pennsylvania. Journal of the Lepidopterists’ Soceity, 16, 174.

Hovanitz, W. (1963) Geographical distribution and variation of the genus Argynnis. III. Argynnis diana. Journal of Research on the Lepidoptera, 1, 209-221.

Ross, G. N. & Lambremont, E. N. (1963) An annotated supplement to the state list of Louisiana butterflies and skippers. The Journal of the Lepidopterists’ Society, 17, 148- 158.

Masters, J. H. (1968) Euphydryas phaeton in the Ozarks (Lepidoptera: Nymphalidae). Entomological News, 79, 85-91.

117

Masters, J. H., & Masters, W. L. (1969) An annotated list of the butterflies of Perry County and a contribution to the knowledge of Lepidoptera in Indiana. Association of Minnesota Entomologists, 6, 1-25.

Shull, E. M., & Badger, F. S. (1971) Annotated list of the butterflies of Indiana. Journal of the Lepidopterists’ Society, 26, 13-24.

Harris, L. (1972) Diana. Speyeria diana Cramer. Pages 276-278 In: Butterflies of Georgia. University of Oklahoma Press. Norman, OK..

Irwin, R.R., & Downey. J.C. (1973) Annotated Checklist of the Butterflies of Illinois. Illinois Natural History Survey, Biological Notes No. 81.

Howe, W. H. (1975) The butterflies of North America. Doubleday & Company, Inc. Garden City, NY.

Nelson, J. M. (1979) A preliminary checklist of the skippers and butterflies of Oklahoma. Proceedings of the Oklahoma Academy of Science, 59, 41-46.

Schowalter, A. H. & Drees, B. M. (1980) Bilateral gynandromorphic Speyeria diana (Nymphalidae). Journal of the Lepidopterists’ Society, 34, 340-344.

Pyle, R. M. 1981. National Audubon Field Guide to Butterflies. Alfred Knopf, New York, NY.

Hammond, P. C. & McCorkle. D.V. (1983) The decline and extinction of Speyeria populaions resulting from human environmental disturbances (Nymphaldae: Argynninae). The Journal of Research on the Lepidoptera, 22, 217-224.

Opler, P. A. (1983) County atlas of eastern United States butterflies (1840-1982). U.S. Fish. Wildlife Service Division of Biological Services, Washington, D.C. 86 pp.

Opler, P. A. & Krizek, G. (1984) Butterflies east of the Great Plains. Johns Hopkins University Press, Baltimore, MD.

Scott, J. A. (1986) The butterflies of North America. A natural history and field guide. Stanford University Press, Stanford, CA.

Shuey, J.A., Calhoun, J.V. & Iftner, D.C. (1987) Butterflies that are endangered, threatened, and of special concern in Ohio. The Ohio Journal of Science, 87, 98-106.

Shull, E. M. (1987) The Butterflies of Indiana. Indiana University Press, Bloomington, IN.

Watson, C. N. & Hyatt, J. A. (1988) Butterflies of northeast Tennessee. Journal of the Lepidopterists’ Society, 42, 19-31.

118

Kohen, L. C. (1989) An aberrant male Speyeria diana (Nymphalidae) from Kentucky. The Kentucky Lepidopterist, 15, 2.

Cohen, E. & Cohen, J. (1991) A collecting adventure in the George Washington National Forest. News of the Lepidopterists Society, 1, 3-4.

Krizek, G.O. (1991) Is Speyeria diana attracted by manure? News of the Lepidopterists’ Society, 6, 83.

Adams, J. K. (1992) My first Diana female. The Kentucky Lepidopterist, 18, 11-12.

Opler, P., & Malikul, V. (1992) A field guide to eastern butterflies. Houghton Mifflin Company, NY, NY.

Skillman, F. W. and Heppner, J. B. (1992) Gynandromorph Speyeria diana from Georgia (Lepidoptera: Nymphalidae). Tropical Lepidoptera, 3, 34.

Carlton, C.E. & Nobles, L.S. (1996) Distribution of Speyeria diana in the highlands of Arkansas, Missouri and Oklahoma with comments on conservation. Entomological News, 107, 213-219.

Allen, J. (1997) The butterflies of West Virginia and their caterpillars. University of Pittsburgh Press, Pittsburgh, PA.

Ross, G. N. (1997) Preliminary inventory of the butterflies of Coweeta hydrologic laboratory, Nantahala National Forest, North Carolina. News of the Lepidopterists’ Society , 39, 70-71, 88.

Ross, G. N. (1998) Definitive Destination: Mount Magazine State Park, Arkansas. American Butterflies, 6, 24–33.

Glassberg, J. (1999) Butterflies through binoculars, the east. Oxford University Press, New York, New York.

Moran, M., & Baldridge, C. (2002) Distribution of the Diana Fritillary, Speyeria diana (Nymphalidae) in Arkansas, with notes on nectar plant and habitat preference. Journal of the Lepidopterists’ Society, 56, 162-165.

Daniels, J. C. (2003) Butterflies of the Carolinas. Adventure Publications, Inc. Cambridge, MN.

Scholtens, B. (2004) Survey for Speyeria diana in Sumter National Forest (Oconee Co., SC). Report to National Forest Service.

119

Cech, R., & Tudor, G. (2005) Butterflies of the East Coast: an observer’s guide. Princeton University Press, Princeton and Oxford.

Vaughn, D. & Shepherd, M. (2005) Species profile: Speyeria diana. In M.D. Shepherd, D.M.Vaughan., & S.H. Black, eds., Red list of pollinator insects of North America. The Xerces Society for Invertebrate Conservation , Portland, OR.

Adams, J.K. & Finkelstein, I. (2006) Late season observations on female Diana fritillary (Speyeria diana) aggregating behavior. News of the Lepidopterists’ Society, 48, 106-107.

Rudolph, D.C., Ely, A.C., Schefer, R.R., Williamson, J.H. & Thill, R.E. (2006) The Diana fritillary (Speyeria diana) and Great Spangled fritillary (S. cybele): Dependence on fire in the Ouachita Mountains of Arkansas. Journal of the Lepidopterists’ Society, 60, 218-226.

Spencer, L. A. (2010) Arkansas butterflies and moths. Ozark Society Foundation, Little Rock, AR.

Campbell, J.W., Hanula, J.L. & Waldrop, T.A. (2007) Observations of Speyeria diana (Diana fritillary) utilizing forested areas in North Carolina that have been mechanically thinned and burned. Southeastern Naturalist, 6, 179-182.

Ross, G.N. (2008) Diana’s Mountain Retreat. Natural History, 72, 24-28.

Wells, C.N., Spencer, L. & Simons, D. (2010) Notes on the breeding behavior of Speyeria diana (Nymphalidae) in Arkansas. Journal of the Lepidopterists’ Society, 65, 51-53.

Wells, C.N., Edwards, .L., Hawkins, R., Smith, L &Tonkyn, D. (2011). A rearing method for Argynnis (Speyeria) diana (Lepidoptera: Nymphalidae) that avoids larval diapause. Psyche, 2011, 1-6.

120