Proceedings of SDAS 1997

Home , Lytechinus pictus, Mike Sturgis, Ptychodus mortoni, Sturgeon chub

Proceedings of the South Dakota Academy of Science

Volume 76 1997

Published by the South Dakota Academy of Science Academy Founded November 22, 1915

Editor Kenneth F. Higgins

Terri Symens, Wildlife & Fisheries, SDSU provided secretarial assistance

Tom Holmlund, Graphic Designer

We thank former editor Emil Knapp for compiling the articles contained in this volume.

TABLE OF CONTENTS

Minutes of the Eighty-Second Annual Meeting of the South Dakota Academy of Science...... 1 Presidential Address: Can we live with our paradigms? Sharon A. Clay ...... 5

Complete Senior Research Papers presented at The 82nd Annual Meeting of the South Dakota Academy of Science Fishes of the Mainstem Cheyenne River in South Dakota. Douglas R. Hampton and Charles R. Berry, Jr...... 11 Impacts of the John Morrell Meat Packing Plant on Macroinvertebrates in the Big Sioux River in Sioux Falls, South Dakota. Craig N. Spencer, Gwen Warkenthien, Steven F. Lehtinen, Elizabeth A. Ring, and Cullen R. Robbins ...... 27 Winter Survival and Overwintering Behavior in South Dakota Oniscidea (Crustacea, Isopoda). Jonathan C. Wright ...... 45 Fluctuations in Daily Activity of Muskrates in Eastern South Dakota. Joel F. Lyons, Craig D. Kost, and Jonathan A. Jenks...... 57 Occurrence of Small, Nongame Mammals in South Dakota’s Eastern Border Counties, 1994-1995. Kenneth F. Higgins, Rex R. Johnson, Mark R. Dorhout, and William A. Meeks ...... 65 Use of a Mail Survey to Present Mammal Distributions in South Dakota. Carmen A. Blumberg, Jonathan A. Jenks, and Kenneth F. Higgins ...... 75 A Survey of Natural Resource Professionals Participating in Waterfowl Hunting in South Dakota. Jeffrey S. Gleason and Jonathan A. Jenks...... 91 The Importance of Conservation Reserve Program Fields to Breeding Grassland Birds at Buffalo Ridge, Minnesota. Krecia L. Leddy, David E. Naugle, and Kenneth F. Higgins...... 105 Effects of Wind Turbines on Nesting Raptors at Buffalo Ridge in Southwestern Minnesota. Robert E. Usgaard, David E. Naugle, Robert G. Osborn, and Kenneth F. Higgins ...... 113 The Origin of Waterfalls in the Black Hills, South Dakota. Charles Michael Ray and Perry H. Rahn...... 119 Stable Carbon Isotopes in Celtis (Ulmacaea) from the Beaver Creek Shelter, Wind Cave National Park, South Dakota and Paleoclimatic Interpretations. Mark L. Gabel, Marlina Cowan, and Larry L. Tieszen...... 131 Comparison of Implements for Improving Sodic Rangeland Soils. John Bechtold, F. R. Gartner, E. M. White...... 139 Atrazine and Alachlor Adsorption Characteristics to Benchmark Soil Series in eastern South Dakota. Z. Liu, S.A. Clay, J. Gaffney, and D. Malo...... 147 Analysis of Spatial Distribution of Canada Thistle (Cirsium arvense) in NoTill Soybean (Glycine max). B.L. Broulik, J. Lems, S.A. Clay, D.E. Clay, and M.M. Ellsbury ...... 159 Leafy Spurge—A review. Sharon A. Clay and Chad M. Scholes ...... 171 Stress Effects and Evolution of Direct Development in Echinoid Echinoderms. Leland G. Johnson and Teresa L. Fitzgerald-Brown...... 189 Comparing Estimates of Some Bivariate Survival Functions Under Random Censorship. Bonaventure A. Anthonio...... 199 Determination of Ozone Optical Depths from Shadowband Radiometer Data. Stephen D. Hawks and Stephen Schiller ...... 211 Strengthening and Deformation Mechanisms in the First Generation of AL-LI Alloys 2091 and 8090. Dajun Chen and Glen A. Stone...... 217 Illuminance is a Point Source is an Oscillating Function of Distance in a Hyperspherical Universe. Richard P. Menzel...... 225 Sources of Natural Resource Information, Non-Consumptive Use of Wildlife and Changes in South Dakota Residents’ Attitudes Toward Hunting Between 1973 and 1989. Nancy J. Dietz, Kenneth F. Higgins, and Robert D. Mendelsohn ...... 229 Student Opinions About Hunting in South Dakota. Robert D. Mendelsohn, Nancy J. Dietz, and Kenneth F. Higgins...... 239 Trace Metals in Water and Sediments of Wetlands in the Rainwater Basin Area of Nebraska. Christine C. Gordon, Lester D. Flake, and Kenneth F. Higgins ...... 253

Abstracts of Senior Research Papers presented at The 82nd Annual Meeting of the South Dakota Academy of Science Determining the Prevalance of Whirling Disease in Idaho Rivers. Bryan Ledgerwood, Dale Droge, and Steve Elle ...... 265 A Characterization of Water Chemistry and Plankton from Four Prairie Lakes. Theodore B. McMillan and Lois Haertel...... 267 Role of Wetlands in Agricultural Systems. D.H. Rickerl and L.L. Janssen ...... 269 Paleontological Resource Survey of State-Administered Lands in South Dakota. Bruce A. Schumacher, James E. Martin, and Curt Johnson ...... 271 Sedimentology of the Coarse-Grained Deposits from the Quaternary Fossil Lake Area in South-Central Oregon. Janet L. Bertog, James E. Martin, and Allen J. Kihm ...... 273 Fossil Invertebrates of the Niobrara Formation in South Dakota. James E. Martin, Bruce A. Schumacher, David C. Parris, and Barbara Smith Grandstaff ...... 275 Additional Quaternary Vertebrates from the Chamber of Lost Souls, Wind Cave National Park, Southwestern South Dakota. James E. Martin and Ruth Anderson...... 277 Preliminary Studies on Distribution of Phyllachora on Poaceae. A.C. Gabel, L.H. Tiffany, and M.L. Gabel...... 279 Studies on a South Dakota Population of Soybean Cyst Nematode, Heterodera glycines. James L. Jones and James D. Smolik ...... 281 Preliminary Examination of the Effect of Direction of Slope of Field on Growth and Yields of Corn Plants in Northwestern Iowa. Troy McKenney and Donna Hazelwood ...... 283 Black Dot Flea Beetle Genetic Diversity as Measured at Two Loci—MDH and PGM. Mark A. Brinkman, Sharon A. Clay, and Nels H. Granholm...... 285 Allelopathy in Echinacea Angustifolia D.C. Roots. Peter A. Jauert and R. Neil Reese ...... 287 Allelopathic Potential of Echinacea angustifolia D.C.’s Root Extracts. Kimberly Piechowski and R. Neil Reese...... 289 Thyroxine Effects on Ivertebrate Development. Leland G. Johnson, Jessica E. Kullbom, Stephanie L. Priesz, Shana N. Strand, and Katherine M. Faber ...... 291 Agouti Gene Regulation of Thiol (CYS amd GSH) Concentrations in Livers of Mice. J.T. Brunz, R. N. Reese, and N. H. Granholm...... 293 Des-Acetyl-Alpha Melanocyte Stimulating Hormone and Melanogenesis in B16 Melanoma Cells. P. Ramasastry and N.H. Granholm...... 295 Serum Estradiol Levels in the Lethal Yellow Mouse. Maureen Diggins, Dustin Dierks, Julia Spiry, and Nels Granholm...... 297 Molecular Characterization of the Bovine Homolog of the Agouti Gene. M.D. Johansen, R.R.R. Rowland, and N.H. Granholm ...... 299 Bovine Viral Diarrhea Virus Infects Camelid Cell Lines. A.A. Ali, A.A. Salama, T.A. Aboellail, and C.C. Chase ...... 301 Use of Digotonin to Differentially Strip the Tegumental Membrane Surrounding the Strobila and Bladder Regions of Taenia Taeniaformis Strobilocerci. E.J. Olson, D.J. Robison, S.R. Duimstra and M.B. Hildreth ...... 303 Mathematical Analysis of a Model for the Growth of a Single Species Cell: Ribosome Dependence Model. A.S. Elkhader ...... 305 Mathematical Models for the Visualization of Juggling. Richard D. Simmons and Douglas J. Peters ...... 307 Modiﬁcation of Carbon Electrode Surfaces. Royce C. Engstrom, Brian D. Lamp and Betsy B. Ratcliff ...... 309 Creating and Imaging Microstructures on Surfaces. Miles Koppang, Robert Carson, Paul Verbanac and Wenchun Chao...... 311

South Dakota Academy of Science 1997 Junior Academy Winning Papers Stuck with Duck Yuck? Justin L. Herreman ...... 315

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 1

SOUTH DAKOTA ACADEMY OF SCIENCE 82ND ANNUAL MEETING APRIL 25-26, 1997 NORTHERN STATE UNIVERSITY ABERDEEN, SOUTH DAKOTA

1996-97 OFFICERS

Executive Committee

PRESIDENT John Thomas, University of South Dakota

PRESIDENT-ELECT Sharon Clay, South Dakota State University

FIRST VICE-PRESIDENT Royce Engstrom, University of South Dakota

SECOND-VICE PRESIDENT Neil Reese, South Dakota State University

SECRETARY-TREASURER Bill Soefﬁng, University of Sioux Falls

PROCEEDINGS EDITOR Emil Knapp, Augustana College

FIRST PAST PRESIDENT Tim Sorenson, Augustana College

SECOND PAST PRESIDENT Arvid Boe, South Dakota State University

MEMBERS AT LARGE Maureen Diggins, Augustana College Tim Mullican, Dakota Wesleyan University Chuck Estee, University of South Dakota Gary Larson, South Dakota State University

Junior Academy Committee

Harlan Heitz (Chair), Ipswich High School Mike Barondeau, Edmunds Central High School Randy Gross, Yankton High School Marv Selnes, Sioux Falls Patrick Henry Middle School Jerry Loomer, Rapid City Stevens High School Cassandra Soefﬁng, Sioux Falls Axtell Park Middle School Rick Simmons, Dakota State University 2 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Membership Committee

Arvid Boe, South Dakota State University Carroll Hanten, S.D. Dept. of Agriculture Tom Cox, Black Hills State University Bob Tatina, Dakota Wesleyan University Seema Bharathan, Northern State University Tom Durkin, S.D. Dept. of Envir. and Nat. Resources

Nominations Committee

Fred Peabody, University of South Dakota John Naughten, Northern State University Mike Hildreth, South Dakota State University Jim Lefferts, Dakota Wesleyan University Curtis Card, Black Hills State University

Local Arrangements

Lenore Koczon and Susan Landon Northern State University

1997 AUDIT COMMITTEE REPORT

The Committee has examined the records of the Treasurer and finds them to be in excellent order and in documents agreement with the distributed statement for the 1996 fiscal year. The Committee wishes to comment Treasurer Bill Soeffing for his careful and detailed record keeping, and recommends a continuation of the summary reporting form now being utilized. The Committee further recommends that a committee composed to the Treasurer, President, and President-Elect investigate the possibilities for rein- vestment of the certification of Deposit maturing in August in an insured in- strument more advantageous to the Academy, taking whatever action they deem advisable.

Respectfully Submitted,

Audrey Gabel Mark Gabel Bill Jensen Chuck Estee Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 3

STATEMENT OF RECEIPTS, DISBURSEMENTS AND CHANGES IN CASH BALANCES FOR FISCAL YEAR 1996

Cash Balance on 1 January 1996 Certiﬁcate of Deposit $ 4898.35 Savings Account $ 9.15 Checking Account $ 3291.30 TOTAL BEGINNING CASH $ 9198.80

Receipts Membership Dues (Life-34/Regular-104/Associate-6) $ 2110.00 AAAS $ 1500.00 Gifts to Jr. Academy $ 120.00 Publications Sales $ 12.12 Page Charges $ 4027.00 Annual Meeting Registration $ 760.00 Banquet Tickets $ 492.00 Interest on Investments $ 263.12 TOTAL RECEIPTS $ 9284.24

Disbursements SDAS Proceedings Publication $ 5396.50 Annual Meeting $ 885.62 Ofﬁce Supplies $ 63.22 Postage $ 244.62 Duplication $ 114.21 Junior Academy $ 550.00 Refunds $ 270.00 TOTAL DISBURSEMENT $ 7524.17

Cash Balance on 31 December 1996 Certiﬁcate of Deposit $ 5161.47 Savings Account $ 9.15 Checking Account $ 4788.25 TOTAL ENDING CASH $ 9958.87

1997 Encumbrances Augustana College $ 271.52 1996 NAAS Dues $ 67.00 SD Secretary of State Corporate Filing Fee $ 35.00 TOTAL ENCUMBRANCES $ 373.52

AVAILABLE CASH (March 1, 1997) $ 9585.35 4 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

1997 REPORT OF THE RESOLUTIONS COMMITTEE

The membership of the South Dakota Academy of Science thanks Dr. John M. Hutchinson, President of Northern State University, for making the campus facilities available, for his provocative discussion on the thermodynamic status of Hell (exothermic or endothermic), and for hosting the 1997 annual meeting of the Academy. The Academy extends its thanks and appreciation to the local planning committee—Lenore Koczon, Susan Landon, and Seema Bhorathan—for making the excellent arrangements for the meeting. Thanks to President-Elect Sharon Clay for her address “Should We Live With Our Paradigms”. Would that we (South Dakota Academy of Science) could unify and become a significant voice in the affairs of South Dakota. We also wish to thank Dr. Daniel E. Brown for his provocative address, “Stress in Working Women”. President John Thomas is commended for his fine leadership this past year. The Academy is appreciative of the time, effort, and perseverance of our Secretary-Treasurer, Bill Soeffing and our Editor of the Proceedings, Emil Knapp. We extend heartfelt congratulations to Judy VonDruska of Mitchell High School, who was recognized as Physical Science Teacher of the Year and to Bob Sprang of Mitchell High School, recipient of the Biology Teacher of the Year Award. The leadership and participants of the Junior Academy are recognized for their continued fine work. Thanks are extended especially to Harlan Heitz of Ipswich High School (Chair) and all the teachers whose students participated in the Junior Academy. We also want to recognize and encourage the Junior Academy Award win- ners: Teresa Kub, Justin Herremen, and Laura Grant.

Respectfully submitted,

Nels H. Granholm Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 5

PRESIDENTIAL ADDRESS Can we live with our paradigms?

Address to the South Dakota Academy of Science Northern State University April 25, 1997

Presented by Sharon A. Clay South Dakota State University

I will confess to you this morning that I am a weed scientist. Right now, if your the average person, one of two things just popped into your head. The first is “I should ask her how to control ______(you fill in the blank) in my garden, (lawn, field)”. What you really want is a prescription for a herbicide, and if I respond with a mechanical method (hoeing, mowing, pulling), you may be disappointed. The other response to my confession is something like “She sprays chemicals (maybe too many damn chemicals), and is ruining the environment. I should discuss that with her”. Either way (giving herbicide recommendations or spraying lots of chemicals) is not what I really do. What you just did was fall into the Pesticide Paradigm when thinking about weed scientists. I'm not saying that there are not weed scientists who do just the sorts of things you thought of, I'm just saying that not all weed scientists are alike. But let's examine the Pesticide Paradigm further. A paradigm is defined as a pattern, standard, or example. Kuhn defined a scientific paradigm as “a universally recognized scientific achievment that for a time provides model problems and solutions to a community of practitioners”. Well, the community of practioners are the weed scientists. The scientific achievement? What is the time frame? From about 1946 to the present and in- to the foreseeable future. What is the scientific achievement? Herbicides, herbicides, and more herbicides. Which is why, if I answer your control question with a mechanical method, you are disappointed. You have bought into the herbicide example for controlling weeds. Starting with 2,4-D, moving on to dicamba, striding towards the sulfonyl ureas and immidiazaliones and now, the crowning achievement, having herbicide-resistant crops, so that no longer do we have to spend a lot of time thinking about carryover, crop safety, or timing. Herbicides have served us well. Very well. Today, we have less people in farming, farming more acreage, and feeding more people than ever before. Changes in crop varieties, fertilizer practices, and other changes have helped in this revolution. But herbicides and the control they provide for weeds are a large part of the equation. In some cases, crop rotation, cover crops, and mechanical techniques have been scraped. If you pick up weed science journals from the 1950's through about the early 1980's, 80% or more of the articles are devoted to herbicides, rates of 6 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) application, dates, combinations, mode of action, or efficacy. We learned the pesticide model for weed control very well. In 1991, Dr. R. Zimdahl wrote an essay entitled “Weed Science: A Plea for Thought”. In this essay, Dr. Zimdahl asked weed scientists to examine the pesticide paradigm, to see where it has led us, and to examine where it will take the practioners in the future. There has been excellent research that can be traced back to herbicides. For example, plant physiology and biochemistry have been advanced in the areas of the shikimate acid cycle for aromatic amino acid synthesis, the branched chain amino acid synthesis reactions, lipid synthesis, and photosynthesis including the structure of the D1 protein and binding site for PQ. We have also gained understanding in the areas of water flow through soils, particle transport, and microbiology. That is where we have been. Where will the herbicide paradigm lead us in the future? Herbicides will continue to have an important and major role in agriculture. There is nothing at the present to replace the efficiency, cost, and labor savings that herbicides bring to agriculture. But weed scientists should not be blinded by the herbicide light and as Zimdahl puts it the “controlling knowledge” of the tool. We are just beginning to think about weed ecology, biological control, site-specific management, and using long-term (4 to 6 yr) rotations for the primary purpose of weed management (not control). When the herbicide paradigm works so well, why change to something else? There are several reasons. The first is that plants are showing resistance to herbicides. The first plant that was documented to be resistant to triazine herbicides was Senecio vulgaris in Idaho in 1968. Today, hundreds of plants in several species have shown resistance to many different types of chemicals. Today's most promising herbicides for reduction of offsite contamination are used at very low rates (ounces per acre) and only affect a single enzyme in the plant. However, some weeds are selected for resistance after as little as two continuous applications and because of their dispersal mechanism can be transported for miles from the original site of application. In the case of kochia, single mutations confer resistance up to 75 times normal field rate. In order to reduce the chances of selecting for resistant weeds, the label directions indicate that these chemicals should be used only once every 48 months and then in combination with some other herbicide that can control the resistant population. So, resistant weeds may shift the herbicide paradigm. The second reason to examine a new paradigm is societal pressure. In the 1950's, scientists had instrumentation that could detect herbicides in the ppm range. Today's instrumentation can detect ppb, ppt, and in some cases ppq. Society is concerned about herbicides in the environment even at very low levels. Herbicides can be found in surface and ground waters, in rain, and in soils in areas far from cropland. Zimdahl presents three dilemas that may be unresolvable but are nevertheless important. The first is that pesticides, as with other unknowns, makes- people nervous. There is a basic lack of scientific understanding concerning basic environmental and health implications of pesticide use. What level really causes birth defects? cancer? How does long low rate exposure affect us? Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 7

The second dilema is that the full cost of use (on health, environmental quality, and non-target organisms) is impossible to derive. The third is if even if we did identify all the problems, there would still be risk involved and not everone would agree on what is a “safe” risk. Individuals feel helpless because it is not their choice to be exposed to pesticides. When 97% of the corn acres are treated with herbicide, you have very little choice in exposure. I am not advocating tossing out herbicides. That would be akin to throw- ing out the baby with the bath. However, we must rethink our models in weed science. Herbicides are so great a tool and allowed for large farms to be the norm that growers have little time for tillage, crop rotation, and cover crops. We are so used to seeing immediate reults that give 90%+ control that we have no time for establishment of bioagents. Zimdahl argues that we need to stress preventative technology rather than control. We need to think in terms of weed management. Weed scientists need to become better ecologists, weed specialists, and managers not just control specialists. Moving away from a successful paradigm will be difficult. Asking the right questions may be even more difficult. But where will progress be made? Someone said that the world is an ocean of ideas and one just bobs to the surface. But we must have some background knowledge of the subject to be able to understand if the idea is good or bad and we must be able to see beyond the present paradigms to come up with inovative ideas. Often progress is made by people on the fringes of the field of research. This may be because they have few preconcieved ideas of what cannot be done. They don't have a lot of energy, money, or emotional baggage invested in the paradigm and therefore, have less to lose if their hypothesis is proven wrong. Collaboration and interdisciplinary research (ie bringing new ideas from “outsiders” of the science) may help to solve old problems. For example, I work with soil scientists, entomologists, genetists, microbiologists, and remote sensing specialists in integrated research projects to work on weed science problems. What are your paradigms that you are confronted with? Do we need to change other paradigms? One of the paradigms that Dr. Nels Granholm discussed in his SDAS presidental address in 1992 was the “Think Small” Syn- drome for South Dakota. It goes like this “South Dakota is a poor state lacking in many of the resources of our surrounding states. We're at the low end of teachers' salaries, per capita income, industrial output, and gross state product.” As a result we come to the conclusion that “… we cannot compete with people at the national level … It has always been like this and we might as well get used to it.” He also stated “Let’s not be intimidated by the granting process or any other facet of this business of doing science.” I totally agree with him. There are researchers in the state that are being awarded substantial amounts of money. However, in 1997, there are a lot of us that are still wrapped in the think small syndrome. We also cannot ignore what is happening in South Dakota. As Dr. Sword stated in an interview about his retirement as SDSU graduate dean and director of research “Never have I seen morale so low among faculty … It will take many years to rebuild (SDSU) back to where it was a few years ago …” The 8 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) frustration is also seen in students as a recent editorial in the Collegian states … SDSU is a sinking ship in an unfriendly ocean of efficiency …” Between increasing out-of-state tuition, cancelling small class sizes, implimenting the sophomore proficiency exam, and the downsizing of other curriculm, students have become just as frustrated as the faculty. This frustartion appears also to be evident in the number of papers submitted here. In the early 1990’s there were about 100 papers submitted, today there are 45. Perhaps it was this long, cold winter that has also contributed to this. Or perhaps it is the think small paradigm. I have had the opportunity to sit on several granting panels this year from the regional to national level. If we just play the numbers game (ie a certain percentage of the propsals will get funded) South Dakota gets blown out of the water. In North Central IPM funding project, MN had 12 proposals, 5 each from NE and IN, 7 from WI, and 8 from OH. SD submitted one. Is that all the expertise we have in IPM in the state? I don’t think so. Of the 120 grant proposals submitted to NRI, one came from SD. Yes, grants are a pain to apply for. Yes, you do need to have good ideas and develop them. But part of this reflects back on the think small syndrome. We are only one deep in certain disciplines in the state. This makes it hard, but perhaps this is a real opportunity to create truly multidisiplinary teams to begin working in systems research. We don’t have to step across too many departments to work together on a problem. But even this means re- thinking our own paradigms. Arguing and discussing problems and their possible solutions and perhaps using new methods may give us the “competitive edge” over larger more disciplined-orientated institutions. As a scientific society, I hope that we can use each others diverse experience to help resolve problems of today and challenges of tomorrow. Part of the solution is examing the paradigms we have grown to know and love. We must keep the beneficial parts of the model but rethink faulty examples to make real progress toward the goals of our respective disciplines. I hope that today, as you sit through the scientific presentations, you gain new insight in- to your paradigms and begin to expand to ask new questions for progress in our everchanging world of science. Complete Senior Research Papers

presented at The 82nd Annual Meeting

of the South Dakota Academy of Science

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 11

FISHES OF THE MAINSTEM CHEYENNE RIVER IN SOUTH DAKOTA

Douglas R. Hampton Department of Wildlife and Fisheries Sciences and Charles R. Berry, Jr. U.S. Geological Survey South Dakota Cooperative Research Unit South Dakota State University Brookings, South Dakota 57007

Abstract

Nine reaches on the mainstem Cheyenne River were sampled during the summer months of 1996 and 1997 to document the current distribution and abundance of fishes. Thirty species representing ten families were collected. Eighty-five percent of the total catch included flathead chub Platygobio gracilis (28%), plains minnow Hybognathus placitus (24%), western silvery minnow H. argyritis (20%), sand shiner Notropis ludibundus (7%), and channel catfish Ic- talurus punctatus (6%). Five species previously considered rare in the state of South Dakota were collected (sturgeon chub Macrhybopsis gelida, plains topminnow Fundulus sciadicus, flathead chub, plains minnow, and western silvery minnow). Channel catfish were captured in all reaches at rates of 0.3-2.5 fish/trap net set. Relative weight (Wr) of channel catfish averaged 106±2 for substock fish, 83±1 for stock-to-quality fish, and 82±3 for quality- to-preferred fish. This is the first comprehensive fishery survey of the Cheyenne River mainstem downstream from Angustora Reservoir. These results have a bearing on irrigation, grassland management, scenic river designation, rare fish conservation, and recreation. This information will also be useful as baseline data for future analysis of fish community health.

Introduction

Since Lewis and Clark’s Corps of Discovery first passed the mouth of the Cheyenne River on their way up the Missouri, many surveys have been conducted on the fish community in the Cheyenne River Basin. However, most work has been on the coldwater tributaries from the Black Hills where recreational trout fishing is important. Five limited surveys have been conducted on the warmwater mainstem of the Cheyenne River that crosses prairie grasslands (Churchill and Over 1933; Bailey and Allum 1962; Koth and Ford 1980; Green et al. 1990, Cunningham et al. 1995). These authors collectively documented 19 species (Table 1) in the mainstem, of which only one is currently on the state’s threatened list (sturgeon chub Macrhybopsis gelida). These historic studies are of limited use, however, because they are dated, few sites were 12 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) sampled, and little habitat information was collected. In addition, previous authors collected species presence data only and did not quantitatively assess the fishery or make inferences about patterns of community distribution. Conservation of the State’s rivers and riverine biota requires current information about fish communities, habitat conditions, and public issues (SDGFP 1994). Fisheries information is needed for sport fish management, protection of rare fishes, commenting on developments (e.g. irrigation, grazing, scenic river designation) and biomonitoring. Our objectives were to determine the current distribution and relative abundance of fishes in the mainstem Cheyenne River, and measure habitat conditions when and where fish were collected.

Table 1. Fish species reported between Angostura Dam and Lake Oahe in the mainstem Cheyenne River, in South Dakota by ﬁve investigations from 1933 to 1997.

Species name Common name

Species documented during earlier studies* 1. Platygobio gracilis Flathead chub 2. Hybognathus placitus Plains minnow 3. Notropis ludibundus Sand shiner 4. Ictalurus punctatus Channel catfish 5. Moxostoma macrolepidotum Shorthead redhorse 6. Carpiodes carpio River carpsucker 7. Noturus flavus Stonecat 8. Macrhybopsis gelida Sturgeon chub 9. Catostomus commersoni White sucker 10. Rhinichthys cataractae Longnose dace 11. Pimephales promelas Fathead minnow 12. Fundulus zebrinus Plains killifish 13. Ameiurus melas Black bullhead 14. Lepomis cyanellus Green sunfish 15. Lepomis humilis Orangespotted sunfish** 16. Hiodon alosoides Goldeye 17. Micropterus dolomieu Smallmouth bass 18. Cyprinus carpio Common carp 19. Stizostedion canadense Sauger

Species found during the present study not previously recorded 20. Hybognathus argyritis W. silvery minnow 21. Notropis atherinoides Emerald shiner 22. Cyprinella lutrensis Red shiner 23. Morone chrysops White bass 24. Aplodinotus grunniens Freshwater drum 25. Notropis hudsonius Spottail shiner 26. Fundulus sciadicus Plains topminnow 27. Micropterus salmoides Largemouth bass 28. Semotilus atromaculatus Creek chub 29. Ameiurus natalis Yellow bullhead 30. Lepomis macrochirus Bluegill 31. Esox lucius Northern pike

*Churchill and Over 1933; Bailey and Allum 1962; Green et al. 1990, Cunningham et al. 1995 **Not captured during the present study Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 13

STUDY AREA

The Cheyenne River is the largest western tributary to the Missouri River in South Dakota draining 84,434 square kilometers, approximately half of which are in South Dakota (Fig. 1). The Cheyenne originates in eastern Wyoming and northwestern Nebraska before winding around the southern Black Hills and emerging onto the high plains of western South Dakota. From it’s confluence with the Belle Fourche River, the Cheyenne courses east towards Lake Oahe on the Missouri River. The watershed is in the semi-arid section of the western xeric ecoregion (Omernik 1987). The Cheyenne River fits the general characteristics of rivers in this ecoregion in that it receives water from outside the region, or from outlier mountains within the region, and is an influent river (e.g. loss of water from channel to water table). Waters of this ecoregion are generally characterized as being high in nutrients and suspended sediment. Most of the region is in grazing or cropland and water quality problems are associated with natural geological features, agricultural activities and mining (DENR 1994). The Cheyenne River is impounded by Angostura Dam, which was constructed by the Bureau of Reclamation in 1946-49 about 14 km southeast of Hot Springs, South Dakota. Irrigation water is delivered to about 5,500 ha from May through September, so flows are minimal in the 7 km reach below the dam during the irrigation season. Irrigation return flows drain to the Cheyenne River throughout the length (about 48 km) of the project (Greene et al. 1990). Portions of the river lie adjacent to National Forest and Grassland units managed by the U. S. Forest Service. The Pine Ridge and Cheyenne Riv- er Indian Reservations border parts of the river.

Figure 1. Map of South Dakota showing 9 reach locations where ﬁshes and habitat were sampled on the Cheyenne River in 1996 and 1997. 14 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

We sampled ﬁsh and habitat at nine equidistant sites along the 306 km section of river between Angostura Dam and the Cheyenne River arm of Lake Oa- he (Table 2). All sampling was conducted during June, July, and August in 1996, and July, August, and September in 1997.

Table 2. Township, range, section, and universal transverse mercator (UTM) coordinates of nine reaches on the Cheyenne River where ﬁshes were collected and habitat was measured in 1996–1997.

Segment Reach Township Range Section UTM

Upper 1 8S 7E 6 635300 E 4805690 N 2 5S 10E 18 662987 E 4830571 N 3 2S 12E 21 685890 E 4858220 N Middle 4 3N 15E 7 710660 E 4901500 N 5 5N 15E 32 712330 E 4913190 N 6 6N 17E 32 730747 E 4925792 N Lower 7 7N 18E 31 263174 E 4934476 N 8 7N 21E 17 296664 E 4938405 N 9 8N 23E 2 317310 E 4950390 N

METHODS

We made standard habitat measurements (Simonson et al. 1994) in con- junction with fish sampling in order to document the conditions when and where fishes were collected. The basic unit of study was the river reach, which varied in length as a function of river width. Upon arrival at each site 10 measurements of river surface width were made 10 meters apart to calculate mean river width. Transects perpendicular to the direction of flow were spaced three mean river widths apart at all reaches. Cross-sectional profiles were derived by measuring physical habitat variables along five transects at every reach (Gor- don et al. 1995), except at reaches 3, 6, and 9 where 13 transects were used. We measured depth (m) with a wading rod at one-quarter, one-half, and three- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 15 quarters of the stream width on every transect. Discharge (cfs) was measured using a Price flowmeter at 20 evenly spaced points across the wetted portion of the transect exhibiting the most laminar flow in every reach. At each reach we measured total dissolved solids (ppm), conductivity (µS), pH, and temperature (Cº) with hand-held meters (Cole-Parmer models 59000-20, 19800-00, and 19800-30), and water clarity (cm) with a weighted Secchi disk. Photos were taken at each reach (Hampton 1998). Fishes were sampled following water quality measurements, but prior to making habitat measurements in order to minimize disturbance related to our presence. We employed both active and passive gears (i.e. seines and trap nets); at least two samples were taken with both gears at every reach. One of two knotless nylon seines (9-m-long, 1.2-m-deep, 4.7-mm-bar mesh, or 15-m- long, 1.2-m-deep, 4.7-mm-bar mesh) were used depending on river surface width. Seine hauls were conducted in a downstream direction without the use of block nets, and the distance seined (m) was recorded. Trap nets (1.22-m- deep, 1.83-m-long, 1.27-cm-bar mesh) were set facing downstream in deeper areas and left overnight for 12 hours. Fish were held in holding pens and anes- thetized with compressed carbon dioxide. All individuals were identified and counted; channel catfish were weighed (g) and measured (mm) for total length. Most fish were placed in recovery tanks filled with river water and returned to the river. We employed two separate indices to further analyze the fish data we collected. Catch per unit effort (CPUE) is an index of density and was calculated separately by species for every seine haul and trap set as the total number of fish caught per meter seined or per net night, respectively. Relative weight (Wr) was calculated using length and weight data to assess channel catfish condition (Brown et al. 1995). Condition indices (e.g. Wr) measure the robustness of individual fish and are related to fish health, reproductive state, and growth. A curvelinear regression model (Wr = 63.75 + 5,780 / total length, R = 0.80, P < 0.001) was used to predict Wr for given fish lengths. The predicted curve for Cheyenne River channel catfish was then compared to similar curves for the same length categories of channel catfish from the Belle Fourche (Doorenbos 1998) and Moreau (Loomis 1997) rivers. Channel catfish length categories were substock (70-279 mm), stock to quality (280-409 mm), quality to preferred (410- 609 mm), and preferred to memorable (610-709 mm). All channel catfish were sampled during the months of July, August, and September in 1996 and 1997 and were presumed to be post-spawn (Carlander 1969, Gerhardt and Hubert 1990). Sampling after spawning minimizes length and weight differences between sexes (Carlander 1969).

RESULTS

A general pattern of increasing river width and depth was associated with increasing discharge from upstream to downstream reaches (Table 3). The flow of the Belle Fourche River more than doubled the flow of the Cheyenne River between reaches five (165 cfs) and six (454 cfs). Mean stream width increased from 15.5 m at reach one to 74.1 m at reach nine downstream. Mean 16 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) depth ranged from 0.28 m at reach three to 0.61 m at reach nine. Water temperatures varied from 15º to 27º C , and pH remained fairly constant between 7.8 and 8.4. Specific conductance ranged from 1,600 to 2,100 µS, and total dissolved solids fluctuated between 870 and 1,290 ppm.

Table 3. Habitat characteristics measured at nine reaches on the Cheyenne Riv- er in South Dakota between Angostura Dam and Lake Oahe in 1996.

Reach Mean Mean Discharge Temperature pH Total Conductivity and surface depth (cfs) (ºC) dissolved (µS) Date width) (m) solids (m) (ppm)

1 15.5 ± 1.65 0.39 ± 0.03 36 26 8.0 870 1900 Jul 10 2 16.1 ± 1.46 0.39 ± 0.02 97 27 7.9 1290 2100 Aug 24 3 32.8 ± 3.27 0.28 ± 0.02 117 15 8.4 870 1600 Jul 17 4 47.7 ± 7.02 0.29 ± 0.04 156 19 8.0 940 2000 Aug 23 5 39.1 ± 5.07 0.38 ± 0.03 166 24 8.0 1010 1900 Aug 21 6 61.5 ± 5.69 0.55 ± 0.04 454 20 8.4 1050 2100 Jul 25 7 70.0 ± 9.95 0.61 ± 0.06 529 21 7.9 1160 2400 Aug 14 8 69.9 ± 8.78 0.52 ± 0.04 591 21 7.8 1080 2200 Aug 15 9 74.1 ± 3.92 0.61 ± 0.04 493 24 8.2 1090 2000 Aug 5

In our total catch of 3,896 fish were 30 species representing 10 families (Table 4). We observed what was probably a shortnose gar (Lepisosteus platostomus), but none were captured. The flathead chub had the highest CPUE in seines (0.35), whereas 19 species had CPUE values of <0.01. The coefficient of variation for means of seining CPUE ranged from 19% - 48% for 13 of the most frequently captured species (i.e. flathead chub, sand shiner, western silvery minnow, sturgeon chub, channel catfish, stonecat, river carpsucker, shorthead redhorse, plains minnow, red shiner, longnose dace, smallmouth, and white bass), whereas those of all other species exceeded 50%. CPUE data were highly variable among specific reaches (Appendix Tables A1, A2). Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 17

Table 4. Summary of catch data from the Cheyenne River in South Dakota for 1996-1997. Catch per unit effort (CPUE) is expressed as ﬁsh/m for seine hauls and ﬁsh/net night for trap nets.

Family Number of River Mean Mean Total and reaches segments CPUE (SE) CPUE (SE) number species where where using using of ﬁsh present captured* seines** trap nets** caught

Cyprinidae Common carp 5 U, M, L 0.001 ± 0.001 0.05 ± 0.03 6 Creek chub 2 U 0.001 ± 0.001 -- -- 2 Emerald shiner 2 M, L 0.014 ± 0.009 -- -- 67 Fathead minnow 4 U, M 0.004 ± 0.002 -- -- 14 Flathead chub 9 U, M, L 0.353 ± 0.066 0.32 ± 0.14 1078 Longnose dace 6 U, M, L 0.005 ± 0.002 -- -- 17 Plains minnow 6 U, M, L 0.206 ± 0.099 -- -- 940 Red shiner 5 U, M 0.017 ± 0.008 0.02 ± 0.02 42 Sand shiner 7 U, M, L 0.089 ± 0.034 -- -- 265 W. silvery minnow 8 U, M, L 0.206 ± 0.071 -- -- 784 Sturgeon chub 5 U, M, L 0.008 ± 0.002 -- -- 26 Spottail shiner 3 L 0.003 ± 0.002 0.02 ± 0.02 9 Ictaluridae Black bullhead 3 U, M 0.002 ± 0.001 0.02 ± 0.02 3 Channel catfish 9 U, M, L 0.064 ± 0.013 0.97 ± 0.19 253 Stonecat 5 M, L 0.009 ± 0.003 0.18 ± 0.09 36 Yellow bullhead 2 U < 0.001 ± < 0.001 0.02 ± 0.02 2 Catostomidae River carpsucker 7 U, M, L 0.009 ± 0.003 0.15 ± 0.06 40 Shorthead redhorse 9 U, M, L 0.035 ± 0.011 1.03 ± 0.38 150 White sucker 2 U 0.010 ± 0.006 0.03 ± 0.02 26 Clupeidae Goldeye 6 U, M, L 0.022 ± 0.013 0.08 ± 0.04 47 Percichthyidae White bass 6 M, L 0.008 ± 0.003 0.07 ± 0.03 26 Centrarchidae Bluegill 1 U < 0.001 ± < 0.001 -- -- 1 Green sunfish 1 U 0.001 ± 0.001 -- -- 2 Largemouth bass 1 L 0.001 ± 0.001 -- -- 6 Smallmouth bass 1 U 0.009 ± 0.004 0.02 ± 0.02 16 Sciaenidae Freshwater drum 5 M, L 0.005 ± 0.004 0.05 ± 0.04 18 Percidae Sauger 4 U, M, L 0.001 ± 0.001 0.08 ± 0.04 10 Cyprinodontidae Plains killifish 1 U 0.002 ± 0.001 -- -- 3 Plains topminnow 1 U 0.004 ± 0.003 -- -- 6 Esocidae Northern Pike 1 U -- -- 0.02 ± 0.02 1

*U, M, L indicates Upper (reaches 1-3), Middle (reaches 4-6), and Lower (reaches 7-9) river segments **-- -- indicates no individuals were caught with this gear type 18 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

The Cyprinidae family had the most taxa of all families with 12 species ac- counting for 83% of all individuals captured. Flathead chub was the most abundant cyprinid, comprising 28% of all individuals captured. Plains and western silvery minnow were also abundant, both species combined making up 44% of all individuals caught. Ictaluridae (four species) and Catostomidae (three species) were the next two most abundant families representing 8% and 6% of all fish caught, respectively. Channel catfish constituted 6% of the total catch, and is the most abundant sport fish in the Cheyenne River mainstem. In addition to channel catfish, we found several other species of game fish, but they made up only 1.7% of the total catch. Northern pike and bluegill were represented by only one specimen each. We captured 5 bullheads (i.e. black and yellow), 33 bass (i.e. white, largemouth, and smallmouth), and 10 sauger. The channel catfish were ubiquitous, but the distribution of other game fish species was spotty. For example, sunfishes, smallmouth bass, and northern pike were found at only one upper basin reach, whereas the white and largemouth bass were found only in the lower reaches. We found five species (i.e. flathead chub, plains minnow, western silvery minnow, plains topminnow, and sturgeon chub) that have been considered rare. At five middle reaches, we found 26 sturgeon chubs, which is a state threatened species in South Dakota and a candidate for federal listing as a threatened or endangered species. Sturgeon chubs were found in runs and rif- fles where depths were 31- 74 cm, water velocity was 0.4-0.9 m/sec, and bottom substrates were 40% silt, 30% sand, and 30% medium to course gravel. Flathead chubs, plains minnow, and western silvery minnow were commonly found at six or more reaches (Table 4). We captured only 6 plains topminnow and these were found only at the most upstream reach. The relative weight (Wr) of Cheyenne River channel catfish (n = 182) varied with fish length (Figure 2). Generally, Wr values decreased with increasing length. Substock length channel catfish (70-279 mm) had Wr values (mean ± SE, N) ranging from 73 to 184 (106 ± 2, 156), whereas mid-sized stock to quality length fish (280-409 mm) and longer (≥ 410 mm) had Wr values ranging from 75 to 96 (83 ± 1, 16) and 59 to 100 (82 ± 3, 10), respectively.

DISCUSSION

Previous collections of fish from the Cheyenne River basin are of limited use because they lacked reference to exact sampling locations (Churchill and Over 1933), were from the upper basin (Koth and Ford 1980), or were for contaminant analysis (Green et al. 1990). These surveys recorded 19 species between Angustora Dam and Lake Oahe. We added 12 species to the list (Table 1) and missed only one species (orangespotted sunfish) that was found in one earlier study. Bailey and Allum (1962) noted that orangespotted sunfish were uncommon in western South Dakota rivers and were probably introduced via accidental stocking. Cunningham et al. (1995) sampled three locations in the middle reaches of the mainstem Cheyenne River in 1994 and documented eleven species. Two of the species they reported (i.e. western silvery minnow, and quillback, Car- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 19

160

S-S S-Q Q-P P-M 140

120

Wr 100

60 0 100 200 300 400 500 600 700 Total Length Belle Fourche Cheyenne Moreau

Figure 2. Predicted relative weight (Wr) of channel catfish for given lengths of fish from the Belle Fourche, Cheyenne, and Moreau rivers in South Dakota, 1995–1997. Abbreviations for length categories are S-S = substock, S-Q = stock to quality, Q-P = quality to preferred, and P-M = preferred to memorable. piodes cyprinus) were previously undocumented in the mainstem. We examined their voucher specimens and believe that their quillback was actually a river carpsucker. River carpsucker has been previously documented in the mainstem Cheyenne River, whereas western silvery minnow were probably present during earlier investigations but identified as Hybognathus species. Fish assemblages differ between western and eastern tributaries to the Mis- souri River in South Dakota. Comparisons were made between the species identified in the Cheyenne River during this study and species inventories from the Little Missouri (Bich and Scalet 1977), Moreau (Loomis 1997), Big Sioux (Di- eterman 1995), and James (Berry et al. 1993) rivers in South Dakota. Eleven species were common to both eastern and western rivers (i.e. channel catfish, sand shiner, shorthead redhorse, river carpsucker, goldeye, white sucker, fathead minnow, common carp, black bullhead, green sunfish, and largemouth bass), although largemouth bass were rarely found in any of the western rivers (i.e. Cheyenne, Little Missouri, and Moreau). Northern pike were found in all rivers except for the Moreau, and were scarce in the Cheyenne and Little Mis- souri rivers. We found three yellow bullhead in the upper segment of the Cheyenne River (Tables A1, A2), which is out of the range for this species as reported by Scott and Crossman (1973). Smallmouth bass and plains topminnow were only documented in the James and Cheyenne rivers. Numerous species reported in the eastern tributaries were not documented in the western rivers. Flathead chub, sturgeon chub, longnose dace, and 20 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) plains minnow were recorded only in western South Dakota rivers, with the single exception of flathead chub being reported in the Big Sioux River as late as 1956 (Nickum and Sinning 1971). Flathead chub have not been reported in the Big Sioux since that time, however. Among the western rivers (i.e. Cheyenne, Little Missouri, and Moreau), plains killifish were unique to the Cheyenne River, sturgeon chub undocumented in the Moreau River, and western silvery minnow found in the Cheyenne and Moreau rivers , but not in the Little Missouri. Species of Hybognathus are difficult to identify (Loomis 1997), so the plains or central silvery minnow (Hybognathus nuchalis) reported by Bich and Scalet (1977) in the Little Missouri River may have been western silvery minnow. Flathead chub, plains minnow, and western silvery minnow were previously listed as state species of concern in South Dakota, but in 1995 they were removed from this list (D. Backlund, South Dakota Natural Heritage Program, personal communication). The abundance of these three species, and especially flathead chub, in the Cheyenne River supports this decision. Similarly, plains topminnow was previously listed as a state threatened species in South Dakota, but was removed from this listing in 1996 because they were more widespread and not as restricted by habitat requirements as previously thought (D. Backlund, South Dakota Natural Heritage Program, personal communication). The upper reaches of the Cheyenne River are on the northern periphery of the range for plains topminnow. We found the plains topminnow only in the first reach where the water was relatively clear and slow flowing due to the low discharge from Angostura Dam. The low flow conditions at the first reach were generally similar to the preferred habitat conditions reported for plains topminnow (Pflieger 1971). Sturgeon chub are currently listed as a threatened species in South Dakota and are a candidate for Federal listing. The presence of 26 sturgeon chub in five of the middle reaches of the study area may indicate that this species is also more common and widely distributed than once thought. We present the predicted Wr curve for channel catfish as a benchmark for comparison with data from future studies in the Cheyenne River. The common notion that the “target” for Wr values is between 95 and 105 should be ques- tioned in this case. The percentage of channel catfish subpopulations having Wr values between 95 and 105 is less than 45% (Brown et al. 1995), a finding that agrees with our data. Channel catfish Wr values change as fish grow (Brown et al. 1995), but the model for riverine populations in not well understood. The higher Wr values of the shorter length categories within the Cheyenne River population may indicate a greater prey availability for substock length, predominantly insectivorous fish, than for longer, typically omnivorous or piscivorous fish (Carlander 1969). Declining Wr’s for the longer length groups may also be due to our small sample size of fish longer than 280 mm (n = 26). The Wr values for specific length groups may reflect characteristics of the environment such as habitat quality and accessibility, water quality, and prey availability, but the relationships have not been well established (Liao et al. 1995). Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 21

The predicted Wr values for Cheyenne River channel catfish are similar to those from the Moreau (Loomis 1997) and Belle Fourche (Doorenbos 1998) rivers that were sampled concurrently (Figure 2). The Moreau River is in a relatively undisturbed basin (Loomis 1997), whereas the Belle Fourche (Dooren- bos 1998) and Cheyenne River basins have greater human populations and more industrial (e.g. mining) and agricultural (e.g. irrigation) use. The Belle Fourche, Moreau, and Cheyenne rivers are located within the same ecoregion and exposed to similar climatic conditions, so our Wr data should be comparable among rivers (Gabelhouse 1991). Our data relate to several river management issues. Our fish species list and river habitat data may be useful to the Bureau of Reclamation as it works to relicense the Angustora Reclamation Unit (described by Green et al. 1990). Fish communities and fishing recreation in the Cheyenne River will be considered when grassland management plans are assessed (USFS 1999). The Cheyenne River offers recreational fishing opportunities for channel catfish in the stock, quality, and perhaps preferred size classes. We observed very few anglers using the Cheyenne River other than local landowners and their families. The river offers a unique opportunity for float fishing through a scenic wilderness setting, but public access is limited (Hansen 1998). About 13 km of river could be classified by the U. S. Forest Service as scenic. We commonly found some fish species that are declining elsewhere, even though river water quality is termed “generally poor” (DENR 1994). Overall river conditions seem to have been suitable for maintaining the fish community. Our habitat data may help establish habitat management plans for rare species, and allow future surveys to be conducted under similar conditions, thus reducing variability. Population trends will be easier to identify for the dozen or so species with the least amount of variation in the CPUE data. For other species, increased sampling effort in specific macrohabitats will be needed to increase catch so that the persistence and stability of fish populations can be determined in the future.

ACKNOWLEDGMENTS

We would like to thank N. Smith and H. Hendrickson for their technical assistance, and R. Doorenbos, T. Loomis, and C. Milewski for sharing their data and collaborative efforts. Funding was provided by Federal Aid in Sport Fish Restoration under D. J. Project # F-15-R, Study # 1563. The South Dakota Co- operative Research Unit is jointly supported by the South Dakota Department of Game, Fish & Parks, the U. S. Geological Survey, South Dakota State Uni- versity, and the Wildlife Management Institute. 22 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

LITERATURE CITED

Bailey, R. M., and M. O. Allum. 1962. Fishes of South Dakota. University of Michigan Museum of Zoology, Miscellaneous Publications 119, Ann Arbor. Berry, C. R., W. G. Duffy, R. Welsh, S. Kubeny, D. Schumacher, and G. Van Eeckhout. 1993. The James River of the Dakotas. Pages 70-86 in L. W. Hesse, C. B. Stalnaker, N. G. Benson and J. R. Zuboy, editors. Restoration planning for the rivers of the Mississippi River ecosystem. National Bio- logical Survey, Biological Report 19, Washington, D.C. Bich, J. P., and C. G. Scalet. 1977. Fishes of the Little Missouri River, South Dakota. Proceedings of the South Dakota Academy of Science 56:163-177. Brown, M. L., F. Jaramillo, Jr., D. M. Gatlin III, and B. R. Murphy. 1995. A revised standard weight (Ws) equation for channel catfish. Journal of Fresh- water Ecology 10:295-302. Carlander, K. D. 1969. Handbook of freshwater fishery biology. The Iowa State University Press, Ames, Iowa. Churchill, E. P., and W. H. Over. 1933. Fishes of South Dakota. South Dako- ta Department of Game, Fish, and Parks, Pierre. Cunningham, G. R., R. D. Olson, and S. M. Hickey. 1995. Fish surveys of the streams and rivers in south central South Dakota west of the Missouri Riv- er. Proceedings of the South Dakota Academy of Science 74:55-64. DENR (Department of Environment and Natural Resources). 1994. The 1994 South Dakota Report to Congress: 305(b) Water Quality Assessment. South Dakota Department of Environment and Natural Resources, Pierre. Dieterman, D. J., 1995. The influence of the Clean Water Act and tributaries on the fish community of the Big Sioux River. M. S. Thesis. South Dakota State University, Brookings. Doorenbos, R. 1998. Fishes and habitat of the Belle Fourche River, South Dakota. M. S. Thesis. South Dakota State University, Brookings. Gabelhouse, D. W. 1991. Seasonal changes in body condition of white crap- pies and relations to length and growth in Melvern Reservoir, Kansas. North American Journal of Fisheries Management 11:50-56. Gerhardt, D. R., and W. A. Hubert. 1990. Spawning habitat of channel catfish in the Powder River system, Wyoming-Montana. Prairie Naturalist 22:155- 164. Gordon, N. D., T. A. McMahon, and B. L. Finlayson. 1995. Stream hydrology an introduction for ecologists. John Wiley and Sons, New York. Green, Earl, C. Sowards, and E. Hansmann. 1990. Reconnaissance investigation of water quality, bottom sediment, and biota associated with irrigation drainage in the Angustora Reclamation Unit, Southwestern South Dakota, 1988-9. Water Resources Investigations Report 90-4152, U. S. Geological Survey, Rapid City, South Dakota. Hampton, Douglas R. 1998. A survey of the fishes and habitat of the Cheyenne River in South Dakota. M.S. Thesis. South Dakota State Uni- versity, Brookings. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 23

Hansen, M. R. 1998. Canoeing the Cheyenne River. South Dakota Conserva- tion Digest 65(3):10-14. Koth, R. M. and R. C. Ford. 1980. Fisheries resources of the Cheyenne River basin Fall River County, South Dakota, July 7-9, 1980. Report 80-8. South Dakota Department of Game, Fish, and Parks, Pierre. Liao, H., C. L. Pierce, D. H. Wahl, J. B. Rasmussen, and W. C. Leggett. 1995. Relative weight (Wr) as a field assessment tool: relationships with growth, prey biomass, and environmental conditions. Transactions of the American Fisheries Society 124:387-400. Loomis, T. M. 1997. Survey of the fishes and habitat in the upper Moreau Riv- er, Perkins County, South Dakota. M. S. Thesis. South Dakota State Uni- versity, Brookings. Nickum, J. G. and J. A. Sinning. 1971. Fishes of the Big Sioux River: An an- notated list. Proceedings of the South Dakota Academy of Science 50:143- 154. Omernik, J. M. 1987. Ecoregions of the conterminous United States. Annals of the Association of American Geographers 77:118-125. Pflieger, W. L. 1971. A distributional study of Missouri fishes. University of Kansas, Lawrence. Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Fisheries Research Board of Canada, Ottawa. Simonson, T. D., J. Lyons, and P. D. Kanehl. 1994. Quantifying fish habitat in streams: transect spacing, sample size, and a proposed framework. North American Journal of Fisheries Management 14:607-615. SDFGP (South Dakota Department of Game, Fish, and Parks). 1994. Stream fisheries program strategic plan. South Dakota Department of Game, Fish, and Parks, Pierre. USFS (U. S. Forest Service). 1999. Summary of the Draft Environmental Impact Statement for the Northern Great Plains Management Plans Revision. US- DA Forest Service, Chadron, Nebraska. 24 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table A1. Catch per unit effort (CPUE) of ﬁshes seined in the Cheyenne Riv- er, South Dakota during 1996–1997 expressed as ﬁsh per meter (number) by species and reach*.

Common name Year Reach

123456789

Flathead chub 96 0.11 (11) 0.23 (46) 1.42 (142) 1.33 (114) 0.04 (8) 0.27 (35) 0.14 (15) 0.18 (27) 97 0.03 (4) 0.71 (138) 0.17 (40) 0.27 (67) 0.35 (77) 0.54 (157) 0.15 (37) 0.59 (135) Channel catﬁsh 96 0.01 (1) 0.03 (5) 0.07 (7) 0.08 (7) 0.20 (26) 0.22 (23) 0.16 (24) 97 0.03 (4) 0.01 (1) 0.04 (10) 0.12 (34) 0.11 (26) 0.12 (27) Shorthead redhorse 96 0.16 (31) 0.01 (1) 0.01 (1) 0.01 (1) 0.01 (1) 0.03 (3) 0.01 (1) 97 0.17 (22) 0.02 (4) 0.02 (4) < 0.01 (1) 0.04 (13) 0.02 (4) < 0.01 (1) Plains minnow 96 0.28 (28) 0.09 (18) 0.37 (37) 0.05 (6) 0.03 (4) 97 0.77 (178) 0.30 (75) 0.03 (6) 0.31 (77) 2.24 (511) W. silvery minnow 96 0.04 (4) 0.01 (1) 0.04 (4) 0.01 (1) 0.39 (51) 0.03 (3) 0.17 (26) 97 0.04 (8) 0.42 (96) 0.17 (41) 1.05 (233) 0.06 (16) 0.36 (88) 0.93 (212) Sand shiner 96 0.33 (66) 0.10 (1) 0.01 (1) 0.16 (16) 97 0.38 (50) 0.05 (9) 0.01 (3) 0.04 (9) 0.07 (16) < 0.01 (1) 0.34 (84) Red shiner 96 0.01 (1) 0.19 (19) 0.02 (2) 0.01 (2) 97 0.07 (13) 0.01 (3) < 0.01 (1) Goldeye 96 0.01 (2) 0.04 (4) 0.01 (1) 0.01 (1) 0.05 (7) 97 0.15 (19) 0.02 (4) 0.01 (3) < 0.01 (1) White sucker 96 0.07 (13) 97 0.08 (10) 0.01 (1) Fathead minnow 96 0.02 (3) 0.01 (1) 97 0.02 (2) 0.01 (1) 0.03 (6) < 0.01 (1) Creek chub 96 0.01 (1) 97 0.01 (1) Common carp 96 0.01 (1) 97 0.01 (1) < 0.01 (1) Longnose dace 96 0.03 (3) 0.01 (1) 0.01 (1) 0.02 (2) 0.02 (2) 97 0.02 (5) 0.01 (3) River carpsucker 96 0.03 (5) 0.03 (3) 0.05 (7) 0.05 (5) 0.01 (2) 97 0.01 (2) < 0.01 (1) 0.03 (6) Sturgeon chub 96 0.02 (2) 0.05 (6) 0.03 (3) 97 0.03 (7) 0.01 (2) 0.01 (3) 0.01 (3) Stonecat 96 0.01 (1) 0.01 (1) 0.01 (1) 97 < 0.01 (1) 0.07 (20) < 0.01 (1) Black bullhead 96 0.01 (1) 97 < 0.01 (1) Freshwater drum 96 0.01 (1) 0.01 (1) 0.08 (12) 97 < 0.01 (1) White bass 96 0.01 (1) 0.07 (11) 97 0.01 (3) 0.03 (7) Sauger 96 0.02 (2) 97 0.01 (3) Spottail shiner 96 0.01 (1) 0.01 (1) 0.03 (5) 97 < 0.01 (1) Emerald shiner 96 0.01 (1) 97 < 0.01 (1) 0.29 (65)

*Other species (n) caught over both years at Reach 1 were smallmouth bass (15), plains topminnow (6), green sunﬁsh (2), and bluegill (1). Additionally, plains killiﬁsh (3) were caught at Reach 2, yellow bullhead (1) at Reach 3, and largemouth bass (6) at Reach 7 in 1997. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 25

Table A2. Catch per unit effort (CPUE) of ﬁshes caught in trap nets in the Cheyenne River, South Dakota during 1996-1997 expressed as ﬁsh per net night (number) by species and reach.

Common name Year Reach

1 2 3 4 5 6 7 8 9

Channel catﬁsh 96 1.00 (2) 0.50 (1) 1.50 (3) 0.70 (3) 2.00 (6) 0.50 (2) 2.50 (5) 97 1.00 (4) 0.25 (1) 1.00 (4) 2.00 (8) 1.00 (4) 1.50 (6) 2.25 (9) Shorthead redhorse 96 3.50 (7) 2.50 (5) 0.50 (2) 0.50 (1) 1.33 (4) 0.50 (1) 97 0.50 (2) 1.00 (4) 6.00 (24) 1.00 (4) 1.50 (6) 0.50 (2) River carpsucker 96 2.00 (4) 0.33 (1) 0.25 (1) 97 0.50 (2) 0.25 (1) White sucker 96 0.50 (2) 97 0.25 (1) Common carp 96 0.25 (1) 97 0.25 (1) 0.25 (1) Goldeye 96 0.50 (1) 97 0.50 (2) 0.25 (1) 0.25 (1) Sauger 96 0.50 (1) 0.50 (1) 0.50 (1) 0.33 (1) 97 0.25 (1) White bass 96 0.50 (1) 0.25 (1) 0.50 (2) 97 Freshwater drum 96 0.50 (2) 97 0.25 (1) Stonecat 96 0.67 (2) 1.50 (6) 97 0.25 (1) 0.25 (1) 0.25 (1)

Flathead chub 96 97 0.50 (2) 1.25 (5) Red shiner 96 0.50 (1) 97 Spottail shiner 96 0.25 (1) 97 Northern Pike 96 0.50 (1) 97 Yellow bullhead 96 97 0.25 (1) Black bullhead 96 97 0.25 (1) Smallmouth bass 96 97 0.25 (1)

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 27

IMPACTS OF THE JOHN MORRELL MEAT PACKING PLANT ON MACROINVERTEBRATES IN THE BIG SIOUX RIVER IN SIOUX FALLS, SOUTH DAKOTA

Craig N. Spencer, Gwen Warkenthien, Steven F. Lehtinen, Elizabeth A. Ring, and Cullen R. Robbins Biology Department Augustana College Sioux Falls, South Dakota 57197

ABSTRACT

We measured significant changes in the macroinvertebrate community in the Big Sioux River downstream from the John Morrell meat packing plant in Sioux Falls, SD. We used a biomonitoring approach, placing artificial substrates in the river at sites above and below the plant allowing colonization by macroinvertebrates over a 37-day period during the Fall of 1996. We documented a classic water pollution response below the plant, evidenced by declines in the more sensitive macroinvertebrate taxa together with increases in the more tolerant taxa. The density of Trichoptera declined significantly at all three sites below the plant, falling to 6.7 organisms per sampler at the most downstream site 3.5 km below the plant, compared to average densities of 59- 91 organisms per sampler at upstream reference sites. By contrast, the density of Chironomidae increased more than three-fold at the site immediately below the plant, reaching a mean of 185 organisms per sampler compared to 33-67 organisms per sampler at the reference sites. The total biomass of all macroinvertebrates was significantly lower at all three sites below the plant, compared to the upstream sites, suggesting that impacts on the invertebrate community were caused by toxicity problems in the effluent rather than changes in food availability. The John Morrell plant in Sioux Falls seems to have satisfied requirements for water quality monitoring and effluent discharge standards during the time period of this study. However, results from our research lead us to conclude that existing monitoring programs, conducted by regulatory agencies as well as point source dischargers, are not sufficient to protect water quality in South Dakota or to ensure that existing state antidegradation regulations are met. We urge adoption of biomonitoring studies as routine complements to existing water quality monitoring efforts.

INTRODUCTION

In February of 1996, John Morrell and Company of Sioux Falls, South Dakota, pleaded guilty to polluting the Big Sioux River (Walker, 1996). Officials from the meat packing plant adrnitted to filing false discharge monitoring re- 28 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) ports as well as violating the Clean Water Act, including repeatedly violating the company's permit for discharge of ammonia into the Big Sioux River. John Morrell and Company has since been sold by its parent company, Chiquita Brands International Inc., to Smithfield Foods Inc., and efforts have been made to improve wastewater treatment at the plant (Pieterick,1997). The present study was initiated to evaluate the impact of current operations at the John Morrell plant on water quality in the Big Sioux River. This large meat packing plant processes over 10,000 hogs per day and discharges an average of 2.2 million gallons (8.3 million L) of treated wastewater per day into the Big Sioux River. We utilized a biological monitoring approach to evaluate the impacts of this wastewater through a comparison of the aquatic macroinvertebrate community living in the Big Sioux River above and below the plant. Biomonitoring of organisms in their natural environment is becoming increasingly used for water quality assessment as a supplement to the more traditional physical and chemical approaches (Rosenberg and Resh, 1996). Cur- rent monitoring requirements for John Morrell and Company, as well as other large pointsource dischargers in South Dakota, rely upon analysis of grab samples of sewage effluent for water quality characteristics and in certain instances determination of toxicity using standard test organisms in laboratory studies. Analyses of grab samples describe conditions that existed at the time of sample collection and, as such, may miss critical discharge events. Goodnight (1973) concluded that the most important sampling times for water quality assessment were not during typical, average conditions, but rather during “extreme conditions,” which may be missed by periodic sampling. Such extremes, however, will not be “missed” by the biota living in the receiving stream, especially the more sensitive taxa. Biomonitoring offers another advantage in that monitoring of the natural biota in the receiving stream largely eliminates the difficult step of attempting to use laboratory data to accurately predict impacts of sewage effluent on biota in the actual stream environment. Furthermore, biological responses may be elicited in the natural stream biota at chemical concentrations below analytical detection limits or after chemical exposure has ceased (Rand et al., 1995). Thus, the natural stream biota can serve as continuous monitors of water quality, subject to both acute and chronic impacts of water pollutants. Macroinvertebrates are the most frequently used group of aquatic organisms used in biomonitoring (Hellawell,1986). Macroinvertebrates offer several advantages over other aquatic biota. “They are ubiquitous and, consequently, are affected by perturbations in many different aquatic habitats; and, the large number of species exhibit a range of responses to environmental stress” (Rosenberg and Resh, 1996). Furthermore, macroinvertebrates are relatively long lived and generally sedentary, unlike fish, which may move out or return as water quality conditions fluctuate (Welch, 1980). Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 29

STUDY SITE AND METHODS

The Big Sioux River originates in northeastern South Dakota and flows through predominately agricultural land and small towns before passing through the city of Sioux Falls. At Sioux Falls, the river has a drainage area of approximately 13,509 km2. The John Morrell meat packing plant is located along the river in the northeastern part of the city. Five monitoring sites were selected near the plant (Fig. 1). Two reference sites were located 100 and 300 meters upstream from the effluent release point, and the other three sites were located 100, 1050 and 3500 meters downstream from the plant. Sites were carefully chosen based upon homogeneity of water velocity, depth, and substrate characteristics. Mean water velocity at the five sites ranged from 0.23 to 0.28 meters per second, and depths ranged from 40 to 50 centimeters. The substrate at each site was primarily sand with lesser amounts of silt and gravel. The macroinvertebrate community was quantified at each site using standard multiple plate samplers (Hester and Dendy, 1962). Figure 1. Map of the study area showing the These samplers were cho- five study sites along the Big Sioux River. Ef- sen because they were easy fluent from the John Morrell meat plant en- to use and they provided ters the river between site two and site three. identical substrate conditions for all study sites. We recognize that macroinvertebrate samples from the natural substrate would likely have yielded a higher diversity of organisms; however, the heterogeneous nature of the riverine environment in our study area would have necessitated the collection of large numbers of replicate samples from each site in order to allow meaningful statistical comparisons to be made among the sites. In addition, some of the study areas included deep, fast flowing waters that would have provided an additional challenge with respect to quantitative sampling of the natural substrate. Each Hester-Dendy sampler was secured in the river on a steel rod driven into the river sediments. The samplers were attached to the rods approximately 10 cm above the river bottom. On September 30,1996, four replicate samplers were placed at each of the five study sites. All samplers were collected on November 5, after 37 days in the river. Upon retrieval, individual samplers 30 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) were placed in plastic bags, preserved in 70% ethanol, and taken back to the laboratory for analysis. The samplers were disassembled and scrubbed with brushes to remove all invertebrates and material (aufwuchs) attached to the artificial substrates. Larger macroinvertebrates were handpicked from the samples, and smaller organisms were removed with the aid of a dissecting microscope. The organisms were identified, enumerated, and dried to constant mass to determine biomass. Most organisms were identified to the genus level based upon taxonomic keys by Merritt and Cummins (1996), Pennak (1989), and Wig- gins (1996). The Chironomidae were identified only to the family level due to time constraints. The Oligochaeta were identified only to class level because the segmented worms became fragmented during the course of preservation and sample processing, making further classification problematic. Statistical comparisons among sites were made using one-way analysis of variance.

RESULTS AND DISCUSSION

Data from this study provide evidence that the John Morrell meat packing plant continues to have an adverse impact on water quality in the Big Sioux River. This impact is evidenced by shifts in the species composition and abundance of macroinvertebrates colonizing samplers located downstream from the plant. Over 96% of the macroinvertebrates collected at each site were from one of three taxa: Trichoptera, Chironomidae, and Oligochaeta. Among these three groups, the Trichoptera (caddis flies) are widely reported to be the least tolerant of poor water quality (Gaufin andTarzwell,1956;Hynes,1960; McGanigle and Lucey,1983). The abundance of Trichoptera declined significantly below the plant, falling from mean densities of 59 and 91 organisms per sampler at the reference sites to 29 organisms per sampler at site three, 100 m downstream from the wastewater outfall pipe (Fig. 2a). Mean densities were further depressed at the lower two sites, falling to 9.7 Trichoptera per sampler at site four and finally 6.7 organisms per sampler at site five. The abundance of Tri- choptera at site five, located 3.5 km downstream from the plant, was 10-fold lower than densities found at the two reference sites upstream from the effluent pipe. The Chironomidae (midges) exhibited a response nearly opposite to the Trichoptera. (Fig. 2b) As a group, the Chironomidae are well known for their tolerance of polluted conditions (Gaufin and Tarzwell,1956; Hynes,1960; Mc- Garrigle and Lucey, 1983), and the density of these organisms increased significantly to 185 organisms per sampler at site three below the plant, more than threefold higher than upstream reference densities of 33-67 Chironomidae per sampler. Unlike the Trichoptera, the Chironomidae returned to background densities relatively quickly. Densities declined to 92 and 27 organisms per sampler at sites four and five, respectively, and these densities were not significantly different from abundances measured at the reference sites. The Oligochaeta (aquatic earthworms), like the Chironomidae, include a number of species that are tolerant of poor water quality (Goodnight and Whit- ley, 1961). Oligochaetes were relatively common at sites three and four below Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 31 the plant (35 and 45 organisms per sampler, respectively), and those densities were comparable to mean densities found at the reference sites of 28 and 49 organisms per sampler. (Fig. 2c) However, the density of Oligochaetes declined to 15 organisms per sampler at the most downstream site, which was significantly lower than the density measured at any of the other sites. Although the absolute density of Oligochactes did not increase below the plant, their relative abundance did. In- creased relative abundance stemmed largely from the significant decline in Trichoptera numbers below the meat packing plant. The net result is that the relative abundance of the- more pollution tolerant taxa, the Oligochaeta and Chironomidae, was very high at the three sites below the sewage outfall, ac- counting for 85 to 90% of all Figure 2. Mean abundance of the three macroinvertebrates on the sam- most common macroinvertebrate taxa: plers, compared to 30 and 66% (a) Trichoptera, (b) Chironimidae and (c) at the two reference sites (Fig. Oligochaeta, colonizing artificial sub- 3). strates at five sites along the Big Sioux The samplers were colo- River. The first two sites were located nized by small numbers of oth- just upstream from the John Morrell meat er macroinvertebrates from a packing plant while the remaining three variety of additional taxa in- sites were downstream from the plant. cluding Amphipoda, Gastropo- Error bars represent standard errors da, Corixidae, Ephemeroptera, around the mean. The asterisks indicate Plecoptera, Diptera (Simuli- that the sites were significantly different idae), and Coleoptera. Taken (p<0.05) from the two upstream refer- together, these taxa accounted ence sites as determined by one-way for less than 4% of all macroin- analysis of variance. vertebrates found on the samplers and their densities were too low to allow meaningful statistical comparisons to be made among the various sites. 32 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Other studies have documented similar changes in the macroinvertebrate community in response to the discharge of wastewater into rivers and streams. In studies of a secondary wastewater treatment plant in Virginia, Kondratieff and Simmons (1982) reported that the macroinvertebrate community in a stream below the outfall was dominated by a few tolerant taxa from the Chirono- midae and Oligochaeta. Up- stream sites were more diverse Figure 3. Mean abundance of all of the and included pollution sensitive various macro invertebrates colonizing aquatic insects in the orders Tri- the artificial substrates at five sites along choptera, Ephemeroptera, and the Big Sioux River. The first two sites Plecoptera. In a classic early were located just upstream from the John study, Gaufin and Tarzwell Morrell meat packing plant while the re- (1956) reported a dramatic remaining three sites were downstream duction in macroinvertebrate from the plant. Error bars represent stan- abundance and species diversi- dard errors around the mean for total ty in an Ohio stream receiving abundance. The asterisks indicate that municipal wastewater. Again, the sum of all the abundances of the var- the more sensitive taxa declined ious macroinvertebrates at these sites below the outfall, and the re- were significantly different (p<0.05) from maining depauperate macroin- the two upstream reference sites as de- vertebrate community was termined by one-way analysis of vari- dominated by a few tolerant ance. The horizontal scale on this figure taxa, including the Oligochaeta. is not continuous. These and other studies of sewage effluent typically report a zone of greatest impact below the outfall followed by a “recovery zone” extending some distance downstream (Hynes, 1960). Data from our study suggest that the wastewater impact zone extends at least 3.5 km downstream from the meat packing plant (Fig. 2a). The more sensitive Tri- choptera showed no evidence of recovery at this location. Additional study sites located further downstream would be necessary to quantify the total extent of the impact and recovery zones.

Mechanisms

The most obvious explanation for the alteration of the macroinvertebrate community documented in this study is a decline in water quality in the Big Sioux River owing to the input of wastewater from the meat packing plant Nev- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 33 ertheless, there are several other factors that should be considered. First, a diversion canal joins the Big Sioux River between sites three and four (Fig. 1). Influx of water from this flood control structure could have altered water quality in the river, which in turn may have affected the invertebrate community. We offer several reasons why it is unlikely that the diversion canal affected our results in a significant way. First, our data indicate significant impacts on the macroinvertebrate community beginning at site three, which lies upstream from the canal and thus remains unaffected by the water from the canal. Second, although the canal could have affected conditions at sites four and five, it is doubtful that water in the canal, which bypasses Sioux Falls, would be of lower quality than the river water flowing through the city. In fact, we expected that this diversion canal might actually ameliorate conditions in the river, by providing additional dilution water to the river. The only other factors we identified that could have influenced water quality conditions in the vicinity of our study sites are several storm drains that enter the river between sites three and five. We do not believe that these drains, which carry surface run-off to the river during rain storms, had a significant impact on the macroinvertebrate community in our study. First we never observed water flowing from these drains during our repeated visits to the river. Even if there was some flow of water through these drains, we believe the potential impact of these inputs would have been minor compared to the 2.2 million gallons per day of wastewater that enters from the meat packing plant. Fi- nally, significant impacts on the macroinvertebrate community were clearly evident at site three which is located upstream from the storm drains and thus outside of their influence. Taken together, all available evidence points to the meat packing plant effluent as the principal cause of the impact on the macroinvertebrate community documented in our study. In his book Ecological Effects of Wastewater, Welch (1980) describes several reasons why the discharge of organic wastewater may lead to characteristic downstream changes in the macroinvertebrate community. First, the increasing severity of the physical and chemical environment below the outfall may eliminate intolerant macroinvertebrate species, allowing the remaining tolerant species to flourish. Second, the effluent may stimulate growth of some taxa due to a more favorable food supply such as organic material in the wastewater or stimulation of microbial growth in the river by nutrients in the effluent. Increased growth of these taxa may result in decreased abundance of other taxa through interspecific competition for resources such as space. Evidence from our study points to a toxicity problem rather than a change in food availability. If the principle mechanism for alteration of the invertebrate community had been due to increased availability of food resources from the efffluent, then we would have expected a stimulation in the growth of macroinvertebrates below the outfall. Since biomass is a better indicator of food availability than the previously described data on abundance, we looked for evidence of enhanced growth by comparing the biomass of macroinvertebrates found above and below the outfall. The total biomass of macroinvertebrates did not increase below the plant; rather, it was significantly reduced at all three 34 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

sites below the outfall compared to upstream reference sites (Fig. 4). Thus, it is unlikely that a more favorable food supply was responsible for the observed alteration of the macroinvertebrate community. Instead, we suggest that the shift in species composition and abundance of macroinvertebrates documented below the outfall was most likely caused by toxicity problems in the effluent. We did not conduct chemi- Figure 4. Mean biomass of all of the var- cal or physical analyses of the ious macroinvertebrates colonizing the effluent from the plant; there- artificial substrates at five sites along the fore we can only speculate as to Big Sioux River. The first two sites were the possible nature of toxic ma- located just upstream from the John Mor- terial(s) in the effluent. We sug- rell meat packing plant while the re- gest several possibilities for maining three sites were downstream consideration. First, reduced from the plant. Error bars represent stan- dissolved oxygen concentra- dard errors around the mean for total tions caused by excessive or- biomass. The asterisks indicate that the ganic material (BOD) in sum of all the biomasses of the various wastewater efffluent may have macroinvertebrates at these sites were significant impacts on aquatic significantly different (p<0.05) from the biota. Oxygen depletion prob- two upstream reference sites as deter- lems were relatively common in mined by one-way analysis of variance. rivers receiving sewage emuent The horizontal scale on this figure is not prior to development of mod- continuous. ern secondary waste treatment facilities (Gaufin and Tarzwell,1956; Welch,1980). However, the waste treatment facility at the John Morrell plant includes secondary treatment, and with proper operation, the final effluent should not produce oxygen depletion problems in the river. Another possibility is excess ammonia in the effluent, which can be toxic to aquatic organisms (Russo, 1985). The John Morrell plant has had problems with elevated ammonia concentrations in the past (Walker, 1996). In addition, most state and federal water quality criteria for ammonia are based upon concentrations of unionized ammonia (NH3), with little concern for the + more common, ionized, form (NH4 ). A recent study indicates that the latter form can be the more toxic form to some macroinvertebrates (Ankley et al., 1995). A third possibility is chlorine or chlorinated organic compounds, which also can be toxic to aquatic organisms (Horne and Goldman,1994; Rand, 1995). These potential toxins can result from the wastewater disinfection process. The Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 35

final effluent at the John Morrell plant is first chlorinated and later dechlori- nated before the wastewater is discharged into the river. It is possible that toxic, chlorinated organic compounds are being produced in the effluent immediately prior to the dechlorination process, or that excess residual chlorine is present in the final effluent. In addition to these possibilities, there could be some other toxic factor(s) unknown to us, that could be present in the wastewater.

Ecosystem Implications and the Fishery

Shifts in the species composition and abundance of macroinvertebrates below the John Morrell meat packing plant are indicative of a deterioration of water quality in the river. In addition to serving as indicators of water quality, the macroinvertebrates are important in their own right as members of the aquatic food web. They are important in the diet of most fishes, often serving as critical intermediate links in the food web between the higher trophic levels and the algae and organic matter at the bottom of the food web. The present study documents impacts on the macroinvertebrate community in the Big Sioux River in Sioux Falls. Evidence from other studies indicates that the fish community in the Big Sioux River also becomes altered below Sioux Falls (Sinning, 1968; Dieterman et al., 1996). Above Sioux Falls, the river supports moderate and balanced densities of walleye and northern pike, whereas densities of these two species are much reduced below Sioux Falls. Common carp and channel catfish serve as the primary gamefish below Sioux Falls. These studies discuss a number of factors that may explain this shift in the fish community below Sioux Falls including lack of deep water habitat, increased siltation, lack of sand/gravel substrate, and decreased water quality. It is important to note that the fishery studies discussed above are broad studies of the entire river rather than specific studies focusing on the John Mor- rell plant, as was the case in our research. If declining water quality does contribute to the observed shift in the gamefish community below Sioux Falls, then numerous sources of water pollution could be involved in addition to the John Morrell plant, such as the Sioux Falls municipal wastewater treatment plant as wel1 as numerous other point and non-point pollution sources along the river.

Upstream Water Quality Not Ideal

The present study provides evidence indicating that effluent from the John Morrell plant is impacting the Big Sioux River ecosystem; however, the river can hardly be considered to be in pristine condition upstream from the meat packing plant. Macroinvertebrates in the orders Ephemeroptera and Plecoptera, which are generally considered to be indicators of good water quality, accounted for less than 1% of the organisms found on samplers at the two reference sites above the meat packing plant. Furthermore, the Trichoptera that were relatively abundant at our reference sites were entirely from the genus Hydropsyche. Among Trichoptera, the non-case building Hydropsychidae, in- 36 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) cluding Hydropsyche spp., are the most tolerant members in this group (Hawkes, 1962; McGarrigle and Lucey, 1983; Gaufin and Tarzwell, 1956). We did not find any of the more sensitive, case-building Trichoptera at our reference sites. In addition tothe present study, there is other evidence of water quality problems in the river above the John Morrell plant. A 1994 assessment of water quality in South Dakota indicated that much of the Big Sioux River “is not supporting its fishable/swimmable beneficial uses” (DENR, 1994). The report indicates that the upper portion of the river above Sioux Falls has problems with excessive total suspended solids, excessive concentrations of un-ionized ammonia, and elevated fecal coliform concentrations. The report also notes a trend of “noticeable nutrient enrichment” for many water bodies in the Big Sioux River Basin. Sources of these problems are identified as discharges of municipal wastewater from a number of communities located along the Big Sioux River and nonpoint source pollution from agriculture, including stream bank erosion and runoff from farmsteads, feedlots, and animal holding sites. Furthermore, the city of Sioux Falls takes approximately 50% of its drinking water from the river just before it enters the city, and water treatment operators periodically have to contend with taste and odor problems from excessive algae growth in the river during low flow periods and from the flush- ing of areas of standing water into the river in early spring (Janssen, 1997). In addition, visitors traveling to Sioux Falls to view the waterfalls are often greet- ed with the sight of a brown, sediment laden river, which during high flow periods is often characterized by a large build-up of foam at the base of the falls and a strong organic odor more commonly associated with sewage aeration basins in wastewater treatment plants. Finally, a hazardous waste site was recently discovered along the river in Sioux Falls, approximately 2 km upstream from our reference sites. The site contains waste from an abandoned coal gasification plant that operated in the late 1800’s, and these wastes have apparently been leaching into the river for decades. Cleanup work on this site began during the summer of 1997 through the U.S.E.P.A.’s Superfund program. Data from our study, taken together with the observations noted above, are indicative of an already stressed riverine ecosystem that is being further im- pacted by wastewater from the John Morrell plant in Sioux Falls. We suspect that the impact documented in our study would have been even greater if the research had taken place during a drought period. The last few years have been characterized by relatively big river flows in the upper Midwest. For example, discharge in the Big Sioux River ranged from 292 to 500 cubic feet per second (cfs) during our study, well above the 26-year median annual discharge of 178 cfs for this site. These values come from the USG gaging station located at North Cliff Ave., which lies just upstream from site four (USGS, 1997). We estimate that wastewater from the John Morrell plant (3.4 cfs) accounted for about 1% of the water volume in the Big Sioux River below the plant during the time period of our study. During low flow periods, the wastewater concentration would climb considerably. For example, USGS data from the last two decades indicate that the river discharge at this site falls below 23 cfs for 10% Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 37 of the time. Under these conditions, the John Morrell effluent would account for 15% or more of the river volume. The concentration would increase to near 50% of the river volume during the Q 7, 10 for this site (minimum flow for seven consecutive days occurring with a 10-year frequency = 7.1 cts). Thus we ex- pect that the negative impacts documented on stream biota during the fall of 1996 would be even more severe during drought periods when the stream biota are subjected to much higher concentrations of wastewater effluent.

State Water Quality Regulations and Implications

The State of South Dakota has existing water quality statutes that include specific antidegradation language. Section 74 of the State Water Quality Stan- dards contains the following provisions. “All waters of the state must be free from substances whether attributable to human-induced point source discharge or nonpoint source activities, in concentrations or combinations which will ad- versely impact the structure and function of indigenous or intentionally introduced aquatic communities... Toxic pollutants which are or may become inju- rious to public health, safety, or welfare; plant, aquatic, and animal life; or the designated uses of waters may not be present in the surface waters ofthe state” (DENR, 1991). Although some exemptions have been made by the state in ap- plying these antidegradation standards to certain water bodies of exceptionally poor quality or “to accommodate important economic or social development,” there seems to be no justification for exempting the Big Sioux River as it is classified for beneficial use as a fishable/swimmable water, a warmwater permanent fishery, and a river which provides drinking water for a number of communities in South Dakota. Results from the present study indicate to us that effluent from the John Morrell meat packing plant is at variance with antidegradation provisions of Section 74 of the South Dakota water quality standards. In contrast to our findings, records provided by the U.S.E.P.A. indicate that water quality monitoring data provided by John Morrell show that the company had no water quality violations during the time period of our study, and that they seem to be in com- pliance with their National Pollutant Discharge Elimination System (NPDES) permit for discharge of wastewater into the Big Sioux River (Heimdal, 1997). There are several possible explanations for the apparent contradiction between our findings and those reported by the company and govemmental regulatory agencies. It is possible that John Morrell and Company provided inaccurate data to governmental regulatory agencies. Although the company pleaded guilty to such activity in the past, we suggest this explanation as an unlikely possibility today for several reasons including the widespread negative publicity, mone- tary fine, and prison terms assessed in the 1996 water quality proceedings against John Morrell and Company, as well as the recent change in company ownership. In our view, a more likely explanation is that current monitoring activities mandated and or conducted by the South Dakota Department of En- vironment and Natural Resources and the U.S.E.P.A. are not adequate to detect the type of water quality degradation documented in our study. 38 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

To our knowledge, neither the company nor the various governmental regulatory agencies have conducted biomonitoring studies on the Big Sioux River in the vicinity of the John Morrell plant. In addition, there seems to be only minimal monitoring of water quality in the river through more traditional analyses of grab samples by the regulatory agencies. The inadequacy of existing monitoring efforts on the Big Sioux River is well illustrated by the fact that government regulatory agencies failed to detect any evidence of past, repeated ammonia discharge violations on the part of John Morrell and Company in Sioux Falls. These violations, which served as the basis for the recent legal action, became apparent only through self-reporting on the part of the company as it was in the process of being sold. Although there seems to be minimal monitoring of water quality in the Big Sioux River, John Morrell and Company is required to periodically monitor the quality of its effluent as part of the company's NPDES permit. As part of the permit requirements, the company routinely collects grab samples of their effluent and analyzes these samples for selected physical and chemical parameters. In addition, the company is required to collect additional grab samples on a quarterly basis for toxicity tests. These tests are conducted by an independent laboratory and include standard, 7-day toxicity tests on one fish species and one invertebrate species. Although the requirement for periodic toxicity tests on whole effluent represents an improvement over simple physical and chemical analysis of the effluent, results from our study suggest that current requirements for toxicity testing are still not sufficient to ensure the nondegrada- tion of water quality mandated by the state's water quality regulations. Again there are several possible explanations for the apparent discrepancy between the results of our macroinvertebrate study (which indicate degradation of water quality) and results of existing toxicity tests and monitoring data from John Morrell’s (which indicate no violation of their NPDES permit). First, current monitoring efforts by the company are based upon analysis of grab samples, which, as discussed in the introduction, may miss key discharge events. For example, no toxicity tests were conducted by the company during our 5-week biomonitoring study. The closest sampling date for their quarterly lab toxicity tests was mid-November of 1996, a week after our study was completed. Second, the toxicity tests currently mandated by state and federal regulations include relatively short-term, 7-day toxicity tests. As noted by Hynes (1960) in his classic book on water pollution, “One of the drawbacks of laboratory tests of toxicity is the fact that it is difficult to run experiments for long periods: thus the poisonous properties of toxins which act very slowly may be entirely overlooked.” For example, in a study of selenium toxicity on fish survival, Hamilton et al. (1986), found no significant effect after 30 days of exposure, but that fish mortality increased significantly to 99.1% after 60 days exposure to selenium compared to only 4.7% mortality in the control group. Third, the test organisms used in lab toxicity tests may not mimic the response of the natural biotic community in the river. The fish species used in the toxicity tests, the fathead minnow (Pimephales promelas), is widely recognized as a “hardy species…known for its tolerance for [adverse conditions in- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 39 cluding] high temperature, extreme turbidity, and low oxygen…During extended dry periods, the fathead minnow along with a few other species including the black bullhead (Ictalurus melas) often comprise the entire fish population in stagnant pools in prairie streams, simply because they are one ofthe few fish species that can tolerate such conditions” (Pflieger,1975). Thus, the in- herent ability of the fathead minnow to tolerate a wide variety of water quality conditions seems to make it useful as a test organism for revealing only the most severe toxicity problems. The species seems to be a poor choice as a test organism for revealing less severe impacts. The other standard test organism used in toxicity tests on the John Morrell effluent is an invertebrate species, Ceriodaphnia dubia, which seems to be more sensitive to a number of toxicants than the fathead minnow (Cooney, 1995). Although the ecology of this species is not especially well known (Cooney, 1995), there are a number of daphnid species that can tolerate relatively low dissolved oxygen concentrations (Pennak, 1989). Furthermore, cladocean zooplankton, such as Ceriodaphnia dubia, are typically found in lakes and ponds (Pennak, 1989). They are uncommon in fast flowing environments, as found below the John Morrell plant and thus are not ideal test organisms for predicting toxicity impacts on riverine species. Finally, current toxicity tests are conducted on grab samples of the plant effluent and could miss the possibility of synergetic interactions between the John Morrell effluent and other pollutants already present in the river. For example, Allen et al. (1946) demonstrated that chlorine from wastewater and or- ganics in the receiving water may interact to produce substances more toxic than the individual compounds. Obviously the macroinvertebrate community living in the river, which formed the basis of our biomonitoring study, would have been subject to these or other potential synergisms between the effluent and the river water.

CONCLUSIONS AND SUGGESTIONS

1. The present study presents quantitative evidence linking wastewater effluent from the John Morrell and Company rneat packing plant in Sioux Falls with adverse impacts on water quality in the Big Sioux River. This evidence includes a decline in the abundance of Trichoptera below the outfall together with an increase in Chironomidae. In addition, there was a significant decline in total macroinvertebrate biomass below the plant. The company should take necessary steps to ensure that their effluent does not continue to harm the biological integrity of the Big Sioux River. 2. Current monitoring efforts by government regulatory agencies together with required monitoring on the part of John Morrell and Company failed to detect any adverse impacts of the effluent on the river. We attribute this fail- ure to shortcomings in existing monitoring efforts and regulations. 3. In order to better protect our valuable multiple-use aquatic resources we feel it is imperative that water quality monitoring efforts be expanded in South Dakota. This recommendation is based not only upon present water quality conditions in the state, but also in light of future threats to water re- 40 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

sources such as the potential influx of large hog confinement facilities into South Dakota. We feel strongly that current chemical and physical monitoring activities by the regulatory agencies should be expanded and that these efforts be coupled with implementation of a new biomonitoring program using macroinvertebrate communities in their natural environment as indicators of water quality. We understand that the South Dakota Department of Environment and Natural Resources is currently working with the U.S.E.P.A. to develop some initial biomonitoring protocols for South Dakota. We ap- plaud these efforts and hope that they lead to widespread application of biomonitoring efforts on rivers and streams throughout the state as a vital complement to more traditional efforts focused on laboratory analyses of grab samples. 4. We suggest that artificial substrates may be appropriate for monitoring macroinvertebrates in some situations, as in the case of our relatively short- term study of a specific point source. In other cases, it may be more appropriate to monitor the macroinvertebrate community on natural substrates, as for example in the case of a long-term monitoring program to be used in quantifying the entire benthic macroinvertebrate community in a watershed. The latter approach would necessitate a much more involved sampling protocol and degree of effort than the use of artificial substrates due to increased sampling difficulties, increased variability resulting from the heterogeneous nature of natural benthic substrates, and the changing physical environment due to periodic, random disturbances such as floods and droughts. In some cases, it may be wise to monitor the macroinvertebrate community using both natural and artificial substrates. 5. We recognize that despite the findings of the present study, efforts have been made to improve wastewater treatment over the last several decades at the John Morrell facility and other large point sources along the Big Sioux Riv- er, most notably through the installation of modern secondary wastewater treatment facilities. These improvements have been beneficial to water quality in the river, especially with regard to dissolved oxygen and ammonia concentrations (Dieterman, 1995). As noted in the past, periodic fish kills resulting from discharge of poorly treated sewage into the river (Sinning, 1968) do not seem to be a major concern now. Further, there is some evidence that the fishery in the river has improved in recent years (Dieterman and Berry, 1995). Nevertheless, our study clearly indicates that point source pollution remains a problem in the Big Sioux River and that more improvements are needed. Other studies describe ongoing water quality problems resulting from nonpoint source pollution in the river basin (DENR,1994). Fi- nally, numerous point sources continue to discharge high concentrations of nutrients (phosphorus and nitrogen) into the Big Sioux River and other waters throughout South Dakota. Indeed, a recent study of water quality trends documents a significant increase in nitrate and phosphorus concentrations in the Big Sioux River below Sioux Falls over the last 20 years (Dieterman, 1995). By contrast, nutrient loadings to surface waters have been reduced in some other parts of the country through various means including phosphate detergent bans, installation of tertiary wastewater treatment facilities, and Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 41

non-point source control measures. In states bordering the Great Lakes various nutrient control measures have been put in practice in the last 2 decades with notable improvements in water quality (Schelske and Hodell, 1995). Thus, much additional work remains before the Big Sioux River and other surface waters in South Dakota comply with the goals of federal water quality legislation initiated over 2 decades ago with the Federal Water Pollution Control Act Amendments of 1972 (PL 92-500) and the Federal Clean Water Act of 1977 (PL 95- 217). These acts had as legislative objectives to "restore and maintain the physical, chemical, and biological integrity ofthe Nation's waters…and further, the elimination of discharges of pollutants to surface waters by 1985."

ACKNOWLEDGMENTS

We thank Steve Hamilton, Federal Ecotoxicology Research Station in Yank- ton, SD, for critically reviewing the manuscript. Data analysis and interpretation beneﬁted from useful information provided by Monica Heimdal, U.S. En- vironmental Protection Agency; Mike Burr, U.S. Geological Survey; Dale Tech, US Army Corps of Engineers; Gene Stueven, SD Department of Environment and Natural Resources; Dennis Holmes, US Attorney's Offce; Charles Berry and Dave Willis, Fisheries Coop. Unit, SDSU; and Allen Knapp, SD Game, Fish, and Parks.

LITERATURE CITED

Allen, L.A., N. Belzard, and A.B. Wheatland.1946. Toxicity to fish of chlorinated sewage effluents. Surveyor Lond., April 19. Ankley, G.T., M.K. Shubauer-Berigan, and P.D. Monson. 1995. Influence of ph and hardness on toxicity of animonia to the amphipod Hyaletia azteca. Can. J. Fish Aquat. Sci. 52:2078-2083 Cooney, J.D. 1995. Freshwater tests. Pages 71-102 IN G.M. Rand, ed. Funda- mentals of Aquatic Toxicology. Second ed., Taylor and Francis, Washing- ton, DC. DENR. 1991. South Dakota Department of Environment and Natural Resources. Surface Water Quality Standards. ARSD Chapt. 74. Pierre, SD. DENR.1994. South Dakota Department of Environment and Natural Resources. The 1994 South Dakota Report to Congress. 305(b) Water Quality Assess- ment. Pierre, SD. Dieterman, D. 1995. The influence of the Clean Water Act and tributaries on the fish community of the Big Sioux River, South Dakota. M.S. Thesis, South Dakota State University. Brookings, SD. Dietennan, D., and C.R. Beny. 1995. The distribution and relative abundance of fishes in the Big Sioux River, South Dakota. South Dakota Department of Game, Fish and Parks. Progress Report No. 95-9. Pierre, SD. Dieterman, D., C.R. Berry, and R. Doorenbos. 1996. Recreational use ofthe Big Sioux River, lowa and South Dakota. South Dakota Department of Game, Fish and Parks. Special Report No. 96-14. Pierre, SD. 42 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Gaufm, A.R., and C.M. Tarzwell.1956. Aquatic macro-invertebrate communities as indicators of organic pollution in Lytle Creek. Sewage and Industrial Wastes 28(7):906-924 Goodnight C.J. 1973. The use of aquatic macroinvertebrates as indicators of stream pollution. Transactions of the American Microscopic Society 92:1 - 13. Goodnight C.J., and L.S. Whitley. 1961. Oligochaetes as indicators of pollution. Engineering Extension Series (Purdue University) 106:139- 142. Hamilton, S.J., A.N. Paimisano, G.A. Wedemeyer, and W.T. Yasutake. 1986. Im- pacts of selenium on early life stages and smoltification of fall chinook salmon. Transactions of the 51st North American Wildlife and National Re- sources Conference 51:343-356. Hawkes, H.A. 1962. Biological aspects of river pollution. Pages 311-432 IN L. Klein, ed. River Pollution, Two Causes and Effects. Butterworths, London. Heimdal, M. 1997. Environmental Engineer, U.S. Environmental Protection Agency, Denver, CO. Personal communication. Hellawell, J.M. 1986. Biological Indicators of Freshwater Pollution and Envi- ronmental Management. Elsevier, London Hester, F.E., and J.S. Dendy. 1962. A multiple-plate sampler for aquatic macroinvertebrates. Transactions of the American Fisheries Society. 91 :420-421. Horne, A. J., and C.R. Goldman. 1994. Limnology. Second Ed. McGraw-Hill, New York. Hynes, H.B.N.1960. The Biology of Polluted Waters. Liverpool University Press, Liverpool, England. Janssen, R.1997. Superintendent Sioux Falls municipal drinking water treatment facility. Personal communication. Kondratieff, P.F., and G.M. Simmons. 1982. Nutrient retention and macroinvertebrate community structure in a small stream receiving sewage effluent., Arch Hydrobiol. 94(1):83-98. McGarrigle, M.L., and J. Lucey.1983. Biological monitoring in freshwaters. Irish Journal of Environmental Science. 2 2:1 - 18. Merritt, R.W., and K.W. Cummins. 1996., An Introduction to the Aquatic Insects of North America. Kendal/Hunt, Dubuque, lowa. Pennak, R.W. 1989. Fresh-Water lnvertebrates of the United States. JohnWiley & Sons, Inc., New York. Pieterick, C.1997. Director of Environmental Affairs, John Morrell and Compa- ny, Sioux Falls, SD. Personal communication. Pflieger, W.L. 1975. The Fishes of Missouri. Missouri Dept. of Conservation, Columbia. MO. Rand, G.M. 1995. Fundamentals of Aquatic Toxicology. Second ed. Taylor and Francis, Washington, DC. Rand, G.M., P.G. Wells, and L.S. McCarty. 1995. Pages 3-67 IN G.M. Rand, ed. Fundamentals of Aquatic Toxicology. Second ed., Taylor and Francis, Washing- ton, DC. Rosenberg, D.M., and V.H. Resh. 1996. Pages 87-97 IN Merritt, R.W. and KW. Cummins, eds. An Introduction to the Aquatic Insects of North America. Kendall/Hunt. Dubuque, lowa. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 43

Russo, R.C. 1985. Ammonia, Nitrite, and Nitrate. Page 456 IN G. Rand and S. Petrocelli, eds. Fundamentals of Aquatic Toxicology. Hemisphere Publish- ing Corporation, Washington, D.C. Schelske, C.L., and D.A. Hodell.1995. Using carbon isotopes of bulk sedimentary organic matter to reconstruct the history of nutrient loading and eu- trophication in Lake Erie. Limnology and Oceanography. 40:918-929. Sinning, J.A. 1968. Fishes of the Big Sioux River. M.S. Thesis, South Dakota State University, Brookings, SD. USGS, 1997. United States Geological Survey, Huron, SD. Walker, C. 1996. Morrell ﬁned $3 million. Argus Leader, Sioux Falls, SD, Feb 8. Welch, E.B.1980. Ecological Effects of Wastewater. Cambridge University Press, Cambridge, England. Wiggins, G.B. 1996. Larvae of the North American Caddisﬂy Genera (Tri- choptera). University of Toronto Press Inc., Toronto, Canada.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 45

WINTER SURVIVAL AND OVERWINTERING BEHAVIOR IN SOUTH DAKOTA ONISCIDEA (CRUSTACEA, ISOPODA)

Jonathan C. Wright Department of Biology Northern State University Aberdeen, SD 57401

ABSTRACT

Six species of terrestrial isopod from the sub-order Oniscidea are currently recorded from South Dakota, all being recent (19th C.–present) introductions from Europe. Although introduced isopods have become naturalized in a range of indigenous habitat types in other regions of North America, distributions in the Great Plains are highly synanthropic. Experimental data presented here shows that moderate cold-tolerance in these species and resultant winter mortality can explain observed synanthropy. Lower lethal temperatures resulting in 50% mortality (LLT50s) vary from -4.3 to -6.3ºC depending on season; only minor interspeciﬁc variation is evident. Seasonal acclimatization depress- es the LLT50 by approximately 2.0ºC in the winter, and laboratory acclimation at 20ºC and 0ºC induces a smaller through signiﬁcant shift in cold tolerance.

Nevertheless, LLT50 determinations are invariably far above winter ambient lows in South Dakota. Subnivean microsites afford some thermal insulation from ambient extremes but sub-lethal temperatures persist only at significant snow depths (>20 cm) or in subterranean refuges. Given that these species are not active burrowers, this restricts viable overwintering refuges to two essentially anthropogenic habitat types: thermally buffered microsites adjacent to heated buildings, and deep accumulations of coarse or friable material permitting vertical migration. The restriction of viable overwintering populations to such habitats is borne out by field observations. Shingle banks beside large or fast- flowing waterways permit deep vertical migration of isopods during cold periods, and are the only semi-natural habitat types known to support isopod populations in South Dakota. Cities are replete with anthropogenic examples such as compost piles and refuse dumps which facilitate overwinter survival and can explain the abundance and wide dispersal of populations typical in areas of extensive human settlement.

INTRODUCTION

At least twenty species of terrestrial isopods of the sub-order Oniscidea, popularly known as woodlice or sowbugs, have been recorded in the northern United States and Canada (Blake, 1931; Hatch, 1939; Hatchett, 1947; Jo- hhansen, 1926; Longnecker, 1924; Van Name, 1936; Walker, 1927). Although 46 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) the North American fauna includes several native oniscideans, the fauna is impoverished by comparison with Europe and the most widespread and familiar species are European introductions. Genetic distances between populations of Armadillidium vulgare indicate that the major avenues of colonization were the St. Lawrence and Mississippi valleys (Garthwaite, Lawson and Sassaman, 1993). In many regions of North America, introduced isopods have become widely naturalized in the native habitat types. In Southern Ontario, at least seven species can be found in the Carolinian woodlands, and naturalized populations are common in limestone screes and riparian woodlands (personal observation). Trachelipus rathkei has even become a widespread component of the macrodecomposer fauna in boreal forest litter in central Ontario (Walker, 1927; personal observation). Hatchett (1947) reports ten terrestrial Oniscidea in Michigan, most of which have invaded a wide range of natural habitat types, and Jass and Klausmeier (1996) have recently documented eleven terrestrial species in Wisconsin. In the Great Plains, isopod distribution and abundance differ markedly from patterns in Ontario and the Great Lakes basin. In South Dakota, only five genera and six species have been recorded: Trachelipus rathkei, Porcellionides pruinosus, Porcellio scaber, Porcellio spinicornis, Cylisticus convexus and Ar- madillidium vulgare. Of these, P. scaber and A. vulgare are currently recorded only from Sioux Falls (Dr Leland Johnson, personal communication) and Porcellio spinicornis is known only from Sioux Falls and Aberdeen. All of these genera belong to the section Crinocheta, a predominantly mesic taxon with species characterized by, among other features, a capacity for the active absorption of water vapor from sub-saturated humidities (Wright and Machin, 1993a,b). Prior to the present study, there were no recordings of Oniscidea in the Dakotas, and only scant records exist for the adjacent mid-western states (Longnecker, 1924; Richardson, 1905). This is most likely to reflect a paucity of regional interest rather than very recent introductions since three of the species (T. rathkei, Porcellionides pruinosus and C. convexus) are generally widespread. All six species are, however, strikingly synanthropic, frequently abundant in towns but virtually absent from the natural habitat types. The present author has never found animals in the mixed-grass prairie or in riparian woodlands, except where these adjoin areas of human settlement. Isopods appear similarly excluded from agricultural soils and ranchland. The only recordings of isopods from semi-natural habitat types are findings of Trachelipus rathkei in riverine shingle banks near Mobridge (shores of the Missouri) and in Spearfish Canyon in the Black Hills. The significance of this habitat is considered further in the Discussion. Synanthropic distributions may occur as a simple consequence of recent human introduction—characterized by a progressive decline of synanthropy over time—or as a result of human activities favoring the survival of particular species. Where synanthropy reflects an obligatory dependence, no dissemina- tion from regions of human settlement will occur. The abrupt disappearance of isopods at the edges of South Dakota towns indicates such an obligatory association. Recent introduction to those towns may explain the restricted ranges of Porcellio spinicornis, P. scaber and A. vulgare. Porcellio spinicornis and A. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 47 vulgare may also have a lower dispersal index by virtue of their pronounced calciphily, both species being closely associated with limestones or with anthropogenic sources of lime such as cement and mortar. But the catholic habitat preferences of the remaining species, and their extensive dispersal within areas of human settlement, strongly suggest obligatory synanthropy. Given the ability of introduced isopods to disperse many miles from human setlements in other regions of North America, distributions in South Dako- ta indicate that one or more environmental factors fall outside the species’ ranges of tolerance. Low rainfall is one possibility but this does not explain the inability of these isopods to colonize flood plains and riparian woodlands. Furthermore, these and related species have successfully colonized markedly xeric habitats in the Southwestern USA and Mexico (Miller, 1936; Warburg, 1965). A more plausible limiting factor is low-temperature tolerance as a de- terminant of winter survival. In the only studies performed to date on isopod thermal tolerance, Edney (1964a) and Tanaka and Udagawa (1993) both showed only modest cold tolerance (-1.4 to -4.6ºC) for Porcellio scaber. Sev- eral isopods have European ranges that follow northern isotherms closely indicating probable significance of low-temperature tolerance in limiting north- ward range expansion. The present study was conducted to determine the lower lethal temperatures (LLT50s) of isopod species in Aberdeen, South Dakota. Comparisons of these data with field temperature measurements in various microsites, and identification of overwintering refugia, provide a basis for testing the hypothesis that cold-tolerance is an ultimate factor explaining the synanthropy of introduced isopods in the region.

MATERIALS AND METHODS

Specimens of the three common oniscideans in Aberdeen—Trachelipus rathkei, Porcellionides pruinosus, and Cylisticus convexus—were used for study, together with commercially available Porcellio scaber (Carolina Biologi- cal) which were used for acclimation studies. Porcellio spinicornis, which is known to occur at one locality within Brown County, was not available in suf- ﬁcient numbers for study. For determinations of lower lethal temperatures

(LLT50), animals were either collected promptly from field populations, or maintained at controlled acclimation temperatures for various periods prior to study. The effect of acclimation was studied using temperatures of 20ºC and 0ºC. Nat- ural acclimatization in field populations was studied by comparing the LLT50 at different times of the year and comparing these with weekly ambient temperature records from the area meterological office. Separate sub-samples of animals were used for all cooling trials.

LLT50 determinations used a fixed cooling-warming cycle starting at 5ºC, cooling with a linear curve to the test temperature over a period of 5 hours, holding the test temperature for 10 hours, then warming at a uniform rate to 5ºC again over a further 5-hour period. The bath was subsequently maintained at 5ºC. Selection of this cooling cycle gives the data complementarity with those of Edney (1964a), and the time-course provides an approximate simula- 48 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) tion of the nocturnal cooling regimes encountered in vivo. Test animals were transferred to individual 50 ml Plexiglas vials containing moistened filter paper discs and sealed with foam plugs to reduce evaporative water loss but allow for diffusional gas exchange. Vials were immersed to a depth of 8 cm in a Cole Permer ‘Polystat’ programmable heating-refrigerating water bath containing a 1:1 mixture of ethylene glycol and water. Survival was assessed as soon as possible following the completion of the cooling cycle and was based on the re- sumption of normal locomotory ability on return to laboratory temperature. Field measurements of microclimate temperatures used a Barnant 115 Thermocouple Thermometer connected to an insulated copper-constantan thermocouple. The insulated cable was taped to a 2 mm diameter steel rod with the terminal 2 cm of the thermocouple projecting beyond the tip. This probe was used to measure subterranean and subnivean temperatures at various depths and at a range of distances from north-facing walls of heated buildings. Temperature recordings were confined to areas where isopods are active during the summer months. Winter recordings of ambient temperatures and soil temperatures were also obtained from the US National Weather Service, the National Climate Data Center, and the High Plains Climate Center (University of Nebraska at Lincoln, NE). Microsites were studied throughout the fall, winter and early spring months to assess which sites were used as overwintering refugia, and to observe the viability of those populations in the spring.

RESULTS

A representative plot of survivorship as a function of cooling temperature is shown in Figure 1. Isopod populations showed minor intraspecific variation in cold tolerance, with mortality increasing abruptly over a small temperature range (0.2–0.4ºC). This facilitates accurate determination of LLT50s without pro- bit transformations, and these could readily be assessed from survivorship plots to within 0.05ºC. Field-collected populations of the three study species revealed only modest cold tolerance, with LLT50 values ranging from -4.8 to -6.3ºC in early spring (cold-acclimatized), and -4.3 to -4.5ºC in summer populations. Collective data for Cylisticus convexus and Trachelipus rathkei are shown in Figure 2 (a,b). Animals were not collected during the winter months, primarily because of the difficulty of finding them, but specimens active in the early spring revealed significant depression of the LLT50 (by approximately 2.0ºC) compared with animals field-collected in the summer months (Cyclisticus convexus: t = 4.7, p < 0.001; Trachelipus rathkei: t = 5.4, p < 0.0001). Laboratory acclimation studies revealed a small but significant effect, with

1-week acclimation at 20ºC and 0ºC resulting in LLT50s differing by 1.0–1.2 ºC in both Porcellio scaber and Cylisticus convexus. Results are summarized in

Table 1. Long-term acclimation (>1 month) to 20ºC elevated the LLT50 only slightly more than a 1-week acclimation period. Daily high and low ambient temperatures for the winter months in years 1995-1996 in Aberdeen, South Dakota, are shown in Figure 3. Ambient daily maxima and minima routinely fell below observed LLT50s for isopods during the Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 49

Figure 1. An example of a survivorship curve (Porcellionides pruinosus) showing the % survival of animals following controlled cooling to different temperatures and the sharply deﬁned LLT50, in this case approximately -3.25ºC.

Figure 2. Seasonal variation in LLT50 showing acclimatization of cold tolerance in ﬁeld-collected populations of Cylisticus convexus (a) and Trachelipus rathkei (b). Cold tolerance shows a long-term acclimatory depression of approximately 2ºC between the mid-summer and winter months. Variation in LLT50 is closely correlated with seasonal changes in the high and low ambient temperature. 50 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 1. Determinations of LLT50 in isopods following laboratory acclimation to 20ºC and 0ºC. All determinations were made from 4-8 trials using a minimum of 10 animals per cooling cycle.

Species Acclimation regime 20ºC (>1 month) 20ºC (1 week) 0ºC (1 week)

Porcellio scaber -4.0 -4.2 -5.1, -5.3 Porcellionides pruinosus -3.25 - - Cyclisticus convexus -4.0 -4.2 -5.4

Figure 3. Daily high and low temperatures for the study area (Aberdeen, South Dakota) for 1996. (Data, courtesy of the High Plains Climate Data Center). months of November through March. Soil temperatures, even at modest depths, reveal much smaller fluctuations. Soil temperature records from fixed- site thermistors (20 cm depth) at Redfield, South Dakota, are plotted in Figure 4. Data show winter soil temperatures (November through March) and mean monthly ambient lows. Soil temperatures at 20 cm never fell below -2.6ºC over this period. The significance of snow as an insulator to the litter fauna, and the effects of heated buildings on local microclimate, are shown in Figure 5 (a-c). These plots show temperature recordings for different subnivean depths and various distances from heated buildings in Aberdeen during three mornings in the win- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 51

Figure 4. Soil temperatures for Redﬁeld, South Dakota, for the winter months (November-March) from 1993-1997. All measurements were from ﬁxed-site thermistors at a depth of 20 cm. Monthly mean ambient lows are also plotted for the winters of 1993-5 (dark bars).

A B Figure 5. Subnivean temperatures recorded with a thermocouple probe at various distances from North-facing walls of heated buildings (x axis) and at various depths (z axis) in March 1997. Temperatures were measured in depth increments of 5 cm from 0 cm to 25 cm which was the deepest snow cover during any of the measurement periods. Snow depth data illustrate the insulating effect of snow cover and the abrupt temperature elevation (to approximately 0ºC) at the C base of exterior walls. 52 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) ter of 1996-7. All measurements were made in shaded locations in the proximity of north-facing walls. While snow exerts a pronounced insulating effect, it is clear from these data that, during the winter months, sub-lethal temperatures persist only at signiﬁcant subnivean depths (>10 cm) or within a few centimeters of the walls of heated buildings. Field observations early and late in the year revealed a number of overwintering refugia. While summer populations use a wide range of microsites as diurnal retreats, winter populations show a very restricted distribution. Overwintering sites which supported viable spring populations could be clas- siﬁed under two basic categories: deep accumulations of friable or coarse materials permitting deep vertical migration; and shallow retreats around the perimeters of heated buildings. Examples of the former are municipal dumps, abandoned tips or rubble piles, and deep accumulations of garden compost or leaf litter. Possible overwintering refugia around the perimeters of buildings are much more widespread and include marginal cracks, stone or gravel margins, and essentially any retreat permitting vertical migration of a few centimeters at a wall edge. One overwintering site, comprising a lagged stone border and adjacent crevices at the North edge of a large house (ca. 1904), supported large populations of Porcellionides pruinosus and Cylisticus convexus in the winters of 1995-6 and 1996-7. Renewed surface activity in the Spring of 1997 was observed on March 30 at an ambient temperature of 5ºC following an overnight low of -3ºC. Renewed surface activity of C. convexus at an abandoned, con- crete-lined dump was observed on March 31 as soon as the overlying snow accumulations had melted. Abundant air spaces among the soaked elm litter and underlying refuse provided refuges for animals away from melt water. In the same site, subnivean activity was observed beneath remnant patches of snow cover. Certain microsites, in both late fall and spring surveys, showed mass mortality of overwintering populations. In some cases, these were beneath rocks or wood, at considerable distances from buildings, and lacked coarse underlying substrates which could have permitted vertical migration. In other sites, mass mortality was noted in deeper refugia, apparently the result of drowning during snow-melt.

DISCUSSION

It is evident from observations of fall and spring populations that onis- cidean populations in South Dakota possess only very modest cold-tolerance. These findings concur with those of Edney (1964a), and of Tanaka and Uda- gawa (1993) who reported LLT50 values for Porcellio scaber of -1.37ºC in Au- gust and -4.58ºC in December. The rather lower values reported in the present study for Cylisticus convexus, Porcellionides pruinosus and Trachelipus rathkei, may be the result of selection. Although these species are only recent introductions, their critical dependence on thermally buffered winter refuges—and the significant winter mortality evident in field populations—would impose a strong directional selection pressure for increased cold-tolerance. Oniscideans Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 53 survive sub-freezing temperatures by supercooling, and are not freeze-resistant (Tanaka and Udagawa, 1993). Cold acclimation in P. scaber is associated with the accumulation of low molecular weight carbohydrates during the winter months which may function as cryoprotectants preventing regional freeze-injury. Composition and millimolar accumulation of cryoprotectants would con- stitute a likely basis for selection in South Dakota populations. More profound supercooling, widespread in insects and arachnids (see Franks, 1985, for a useful review), may be precluded by other factors. Tanaka and Udagawa (1993) showed that Porcellio scaber populations in Hokkaido maintain a whole-body supercooling point (the limit of freeze-avoidance) of approximately -7.0ºC throughout the year which they attributed to ice-nucleation by gut contents. Although the onset of freezing was not assessed in the present study, it is interesting that our animals also retained gut contents during cooling trials and

LLT50 values never fell below -7.0ºC. However, when experimental animals were fasted prior to cooling to clear the gut, there was no evident depression of the LLT50.

Acclimation at 20ºC in the laboratory elevated the LLT50 signiﬁcantly above that of summer- acclimatized populations (p < 0.01; 2-sample t-test comparing data for all species) in spite of the fact that mean ambient summer temperatures exceed 20ºC. This supports the well established point that acclimatization is a complex process, depending on the range and incidence of temperature variation as well as the mean temperature over the acclimatization period. In the only other studies to date exploring acclimation of cold-tolerance in terrestrial isopods, Edney (1964a) obtained respective LLT50 estimates for P. laevis and A. vulgare of -2.4ºC and -2.7ºC following 14 days acclimation at 10ºC, -0.7ºC and -1.7ºC following 14 days acclimation at 20ºC, and +5.5 and +3.0ºC following 14 days acclimation at 30ºC. Edney’s isopods were collected from the University of California at Riverside. The fact that long-term acclimation at

20ºC did not elevate the LLT50 above -3.3ºC in the three species examined in the present study again suggests the likely impact of selection (the P. scaber used in the present study came originally from populations in Burlington, NC).

It is clear that ambient winter lows in South Dakota fall far below the LLT50 for any of the study species—exceptionally by as much as 40ºC. Although snow cover affords significant thermal insulation, prolonged temperatures below -10ºC readily depress the ground temperature below LLT50s, even with 80 cm of snow cover. Fixed-site thermistors in Bismark, ND, indicate that with more modest snow cover (12 cm) the frost line occurs at soil depths over 140 cm, and animals would need to burrow below 20 cm to evade lethal temperatures. With a few exceptions, such as the desert species Hemilepistus reau- muri, oniscideans are not active burrowers and can only migrate vertically in coarse or friable substrates. Winter temperatures in the Dakotas will therefore preclude them from the indigenous prairie soils and compacted agricultural soils. These predictions are borne out by field observations. Overwintering isopod populations in South Dakota are apparently confined to two major habitat types: deep accumulations of coarse or friable material permitting deep vertical migration, and thermally buffered microsites around the edges of heated 54 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) buildings. Even at depths of only a few centimeters, crevices at the perimeter of a heated building remain at or above freezing temperatures. Deep accumulations of coarse material allow isopods to escape from the confines of heated buildings but are of essentially anthropogenic origin. They include leaf and compost piles, dumps and rubble piles, and landfills. One naturally occurring example is the coarse shingle banks built up by large or fast-flowing waterways and which support adventitious populations of Trachelipus rathkei in two known locations. Coarse materials not only permit vertical migration, but also provide for drainage of snow-melt. Terrestrial oniscideans generally only survive submergence for a few hours at most (Taylor and Carefoot, 1993) and drowning during snow-melt is probably a significant contributing factor to overwinter mortality; recently drowned populations are a common sight in the early spring. Following the spring emergence of overwintering populations, terrestrial isopods may disperse quite widely. Summer populations beneath wood and rocks have been found over 50 meters from the closest known overwintering sites, and it is likely that isopods disperse much more extensively than this. Re- turn to overwintering refugia in the fall could involve gradual movements toward a ‘preferred temperature’ and resultant aggregation in deep or locally heated microsites. Aggregation is largely olfactory (Kuenen and Nooteboom, 1963) and is augmented by one or more short-range pheromones (Takeda, 1980, 1984). The exaggerated effect of olfactory cues in large populations may alleviate ‘stranding’ of remnant populations in unsuitable locations. Once amassed, vertical migration to avoid lethal cooling presumably involves movement up the temperature gradient, whether by a thermokinetic or thermotactic orienting mechanism (Cloudsley-Thompson, 1952, 1977). The Crinocheta appear to have originated in the Mediterranean (Vandel, 1960), and species ranges have been greatly extended by man. Natural ranges of species used in the present study do, however, extend into Northern Europe. Why they should not have evolved the remarkable supercooling capacities of many temperate insects and arachnids is unclear. One possibility is that viable overwintering habitats in low temperature extremes are dictated as much by water balance requirements as by thermal tolerance. The integumental per- meabilities of oniscideans are substantially higher than those of most insects and arachnids (Wright and Machin, 1993a; Hadley, 1994). A typical value is 0.8 µg h-1 cm-2 kPa-1 which would result in water losses of approximately 1% h-1 for a 15 mm (80 mg) animal in 50% RH at 0ºC. Such a flux would result in lethal desiccation within 30 hours. During the summer months, diurnal refugia will tend to be cooler than the ambient air and thus maintain elevated humidities. In overwintering habitats the opposite is true. Furthermore, the ability of isopods to replenish water losses by active absorption of water vapor in elevated humidities (Wright and Machin, 1993a,b) will be greatly reduced or eliminated at low temperatures. This is a consequence both of metabolic depression (Edney, 1964b; Wieser, 1984) and the reduction in vapor density (‘absolute humidity’) at low temperatures. If one assumes a Q10 of 1.72 (data from Ed- ney, 1964b), metabolism would be depressed 4.3-fold on cooling from 20ºC to -5.0ºC. When this is combined with the approximately 6-fold decline in vapor Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 55 density over this temperature range, the overall capacity for water recovery by active water vapor absorption would be reduced by a factor of twenty six. It is thus clear that maintenance of near-saturated humidities by evaporation from local water or ice will be an essential feature of overwintering microsites. The vapor pressure of ice is very similar to that of supercooled water at the same temperature (Wagner et al., 1994) and animals resting beneath snow cover will not be exposed to significant vapor pressure gradients. In this respect, as well as serving as a thermal insulator, snow cover is likely to facilitate winter survival. It does however impose the dangers of drowning during spring melt. Water-balance requirements would preclude isopods from overwintering in thermally harsh above-ground habitats, such as bark or rock crevices, which are exploited by many insects and arachnids. These findings provide an explanation for the extreme synanthropy of oniscideans in South Dakota.

ACKNOWLEDGEMENTS

I extend my thanks to Melissa Albers and Michael Harlow who provided technical assistance during various stages of this work. I am also indebted to Matt Werner of the High Plains Climate Data Center, Lincoln, NE, and Sara Walker of the Sioux Falls Metereological Ofﬁce, for supplying me with the regional soil temperature data.

REFERENCES

Blake, C.H. (1931). New land isopods from New England. Occasional Papers of the Boston Society of Natural History 5: 341-348; 349-355. Cloudsley-Thompson, J.L. (1952). Studies in diurnal rhythms II. Changes in the physiological responses of the woodlouse Oniscus asellus to environmental stimuli. Journal of experimental Biology 29: 295-303. Cloudsley-Thompson, J.L. (1977). The Water and Temperature Relations of Woodlice. Shildon: Meadowﬁeld Press Ltd. Edney, E.B. (1964a). Acclimation to temperature in terrestrial isopods. I. Lethal temperatures. Physiological Zoology 37: 364-377. Edney, E.B. (1964b). Acclimation to temperature in terrestrial isopods. II. Heart rate and standard metabolic rate. Physiological Zoology 37: 378-394. Franks, F. (1985). Biophysics and Biochemistry at Low Temperatures. Cam- bridge University Press, UK Garthwaite, R.L., Lawson, R., and Sassaman, C. (1993). Population genetics of Armadillidium vulgare in Europe and North America. In Terrestrial Isopod Biology (M.A. Alikhan, ed.). Crustacean Issues 9: 145-199. Hadley, N.F. (1994). Water Relations of Terrestrial Arthropods. Academic Press, New York. Hatch, M.H. (1939). Records of terrestrial Isopoda or sow bugs from North America. American Midland Naturalist 21: 256-258. Hatchett, S.P. (1947). Biology of the Isopoda of Michigan. Ecological Mono- graphs 17: 47-79. 56 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Jass, J., and Klausmeier, B. (1996). Terrestrial isopods (Isopoda: Oniscidea) of Wisconsin. The Great Lakes Entomologist 29: 11-20. Johhansen, F. (1926). On the woodlice (Oniscoidea) occurring in Canada and Alaska. Canadian Field Naturalist 40: 165-167. Kuenen, D.J. and Nooteboom, H.P. (1963). Olfactory orientation in some land isopods (Oniscoidea, Crustacea). Entomologica Experimental and Applied 6: 133-142. Longnecker, M. (1924). The terrestrial isopods of Iowa. Proceedings of the Iowa Academy of Sciences 30: 197-199. Miller, M.A. (1936). California isopods of the genus Porcellio with descriptions of a new species and a new subspecies. University of California Publica- tions in Zoology 41: 165-172. Richardson, H. (1905). A monograph on the isopods of North America. Bulletin of the US National Museum 54: 1-727. Takeda, N. (1980). The aggregation pheromone of some terrestrial isopod crus- taceans. Experientia 36: 1296-1297. Takeda, N. (1984). The aggregation phenomenon in terrestrial isopods. Sym- posium of the Zoological Society of London 53: 381-404. Tanaka, K. and Udagawa, T. (1993). Cold adaptation of the terrestrial isopod, Porcellio scaber, to subnivean environments. Journal of Comparative Phys- iology B 163: 439-444. Taylor, B.E. and Carefoot, T. H. (1993). Terrestrial life in isopods: evolutionary loss of gas- exchange and survival capability in water. Canadian Journal of Zoology 71: 1372-1378. Van Name, W.G. (1936). The American land and freshwater isopod Crustacea. Bulletin of the American Museum of Natural History 71: 1-535. Vandel, A. (1960). Isopodes Terrestres. Faune de France 64: 1-416. Wagner, W., Saul, A., and Pruss, A. (1994). Journal of Physical Chemistry Ref- erence Data 23: 515. Walker, E.M. (1927). The woodlice or Oniscoidea of Canada. Canadian Field Naturalist 41: 173-179. Warburg, M.R. (1965). The microclimate in the habitats of two isopod species in southern Arizona. American Midland Naturalist 73: 363-375. Wieser, W. (1984). Ecophysiological adaptations of terrestrial isopods: a brief review. Symposium of the Zoological Society of London 53: 247-265. Wright, J.C. and Machin, J. (1993a). Atmospheric water absorption and the water budget of terrestrial isopods (Crustacea, Isopoda, Oniscidea). Biological Bulletin 184: 243-253. Wright, J.C. and Machin, J. (1993b). Energy-dependent water vapor absorption (WVA) in the pleoventral cavity of terrestrial isopods (Crustacea, Isopoda, Oniscidea): evidence for pressure cycling as a supplement to the colliga- tive uptake mechanism. Physiological Zoology 66: 193-215. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 57

FLUCTUATIONS IN DAILY ACTIVITY OF MUSKRATS IN EASTERN SOUTH DAKOTA

Joel F. Lyons, Craig D. Kost, and Jonathan A. Jenks Department of Wildlife and Fisheries Sciences, South Dakota State University Box 2140B, Brookings, SD 57007

ABSTRACT

Daily muskrat (Ondatra zibethicus) activity was assessed in the fall of 1995 adjacent to the Big Sioux River, in south central Brookings County, South Dakota. Activity was determined using direct observation and trapping. Muskrats were observed from 7 to 28 October and trapped from 6 to 10 and 16 to 21 November. During ice-free periods muskrats were active during nocturnal and crepuscular hours, and were rarely observed during diurnal hours. Peak activity occurred between 0401 and 0500 hours. Trapping data collected during the onset of ice-over showed a signiﬁcant increase in diurnal activity. Our results support other ﬁndings that indicate that muskrats display a bimodal activity pattern and that they modify their activity patterns from nocturnal during summer to diurnal during winter.

INTRODUCTION

Seasonal movements and migrations of muskrats (Ondatra zibethicus) have been well documented (Errington, 1939; Sather, 1958; Boutin and Birken- holz, 1987). Migrations generally occur in spring and fall or during times of overcrowding or drought (Errington, 1939). Muskrats are known to be territo- rial (i.e., family groups actively defend their habitations within their home range) and dispersing juveniles tend to remain close to their natal territories (Sather, 1958). Muskrats are predominantly nocturnal, but may occasionally be observed during the day in spring and fall (Boutin and Birkenholz, 1987). MacArthur (1980) reported that muskrats in Manitoba, Canada displayed a bimodal trend in activity during summer, with periodic activity throughout the day, and peak levels of activity occurring between sunset and sunrise. During winter, muskrats also displayed periodic daytime activity, but a bimodal activity pattern was less evident. Peak winter activity was observed from late afternoon to early evening. Bimodal daily and feeding activity patterns of muskrats were documented by Van Horn (1975) in Wisconsin. During ice-free periods, peaks in activity occurred one to two hours prior to sunrise. After ice-over he documented an increase in the proportion of daylight activity, with the peak in activity occurring shortly after sunset. 58 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Sokolov et al. (1983) recorded change in muskrat cardiac rhythm in Rus- sia, using a radiotransmitter, to determine daily activity levels. They determined that muskrats displayed a 3-phase activity pattern. During ice-free periods, the ﬁrst activity period was one or two hours prior to sunset, continuing until sunrise. Diurnal activities were greatest from 0601-0800 to 1001-1100 hours and from 1201-1400 to 1601-1800 hours, second and third phases, respectively. After ice-over, muskrats retained a 3 phase activity pattern, but the major activity period occurred during daylight hours. The objective of our study was to determine activity patterns of muskrats in eastern South Dakota during the fall using direct observation and trapping. We hypothesized that muskrat activity, for a 24 hour period, would be equally distributed over time for observation and trapping periods.

STUDY AREA

Our study site was located adjacent to the Big Sioux River in south central Brookings County, South Dakota (Fig. 1). The area has a continental climate with an annual average rainfall of 59.5 cm and an average annual temperature of 5.78 C (Hadeen, 1994). Muskrat habitat at the study site consisted of a small stream that ﬂowed into a Class IV (Stewart and Kantrud, 1971), partially drained palustrine wetland, 0.75 km in length. Cattails (Typha sp.) and common reed (Phragmites australis) were the dominant vegetation at the study site. Tree species, such as cotton- wood (Populus deltoides) and peachleaf willow (Salix amygdaloides), occur adjacent to wetlands in the region.

METHODS

Activity levels of muskrats were determined by direct observation, from 7 to 28 October 1995. Viewing occurred at four established observation Figure 1. Location of study site adjacent to points around the wetland. the Big Sioux River in south central Brook- Two observation points ings County, South Dakota. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 59 were from trees and two were from points on an adjacent road. The wetland was monitored for a total of 24 hours. Viewing times ranged from one to three hours per day and continued for three weeks until every hour in the 24 hour period had been monitored. Viewing sites were alternated every half hour to avoid location bias. Activity levels were determined by the number of muskrat sightings recorded during the different observation periods. Trapping of muskrats was initiated on 6 November, two days after the opening of the 1995 trapping season. Bodygrip, colony, and leghold traps were set from 6 to 10 and 16 to 21 November at the study site and at other wetland locations along the travel route from Brookings to the study site. Traps were set in locations that required muskrats to leave their dens or houses and exhibit activity to be captured. Traps were checked three times a day; 0800 (midnight to sunrise activity), 1600 (daylight activity), and 2400 (sunset to midnight activity) hours. Traps were checked at these times because sunrise and sunset occurred between 0701-0800 and 1901-2000 hours, respectively. This allowed midnight to sunrise and daylight activity period traps to be checked during daylight hours. Relative frequency of muskrat observations and harvest rates were compared within and across time periods using Pearson chi-square tests (Wilkin- son, 1990). RESULTS

Two hundred seven muskrat sightings were recorded during the study: 123, 0001-0800; 6, 0801-1600; and 78, 1601-2400 hours. Hourly sightings ranged from a maximum of 32 muskrats between 0401-0500 hours to a minimum of 0 muskrats for periods 0801-1200 and 1501-1700 hours. Little (i,e., ≤8 muskrats) to no activity was observed during daylight hours (Fig. 2).

Figure 2. Hourly muskrat sightings by observation over a 24 hour period, from 7 to 28 October 1995 at a wetland in south central Brookings County, South Dakota. 60 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

We observed a reduction in muskrat activity at 2400 hours. Ten muskrat sightings were recorded for each of the two one-hour periods following midnight (0001-0100 and 0101-0200 hours). At 0201 hours, activity gradually increased to the daily high of 32 muskrat sightings, which occurred between 0401-0500 hours. After 0501 hours, activity gradually declined until sunrise, 0701-0800 hours, when observations increased to 19 muskrat sightings. The only observed muskrat activity during daylight hours occurred between 1201- 1500 and 1701-1900 hours. Maximum number of muskrat sightings recorded in any period during the day was three. At sunset, 1901-2000 hours, activity increased to 16 muskrat sightings. The hour following sunset, activity decreased to eight sightings, followed by a gradual increase to 21 sightings between 2301-2400 hours (Fig. 2). One hundred four trap nights (8, bodygrip; 16, colony; 80, leghold) were recorded during the study, with a total of 29 muskrats trapped: 13, 0001-0800; 7, 0801-1600; and 9, 1601-2400 hours. A significant difference (X2=22.076, df=2, P<0.0001) was observed between observation and trapping frequencies when time periods were combined. No difference was observed between observation and trapping rates for time periods 0001-0800 and 1601-2400 hours when data were compared using 90% Bonferroni confidence intervals (24.30- 65.11%, 0001-0800 hours; 8.83-44.51%, 0801-1600; 13.57-51.77%, 1601-2400 hours). However, observation activity for time period 0801-1600 hours, which occurred prior to ice-over, differed from trapping activity which occurred during a transition to ice-over (Fig. 3). A significant difference (X2 =100.96, df=2, P<0.0001) also was observed between time periods for observation data (51.67- 66.60%, 0001-0800 hours; 0.90-6.44%, 0801-1600 hours; 30.41-45.14%, 1601- 2400 hours). No difference (X2=1.93, df=2, P=0.381) was observed between time periods for trapping data (Fig. 3).

Figure 3. Percent of muskrat observations, 7 to 28 October, and muskrats trapped, 6 to 10 and 16 to 21 November 1995, during a 24 hour period at a wetland and along a trapping route in south central Brookings County, South Dakota. (* % Ob- served signiﬁcantly differed from % trapped during 0801-1600 hour period.) Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 61

DISCUSSION

Muskrat observation (ice-free) data indicated that muskrats were active at night and inactive during the day. Our findings are supported by Van Horn (1975) and MacArthur (1980), who reported a bimodal activity pattern during ice-free periods with a majority of the activity occurring at night. MacArthur (1980) recorded peak levels in activity between sunset and sunrise. Van Horn (1975) noted that during ice-free periods a dominant peak in activity occurred one to two hours prior to sunrise. We recorded a similar occurrence three hours prior to sunrise, with a daily high of 32 muskrat sightings recorded. Sokolov et al. (1983) reported that muskrats (both captive and wild) had a 3- phase activity pattern. A distinct 3-phase activity pattern was not evident from our observation or harvest data. However, we did record a reduction in muskrat sightings during the two hour period following midnight (0001-0200 hours), which was similar to the findings of Sokolov et al. (1983). At the beginning of the 1995 harvest season, average daily temperatures declined from 8.34 to -15.01 C. Although moderately sized lentic bodies of water remained frozen throughout the harvest period, smaller lentic and lotic areas became ice-free during the day. Muskrats in our study were harvested during the ice-free to ice-over transition period and were likely modifying their activity pattern, increasing activity during the diurnal period (based on trapping). Our findings of increased diurnal activity during ice-over are supported by MacArthur (1980) who reported that peak muskrat activity during winter occurred in late afternoon and early evening. He also reported a season difference, summer versus winter, in nocturnalism of muskrats, using automated data recorded every five minutes over a 24 hour period. MacArthur (1980) also reported a reduction in the mean hourly activity index of muskrats, ([total observations per hour in which an animal is active away from a lodge or burrow / total of all observation per hour] X 100), two days after ice-over compared with two days prior to ice-over. In addition, he reported a difference in the N:D ratio of muskrats [mean hourly activity index for all hours between the hour of sunset and the hour of sunrise (N) / mean hourly activity index for all other hours (D)] between summer and winter automated data. Van Horn (1975) noted an increase in the proportion of daylight activity after ice-over when compared to previous ice-free weeks. Sokolov et al. (1983) reported that during ice-over, muskrats shifted their activity pattern from one where the major activity period occurred during nocturnal hours to one where the major activity pattern occurred during diurnal hours. However, their evaluation of the shift from nocturnal to diurnal activity was qualitative and not supported by our investigation. MacArthur (1980) proposed that reduction in winter light intensity may be the primary factor responsible for increased diurnal activity. Kooyman (1975) reported that Weddell seals (Leptonychotes weddelli) relied heavily on vision for under-ice navigation and that the frequency of dives made by seals was greater during daylight hours. Muskrat houses are occupied by many individuals during winter (Errington, 1963; MacArthur and Aleksiuk, 1979). Increased diurnal 62 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) activity patterns of muskrats during winter would ensure that houses will always be occupied, maintaining the microclimate within at a favorable temperature. This increased activity also ensures that the plunge hole would remain ice-free, allowing access into houses and feeding push-ups (MacArthur, 1978 and 1980). Ice cover also may offer protection from predators (Van Horn, 1975; MacArthur, 1980). Increased diurnal activity of muskrats during winter may aid in the maintenance of body temperatures. This may be accomplished by increasing the frequency of foraging bouts during diurnal hours. MacArthur, (1980) reported that winter feeding of muskrats is intermittent, with individuals leaving their houses for periods greater than one hour. Energy gained from these feeding bouts would likely last long enough to be of thermoregulatory benefit to muskrats, generating metabolic heat and raising metabolic rates for up to five hours post feeding (MacArthur and Campbell, 1994). Our observation technique was useful in documenting activity of muskrats during ice-free periods, but would not be useful for observing muskrats during periods of ice cover. However, the trapping technique, although somewhat difficult to use during periods of ice cover, was useful in documenting muskrat activity. During other seasons, a combination of observation, trapping, and radiotelemetry techniques would yield greater insight on seasonal fluctuations in daily muskrat activity.

ACKNOWLEDGMENTS

We thank R. A. Fletcher, D. C. Franke, and J. S. Gleason for their review of our manuscript. We also thank South Dakota State University for support of our study.

LITERATURE CITED

Boutin, S. and D. E. Birkenholz. 1987. Muskrat and round-tail muskrat. 315- 322 IN Novak, P., J. A. Baker, M. E. Obbard, and B. Mallooch (eds.), Wild Furbearer Management and Conservation in North America. Ministry of Natural Resources Ontario, Canada. Errington, P. L. 1939. Reactions of muskrat populations to drought. Ecology 20:168- 186. Errington, P. L. 1963. Muskrat populations. Iowa State University Press. Hadeen, K. D. 1994. Climatological data annual summary South Dakota. Cli- matological Data South Dakota 99:7-11. Kooyman, C. L. 1975. A comparison between day and night diving in the Weddell seal. J. Mammal. 56:563-574. MacArthur, R. A. 1979. Winter movements and home range of the muskrat. Can. Field Nat. 92:345-349. MacArthur, R. A. 1980. Daily and seasonal activity patterns of the muskrat (On- datra zibethicus) as revealed by radiotelemetry. Holartic Ecology. 3:1-9. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 63

MacArthur, R. A., and M. Aleksiuk. 1979. Seasonal microenvironments of the muskrat (Ondatra zibethicus) in a northern marsh. J. Mammal. 60:146- 154. MacArthur, R. A., and K. L. Campbell. 1994. Heat increment of feeding and its thermoregulatory beneﬁt in the muskrat (Ondatra zibethicus). J. Comp. Physiol. B. 164:141-146. Slather, J. H. 1958. Biology of the great plains muskrat in Nebraska. Wildl. Monogr. No. 2. Sokolov, V. E., V. P. Sukhov, and G. S. Sukhova. 1983. Radiotelemetric investigation of muskrat activity by change in cardiac rhythm. Dokl. Biol. Sci. 266:500-504. Stewart, R. E., and H. A. Kantrud. 1971. Classiﬁcation of natural ponds and lakes in the glaciated prairie region. U.S. Fish and Wildl. Ser. Res. Publ. No. 92. Van Horn, S. D. 1975. Activity and feeding rhythms in the muskrat, (Ondatra zibethicus). Ph.D. Thesis. Univ. of Wisconsin, Madison. Wilkinson, L. 1990. SYSTAT: The System for Statistics. SYSTAT, Inc. Evanston, IL.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 65

OCCURRENCE OF SMALL, NONGAME MAMMALS IN SOUTH DAKOTA’S EASTERN BORDER COUNTIES, 1994-1995

Kenneth F. Higgins South Dakota Cooperative Fish and Wildlife Research Unit, USGS/BRD and Rex R. Johnson, Mark R. Dorhout, and William A. Meeks Wildlife and Fisheries Sciences South Dakota State University Brookings, SD 57007-1696

ABSTRACT

Less is known about small, nongame mammals in South Dakota than other mammal groups. During 1994-95, we collected small, nongame mammals in 3 primary habitat types along eastern South Dakota’s 8 border counties. Small mammals were sampled with snap and Sherman live-traps baited with a mixture of oats, peanut butter, and bacon grease. Of nearly 2,000 specimens collected, approximately 200 were archived in the Natural History Museum at the University of Kansas in Lawrence. Sixteen species were captured of which 12 were small nongame species. Species richness did not vary among habitat types but capture rates did vary by habitat type and county.

INTRODUCTION

Considerable information is available on the occurrence and distribution of birds (SDAOU 1991, Peterson 1995), plants (Great Plains Flora Association 1977, Larson 1993), and fish (Neumann and Willis 1994) in South Dakota; however, similar data bases are incomplete for mammals, and especially for the small, nongame species. Several mammalogists have published general distribution maps for species of South Dakota mammals (Over and Churchill 1941, 1945; Hall and Kelson 1959; Chapman and Feldhamer 1982; Jones et al. 1983, 1985; Armstrong et al. 1986), but none map the occurrence of species per county. Except for Blumberg (1993), earlier distribution maps for mammal species in South Dakota were based on specimen collections in museums, a standard procedure for mammalogists but also one which has temporal limitations (Hazard 1982). Collecting and preparing representative specimens for museums or institutional repositories is also expensive. Thus, few intensive collection studies have been conducted on mammals in South Dakota since the 1960's (Blumberg 1993), and particularly for the eastern half of the state. Re- cent collections in eastern South Dakota have been conducted by Lindell (1971), Moe (1974), Searls (1974), Barnes and Linder (1982), Pendleton and Davison (1982), Pendleton (1983, 1984), Mullican (1992, 1993), Backlund (1995), Kraft (1996), Meeks (1996) and this study. In addition to collections, 66 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) several checklists of mammals of South Dakota have been compiled (Choate and Jones 1981; Houtcooper et al. 1985; and Sharps and Benzon 1984). The purpose of this study was to sample small mammals in 3 primary habitat types in eastern South Dakota’s counties bordering on Minnesota and Iowa. Information gained will help fill the information void on small mammal distributions in eastern South Dakota. The South Dakota Department of Game, Fish and Parks currently maintains a natural resource data base on the flora and fauna within the state via its Natural Heritage Program in cooperation with The Nature Conservancy (Houtcooper et al. 1985); data from this study are also important to this data base.

ACKNOWLEDGEMENTS

We thank K. Kruse, T. Rhinehart, A. Attrill, and J. Moritz for ﬁeld work assistance; J. Cooper for data entry; R. Seabloom for taxonomy assistance; and E. Dowd-Stukel and D. Backlund for overall project assistance. Funding and support for this project was provided by Federal Aid to Wildlife Restoration (W-107-R, Study No. 7557) via the South Dakota Cooperative Fish and Wildlife Research Unit in cooperation with the U.S. Fish and Wildlife Service, the U.S. Geological Survey, the National Biological Service, South Dakota Department of Game, Fish and Parks, South Dakota State University and the Wildlife Management Institute.

STUDY AREA AND METHODS

Our study area included 8 eastern South Dakota counties, from north to south, (Roberts, Grant, Deuel, Brookings, Moody, Minnehaha, Lin- coln and Union) bordering on western Minnesota and Iowa (Fig. 1). Within each county 3 fields were selected for each of 3 primary habitat types (grassland, wetland-edge, tree belts). Two additional habitats (cropland and woodland) were sampled in Deuel County. Trapping was conducted between mid-May and 31 July in 1994 and 1995 to inventory nongame small mammals. Medium- and large-sized mammals and bats were not sampled. Trap types included Museum Spe- cial, regular snap traps and small-sized Sherman- live traps. All traps were baited daily with a mixture of rolled oats, peanut butter and bacon Figure 1. The 8 county grease. Three traps were set per station (2 snap study area in eastern and 1 live trap) with 10 stations per trap line. South Dakota. One trap line was placed in each of 3 fields per Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 67 habitat type (n = 3) per county (n = 8). Data for 1994 and 1995 were pooled for analysis. Capture rates (indices of relative abundance) were calculated as the number of individuals captured/100 operable trap nights. An operable trap night was defined as any of those traps containing a specimen or those still set mi- nus those snapped and empty or those containing non-mammal species (e.g., a frog or bird). All mammal specimens were cross-referenced to data sheets by date, trap number, trap line number, habitat type, and county. Specimens were placed in plastic zip-lock bags with their individual number and cross-reference data and frozen until they could be identified to species. Representative individuals per species by county and habitat types were prepared as voucher specimens and deposited at the Natural History Museum, University of Kansas, Lawrence. Dr. Robert Seabloom of the University of North Dakota, Grand Forks, assisted with species identification verification for all specimens and W.W. Goodpaster prepared the skulls and skins for museum vouchering. We calculated mean catch rates and standard errors (SE) for all lines/habitat types/county. The General Linear Models procedure in SAS (1990) was used for analysis of variance. Multiple comparisons were performed using Duncan's Multiple Range procedure in Proc GLM.

RESULTS

A total of 2,629 mammals representing 16 species was captured over 38,365 trap nights on 250 traplines in the 8 county study area during 1994-1995 (Fig. 2). Four of the 16 species, cottontail rabbits (Sylvalagus floridanus), op- possums (Didelphus marsupialis), long-tailed weasels (Mustela frenata), and least weasels (M. nivalis), were not included in the analysis because they were classified as medium-sized game mammals. Distribution and habitats of the remaining 12 small-sized nongame species are discussed below. These 12 species are the: thirteen-lined ground squirrel (Spermophilus tridecemlineatus), white-footed mouse (Peromyscus leucopus), deer mouse (P. maniculatus), meadow jumping mouse (Zapus hudsonicus), prairie vole (Microtus ochrogaster), meadow vole (M. pennsylvanicus), western harvest mouse (Rei- throdontomys megalotis), southern red-backed vole (Clethrionomys gapperi), house mouse (Mus musculus), northern grasshopper mouse (Onychomys leucogaster), masked shrew (Sorex cinereus), and short-tailed shrew (Blarina brevicauda). Mean capture rates (captures/100 trap nights) for all species combined declined with trapping session (F=4.59, df=4, P=0.0021). Mean capture rates were 20.2 for rep 1; 19.1 for rep 2; 16.3 for rep 3; 7.1 for rep 4; and 2.3 for rep 5. Declining capture rates may reflect live trap avoidance, or depletion of the populations of vulnerable species at sites where snap traps were used. De- clines in capture rates were significant for white-footed mice (F=2.57, df=3, P=0.0540), deer mice (F=2.97, df=4, P=0.0195), meadow voles (F=7.13, df=4, P=0.0001), and house mice (F=4.42, df=2, P=0.0574). 68 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 2. Distribution of the (a) house mouse; (b) meadow jumping mouse; (c) prairie vole; (d) meadow vole; (e) southern red- backed vole; (f) northern grasshopper mouse; (g) deer mouse; (h) white-footed mouse; (i) western harvest mouse; (j) thirteen-lined ground squirrel; (k) masked shrew; and (L) short-tailed shrew.

Species richness did not differ among principal habitat types (F=0.84, df=4, P=0.5030). Mean species richness was 3.9 (±SE=0.45) for traplines in wetland edges, 3.8 (±0.37) for traplines in grasslands, and 4.4 (±0.36) for traplines in tree belts. Capture rates for all species combined varied with habitat types (F=4.36, df=4, P=0.0032) and counties (F=3.69, df=7, P=0.0018) (Table 1). Tree belts, the habitat with the highest mean capture rate (19.7 individuals/100 trap nights), and cropland, the habitat with the lowest mean capture rate (4.7 individuals/100 trap nights), differed. Grasslands (16.4), wetlands (12.1), and other woodlands (6.6) did not differ from tree belts or cropland.

House Mouse

The only exotic species sampled was the house mouse. Ten specimens were captured, representing 4 counties and comprising 0.3% of the total small mammal sample (Fig. 2a). House mice were habitat generalists, occurring with Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 69

Table 1. Mean capture rates (captures/100 trap nights) by counties and habitats for all species combined.

HABITAT Other County Wetland Edge1,2 Grassland1,2 Cropland2 Tree Belt1 Woodlands1,2

Brookingsa 7.89 ± 5.07 7.67 ± 2.72 4.46 ± 1.46 Deuelb 27.84 + 8.54 28.19 ± 5.78 4.96 ± 2.18 12.92 ± 2.74 6.58 ± 1.17 Granta,c 8.72 ± 4.16 8.82 ± 3.14 7.67 ± 1.70 Lincolnb 9.77 ± 0.61 21.53 ± 8.64 22.98 ± 6.67 Minnehahab 5.77 ± 1.35 22.38 ± 6.50 43.45 ± 11.25 Moodyb 13.04 ± 2.40 12.00 ± 7.08 31.62 ± 9.27 Robertsa,b,c 16.90 ± 4.35 13.07 ± 1.19 15.20 ± 4.88 Unionb,c 6.60 ± 2.83 16.03 ± 5.22 23.19 ± 4.95 a,b,c Counties with the same letter did not differ in small mammal capture rates. 1,2 Habitat types with the same number did not differ in small mammal capture rates.

near equal abundance, as indexed by capture rates, in all habitat types (F=2.14, df=2, P=0.1737) (Table 2).

Meadow Jumping Mouse

Meadow jumping mice occurred in all 8 counties surveyed (Fig. 2b). A total of 287 individuals were captured, comprising 10.9% of the total sample. This species was more abundant in wetland edge habitats than in other habitats (F=13.12, df=4, P=0.0001). Capture rates in wetlands were greater than in woodlands (other than tree belts) or croplands, but did not differ from capture rates in grasslands or tree belts (Table 2). Jones et al. (1983) also reported that meadow jumping mice typically inhabit moist sites with dense cover.

Prairie Vole

Prairie voles were captured in all counties except Union (Fig. 2c). A total of 46 individuals were captured comprising 1.7% of the total sample. This species commonly occupies upland prairies, although they may occur in riparian or swale habitats when meadow voles are absent (Jones et al. 1983). How- ever, capture rates were similar for all habitat types (F=1.92, df=2, P=0.1565) in our study (Table 2).

Meadow Vole

Meadow voles occurred in all 8 counties (Fig. 2d). A total of 419 individuals were captured comprising 15.9% of the total sample. Unlike prairie voles, meadow voles capture rates varied by habitat type (F=7.66, df=4, P=0.0001) (Table 2). Jones et al. (1983) noted that this species typically in- 70 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 2. Mean captures rates (captures/100 trap nights) of small mammals by habitat type.

HABITAT TYPE Other Species Wetland Grassland Cropland Tree Belt Woodlands

Omnivore S. tridecemlineatus 0.69 ± 0.14 4.50 ± 1.92 0.64 ± 0.20 P. leucopus 3.80 ± 1.34 4.74 ± 1.54 1.19 ± .a 29.81 ± 9.44 3.09 ± . P. maniculatus 4.35 ± 1.93 9.24 ± 2.73 1.11 ± . 19.74 ± 7.63 Z. hudsonicus 13.86 ± 2.45 4.34 ± 0.92 0.56 ± . 8.15 ± 3.67 1.85 ± . Insectivore S. cinereus 13.09 ± 5.07 13.26 ± 3.80 1.11 ± . 10.15 ± 2.07 5.35 ± . B. brevicauda 1.97 ± 0.89 3.81 ± 1.12 2.73 ± 1.18 1.36 ± . O. leucogaster 0.59 ± 0.13 0.71 ± . 1.05 ± . Herbivore M. ochrogaster 2.39 ± 1.79 2.45 ± 1.36 1.87 ± 0.65 M. pennsylvanicus 11.58 ± 4.43 14.81 ± 4.60 5.95 ± . 4.16 ± 0.97 0.60 ± . Granivore R. megalotis 0.55 ± 0.10 9.30 ± 2.91 1.10 ± 0.39 C. gapperi 1.35 ± . M. musculus 1.31 ± 0.72 0.82 ± 0.27 1.34 ± .

Unknown (999) 6.13 ± 2.34 3.74 ± 1.20 5.52 ± 2.09 7.48 ± . aN=1

habits moist to wet meadows dominated by lush, dense stands of grasses or sedges, characteristic of wetland edges. In this study, capture rates were higher in grasslands than in wetlands or tree belts (Table 2).

Southern Red-backed Vole

South Dakota is at the southern limit of the range of southern red-backed voles. They (n=37) were captured only in tree belts (Table 2) and only in Roberts County, the northern most edge county in eastern South Dakota (Fig. 2e). These specimens comprised 0.1% of the total sample. Southern red- backed voles are considered a woodland species (Jones et al. 1983).

Northern Grasshopper Mouse

A total of 4 northern grasshopper mice were captured in wetland edges, grasslands, and tree belts in Moody, Minnehaha, and Lincoln counties (Fig. 2f). This species is an obligate dust bather, and it characteristically inhabits sites with dry sandy or silty soils (Jones et al. 1983). Habitat associations were not detected in this study (F=3.32, df=2, P=0.1738), probably because of the small number captured (Table 2). Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 71

Deer Mouse

Deer mice occurred in all 8 counties surveyed (Fig. 2g). A total of 437 were captured, comprising 16.6% of the total sample. Deer mice are commonly considered habitat generalists (Jones et al. 1983). Capture rates varied by habitat type (F=4.97, df=3, P=0.0021) in our study, and captures were more abundant in tree belts than in wetlands, grasslands, or croplands (Table 2).

White-footed Mouse

White-footed mice were captured more often than other species in eastern South Dakota border counties. A total of 503 were trapped, comprising 19.1% of the total sample. They were captured in all 8 counties (Fig. 2h) and their capture rates varied by habitat type (F=7.28, df= 4, P=0.0001). Capture rates were greater in tree belts than in any other habitat type (Table 2). Jones et al. (1983) noted that white-footed mice on the northern Great Plains were almost invariably found in woodlands.

Western Harvest Mouse

Western harvest mice were captured in the 4 southernmost border counties (Fig. 2i). A total of 79 were captured comprising 3.0% of the total sample. A preponderance of this species was captured in grasslands, although capture rates did not differ among habitat types (F1.17, df=2, P=0.3153) (Table 2).

Thirteen-lined Ground Squirrel

Thirteen-lined ground squirrels were captured in all 8 counties (Fig. 2j). Their capture rates did not differ among habitat types (F=0.56, df=2, P=0.5766). A total of 56 thirteen-lined ground squirrels were captured, comprising 2.1% of total captures. This species commonly occupies short grass habitats with well drained soils (Jones et al. 1983). In this study a large variance in the mean grassland capture rate (± 1.92) may have masked a preference for grasslands over wetlands and tree belts (Table 2).

Masked Shrew

Masked shrews were captured in all 8 counties (Fig. 2k). A total of 468 were captured, comprising 17.8% of the total sample, and making them the second most common small mammal captured. Masked shrews prefer moist sites, but they occur in a variety of habitats (Jones et al. 1983). In this study, they occurred in all habitats, and captures/100 trap nights were similar for wetland edges, grasslands, and tree belts, which exceeded capture rates for croplands and other woodlands (F=6.43, df=4, P=0.0001) (Table 2). 72 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Short-tailed Shrew

Short-tailed shrews were captured in all 8 counties(Fig. 2L). A total of 111 specimens were captured, comprising 4.2% of the total sample. Principal habitats of this species are deciduous woodlands, although in eastern South Dako- ta, short-tailed shrews occurred in a variety of habitats. In this study, capture rates were greater than in grasslands or tree belts, but did not differ from capture rates in other woodlands (F=9.23, df=3, P=0.0001) (Table 2).

DISCUSSION

Twelve species, including 10 rodents and 2 insectivores, comprised the principal small mammal fauna of the 8 eastern South Dakota border counties. Five species (meadow jumping mouse, meadow vole, deer mouse, white-footed mouse, and masked shrew) comprised over 80% of small mammals captured. As expected, croplands supported the lowest density of small mammals. Hayslett and Danielson (1994) stated that small mammal populations are influenced more by vegetation structure than by plant community composition. Croplands that are seasonally denuded of vegetation would be expected to support impoverished small mammal communities. We did not find differences in small mammal species richness among the 3 principal habitat types. How- ever, traplines in tree belts had higher mean capture rates than traplines in other habitat types. A high occurrence of zero captures and the high degree of variability in capture rates likely limited our ability to detect siginificant differences in some data sets. Of the 12 species captured in eastern South Dakota border counties, 6 (house mouse, western harvest mouse, prairie vole, northern grasshopper mouse, masked shrew, and short-tailed shrew) were habitat generalists. A preponderance of individuals of deer mice and white-footed mice were captured in tree belts. Thirteen-lined ground squirrels were most commonly captured in grasslands (although habitat use differences were not significant), and meadow jumping mice were most abundant in wetlands. Meadow voles were nearly as abundant in wetlands as in grasslands.

LITERATURE CITED

Armstrong, D.M., J.R. Chaote, and J.K. Jones, Jr. 1986. Distributional patterns of mammals in plains states. Texas Tech University Press, Museum Occa- sional Paper No. 105. 27pp. Backlund, D.C. 1995. New records for the dwarf shrew, pygmy shrew, and least shrew in South Dakota. Prairie Nat. 27(1):63-64. Barnes, T.G. and R.L. Linder. 1982. Small mammal occurrence in South Dako- ta shelterbelts and movements of Peromyscus maniculatus. Proc. South Dakota Acad. Sci. 61:56-63. Blumberg, C.A. 1993. Use of a mail survey to determine present mammal distributions by county in South Dakota. M.S. Thesis. South Dakota State University, Brookings. 136pp. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 73

Chapman, J.A. and G.A. Feldhamer, editors. 1982. Wild mammals of North America. Johns Hopkins Univ. Press, Baltimore, MD. 1147pp. Choate, J.R. and J.K. Jones, Jr. 1981. Provisional checklist of mammals of South Dakota. Prairie Nat. 11:65-77. Great Plains Flora Association. 1977. Atlas of the ﬂora of the Great Plains. Iowa State University Press, Ames. 600pp. Hall, E.R. and K.R. Kelson. 1959. The mammals of North America. Ronald Press Co., New York, 2 Vols. (I: xxx + 1-546 + 79; 2: viii + 547-1083 + 79). Hayslett, L.A. and B.J. Danielson. 1994. Small mammal diversity and abundances in three central Iowa grassland habitat types. Prairie Nat. 26:37-44. Hazard, E.B. 1982. The mammals of Minnesota. Univ. Minn. Press, Min- neapolis, MN. 280pp. Houtcooper, W.C., D.J. Ode, J.A. Pearson, and G.M. Vandel, III. 1985. Rare animals and plants of South Dakota. Prairie Nat. 17:143-165. Jones, J.K., Jr., D.M. Armstrong, R.S. Hoffmann, and C. Jones. 1983. Mammals of the northern Great Plains. Univ. Nebraska Press, Lincoln. 379pp. Jones, J.K. and J.R. Choate. 1985. Guide to mammals of the plains states. Univ. Nebraska Press, Lincoln. 371pp. Kraft, C.K. 1996. Mammals inhabiting the land-water interface of Oak Lake as indicators of disturbance. M.S. Thesis. South Dakota State University, Brookings. 59pp. Larson, G.E. 1993. Aquatic and wetlands vascular plants of the northern Great Plains. Gen. Tech. Rep. RM-238. Fort Collins, CO: U.S. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Exper. Sta. 681pp. Lindell, J.R. 1971. Small mammal distribution in relation to marshland vegetation types in southeastern South Dakota. M.A. Thesis. Univ. of South Dakota, Vermillion. 33pp. Meeks, W.A. 1996. Nongame vertebrate survey of Sand Lake National Wildlife Refuge, Brown County, South Dakota. M.S. Thesis. South Dakota State University, Brookings. 102pp. Moe, M.S. 1974. Habitat preferences and food habits of striped skunks in eastern South Dakota. M.S. Thesis. South Dakota State University, Brookings. 34pp. Mullican, T.R. 1992. Distribution of the pygmy shrew in South Dakota. Prairie Nat. 24(4):257-259. Mullican, T.R. 1993. Survey of small mammals on state game production areas and other public lands. Final Report, Submitted to the South Dakota Game, Fish and Parks National Heritage Program. 12pp. Neumann, R.M. and D.W. Willis. 1994. Guide to the common ﬁshes of South Dakota. EC 899. South Dakota State University, Brookings. 60pp. Over, W.H. and E.P. Churchill. 1945. Mammals of South Dakota. Univ. South Dakota Mus., 56pp. (mimeographed). Over, W.H. 1941. Mammals of South Dakota. Univ. South Dakota Mus. and Dept. Zool., 56pp. (mimeographed). Pendleton, G.W. 1983. Northern pocket gopher from Clark County, South Dakota. Prairie Nat. 15:8. 74 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Pendleton, G.W. 1984. Small mammals in prairie wetlands: habitat use and the effects of wetland modifications. M.S. Thesis. South Dakota State Uni- versity, Brookings. 54pp. Pendleton, G.W. and R.P. Davison. 1982. Relative efficiency of three small- mammal traps in a prairie wetland. Prairie Nat. 14:9-12. Peterson, R.A. 1995. The South Dakota breeding bird atlas. S.D. Ornitholo- gists’ Union, Northern State University, Aberdeen, SD. 276pp. SAS Institute Inc. 1990. SAS/STAT user’s guide, version 6. Fourth edition, Vol- ume 2. SAS Institute Inc., Cary, NC. 1739pp. Searls, D.A. 1974. Influence of vegetation on the distribution of small mammals on a Waterfowl Production Area. M.S. Thesis. South Dakota State University, Brookings. 47pp. Sharps, J.C. and T.A. Benzon. 1984. A compiled list of South Dakota Wildlife. S.D. Dept. Game, Fish and Parks Rep. No. 85-11. 27pp. South Dakota Ornithologists’ Union. 1991. The birds of South Dakota: Second Edition. Northern State University Press, Aberdeen, SD. 411pp. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 75

USE OF A MAIL SURVEY TO DETERMINE PRESENT MAMMAL DISTRIBUTIONS IN SOUTH DAKOTA

Carmen A. Blumberg Department of Wildlife and Fisheries Sciences

Kenneth F. Higgins USGS/Biological Resources Division South Dakota Cooperative Fish and Wildlife Research Unit and Jonathan A. Jenks Department of Wildlife and Fisheries Sciences South Dakota State University Brookings, SD 57007

ABSTRACT

A state-wide mail survey was conducted in 1992 to determine the present distribution by county, of 42 mammals. Questionnaires (n = 834) were sent to natural resource managers, trappers, and others with knowledge of the distribution of mammals in South Dakota. Return rate of questionnaires was 37%. Species distribution maps were developed to compare past species distributions to mail survey results. Ranges of nine mammals studied expanded, seven decreased, and 13 were similar to those previously published for the species. Temporal changes in species distribution were uncertain for 13 species because of questionable sightings reported by survey respondents and voids in the literature and museum records. Study results indicated that with some modiﬁcations, a mail survey showed potential as an effective technique to assess the distribution of medium- to large-sized mammals by county.

INTRODUCTION

Except for a few ungulate species (e.g., white-tailed deer; scientific names occur in Table 1) that are annually inventoried to set hunting seasons, information on the current distribution of mammals in South Dakota is generally lacking (Wilhelm et al. 1981). Most publications (Chapman and Feldhamer, 1982; Hall and Kelson 1959; Jones et al., 1983, 1985; Armstrong et al. 1986) that address statewide distribution of mammals provide distribution maps for each species, but none map the occurrence of species by county. Benefits of determining mammal distributions by county are: (1) natural resource professionals usually manage areas by county; and (2) this method enables a more accurate description of species distributions as, for example, there are 66 counties but only 12 physiographic regions in South Dakota. Species distribution maps in earlier accounts from South Dakota were derived mostly from museum specimen collections and records in published lit- 76 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) erature, a standard procedure for accumulating such data (Hazard, 1982:16). However, this procedure has two major limitations; one is the lack of a data set based on one simultaneous, statewide collection effort and the second is the difficulty in accessing museum specimens, which are dispersed across the Unit- ed States. Furthermore, many collection trip diaries or logs of specimens are in repositories located across the United states and many collections are not computerized. Past findings also indicate that the sole use of specimen collections to document mammal occurrence may misrepresent present mammal distribution (Hazard 1982:16). The goal of this study was to determine the current county distribution of 42 of the most common mammal species found in South Dakota (Table 1). Study objectives were to determine the utility of a mail survey for determining mammal distributions, to identify where and when past surveys and major museum or institution collections of mammals were made in South Dakota, and to compare mail survey data to data from published literature and listings of museums or reference collections. Information from this study will identify gaps where future systematic mammal collections should be made in the state.

MATERIALS AND METHODS

Study Area

The study area included all of South Dakota, which encompasses 199,552 km2 (Fig. 1). Before 1941, South Dakota had 69 counties. Presently, the state has 66 counties as Washington County was incorporated into Shannon Coun- ty, Armstrong County became a part of Dewey County, and Washabaugh Coun- ty was added to Jackson County. Because of name changes of counties and

Figure 1. Locations of the 66 counties of South Dakota. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 77 towns since the 1800's, locations of mammal collections and observations indicated by some literature sources and museum listings were sometimes difﬁ- cult to assess. Sneve (1973) was particularly helpful for locating physiographic areas and counties or towns in South Dakota that no longer exist or have had names changed, such as the old forts and ghost towns.

Literature Search

A comprehensive literature search was initiated in 1992 for available information on mammals of South Dakota. Computer and visual searches were used to ﬁnd references on mammals of South Dakota from the Bibliography of Agriculture (1942-1992), Biosis Previews (1969-1992), Zoological Record (1978- -1992), and Wildlife Review (1935-1957, 1985-1993) indexes. Studies or documents written prior to these years were gathered by searching the literature cited sections of natural history studies, books, and documents. We corresponded with government institutions to request bibliographies and publications of available mammal documents or studies conducted in South Dakota. To determine changes in the county distribution of mammals of South Dakota from Jones et al. (1983) until this study, distribution maps for each of the 42 mammal species were xeroxed from that publication, enlarged to a standard size, and overlaid with a mylar template containing county boundaries of South Dakota. Comparisons of mammal distributions from Jones et al. (1983) and our study were used to assess temporal changes in mammal distributions.

Museum Collections

National and state museums and institutions having mammal collections were identiﬁed from the literature (Turner 1974; Yates et al. 1987), from returned questionnaires, and from the South Dakota Game, Fish and Parks De- partment. Lists of museums that have specimens from South Dakota were gathered from the South Dakota Natural Heritage Program in Pierre. Copies of lists and computer printouts of data relative to South Dakota mammals were requested from museums and institutions. In our study, a major mammal collection was deﬁned as one having specimens of more than seven small to large mammals collected by a single person or institution, during one year, and at the same location. The decision to choose seven species was arbitrary. Only collections having species covered in this study (Table 1) were included. Per- sonal visits were made to some museums and institutions in South Dakota when information could not be obtained by other means.

Mail Survey

A statewide mail survey was conducted in 1993 to determine county distributions of the 42 mammals. A pilot-survey form was reviewed by South Dakota Game, Fish and Parks personnel and other professionals relative to the publics surveyed and the objectives of the project. The survey included a cover letter explaining the study and a statement of how the results would be 78 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 1. Mammal species (n = 42) surveyed for distribution in South Dakota. Species common and scientiﬁc names are from Jones et al. (1992).

Common Names Scientiﬁc Names

1. Virginia Opossum (Didelphis virginiana) 2. Desert Cottontail (Sylvilagus audubonii) 3. Eastern Cottontail (Sylvilagus floridanus) 4. Mountain Cottontail (Sylvilagus nuttallii) 5. Snowshoe Hare (Lepus americanus) 6. Black-tailed Jackrabbit (Lepus californicus) 7. White-tailed Jackrabbit (Lepus townsendii) 8. Eastern Gray Squirrel (Sciurus carolinensis) 9. Eastern Fox Squirrel (Sciurus niger) 10. Red Squirrel (Tamiasciurus hudsonicus) 11. Northern Flying Squirrel (Glaucomys sabrinus) 12. American Beaver (Castor canadensis) 13. Common Muskrat (Ondatra zibethicus) 14. Coyote (Canis latrans) 15. Gray Wolf (Canis lupus) 16. Swift Fox (Vulpes velox) 17. Red Fox (Vulpes vulpes) 18. Common Gray Fox (Urocyon cinereoargenteus) 19. Black Bear (Ursus americanus) 20. Common Raccoon (Procyon lotor) 21. American Marten (Martes americana) 22. Ermine (Mustela erminea) 23. Least Weasel (Mustela nivalis) 24. Long-tailed Weasel (Mustela frenata) 25. Black-footed Ferret (Mustela nigripes) 26. Mink (Mustela vison) 27. Wolverine (Gulo gulo) 28. American Badger (Taxidea taxus) 29. Spotted Skunk (civet) (Spilogale putorius) 30. Striped Skunk (Mephitis mephitis) 31. Northern River Otter (Lutra canadensis) 32. Mountain Lion (Felis concolor) 33. Lynx (Lynx lynx) 34. Bobcat (Lynx rufus) 35. Elk (Cervus elaphus) 36. Mule Deer (Odocoileus hemionus) 37. White-tailed Deer (Odocoileus virginianus) 38. Moose (Alces alces) 39. Pronghorn (antelope) (Antilocapra americana) 40. Bison (buffalo) (Bison bison) 41. Mountain Goat (Oreamnos americanus) 42. Bighorn Sheep (Ovis canadensis) Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 79 used, which was brief and self-explanatory (Sommer and Sommer 1991). Re- spondents were asked if each species occurred in their county or in surrounding counties and for supporting evidence for those sightings. Subsequent to the pilot survey, a revised three-page, two-sided questionnaire was designed for the mail survey in the fall of 1992 (Blumberg 1993). The survey targeted natural resource managers and furbearer trappers because those individuals were assumed to be most knowledgeable of the state's mammals. A total of 834 questionnaires was sent statewide to the following selected personnel or targeted publics: South Dakota Game, Fish and Parks Department (n = 192), Cooperative Extension Service (n = 60), trappers (n = 394), and others (n = 188) who were recommended as having knowledge or interest of mammals in South Dakota. "Others" consisted of personnel from the U.S. Forest Service, U.S. Fish and Wildlife Service, Bureau of Indian Affairs, and individuals suggested by the survey respondents after the first mailing. Be- cause nonrespondents are often individuals who have a low interest in the subject matter (Filion 1975), the targeted publics were chosen to obtain high response rates and to avoid receiving misleading survey responses. The first survey mailing of the questionnaire was in February 1993. A second mailing was sent to trappers (n = 297) in April 1993, and to other individuals (n = 8) who did not respond to the first mailing in June 1993. In other surveys, follow-up letters have proven effective at increasing response rates (Kanuk and Berenson 1975; Linsky 1975; Benson 1988) Completed questionnaires were reviewed for obvious errors or abnormalities immediately after receiving them. Conversations with respondents were used to correct obvious errors. Questionable records that had no evidence or weak evidence (e.g., unsure of the species sighted) for sightings were deleted from the final data set. The occurrence of each species by county was mapped on state maps with county boundaries. Mammal observations dated prior to 1990 by respondents were not included in the final distribution maps but were used in the discussion of each species when they provided unique information. Respondents unwilling to complete questionnaires were deleted from the mailing list and from any data analyses. Mammal distributions were compiled as the number of counties in which each species occurred. These data were compared to those of Jones et al. 1983 using chi-square analysis with Bonferonni conifence intervals (Systat 1990). Alpha was set at 0.10.

RESULTS

Literature Search

A total of 521 references was found that contained information on mammals in South Dakota. Many earlier documents listed collections by regions, such as the Black Hills (i.e., Pennington, Custer, and Lawrence counties). References to mammal collections were found for 22 of 66 (33%) counties in South Dakota from 1857 to the present (Fig. 2). Most published accounts of the 42 species were from studies conducted in Custer (n = 45), Pennington (n = 23), Brook- ings (n = 19), and Harding (n = 14) counties. No published accounts were 80 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) found for any of the studied species for Bon Homme and Douglas counties and only one published account was found for each of 11 counties in eastern South Dakota (Fig. 3). Perkins, Meade, Ziebach, Dewey, Hand, Charles Mix, Corson, and Tripp counties are among the largest in area in South Dakota, yet only one to ﬁve published accounts of the 42 species exist for these counties. Published accounts were most frequent for white-tailed deer, mule deer, black-footed ferret, and red fox; these are either federally protected or eco-

Figure 2. County locations of major mammal collections in South Dakota obtained from the literature, 1980 to present.

Figure 3. Literature voids by county for 42 medium to large mammals in South Dakota. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 81 nomically important species in South Dakota. Except for Ashton and Dowd (1991), literature does not exist or is not speciﬁc on county of occurrence for gray squirrels, northern ﬂying squirrels, American marten, northern river otter, mountain goat, mountain lion, or wolverine.

Museum Collections

Major collections of South Dakota mammals that are still available are primarily located in 12 museums and institutions outside of South Dakota (Table 2). The largest collection (n = 4,983) of small- to large-sized South Dakota mammal specimens is located at the University of Kansas. Major collections of the 42 mammal species were mainly from the Black Hills region. From 1946 to 1987, 535 specimens of the 42 mammals were collected from 19 South Dakota counties. Most of these collections occurred in the 1960's; temporal voids in collecting occurred prior to 1940 and after 1990. Mammal specimens from nine counties are represented in major collections at seven museums other than the University of Kansas. On the basis of useable listings provided by museums, institutions, and private individuals with known collections of any South Dakota mammals, major collections (seven or more mammals collected at the same location by the same collector) of the 42 species occurred in 22 of 66 (33%) counties (Fig. 2). Most museum collections occurred from 1951-1975 (n = 46). From 1980 to the present there have been eight major collections. The Black Hills is the location of most mammal collections and studies (1856 to the present) (Fig. 2). The least amount of mammal collecting has occurred in east-central and west-central counties within the state (Fig. 2).

Mail Survey

Three hundred and nine of 834 questionnaires mailed were completed and returned (response rate = 37%) (Table 3). Most (258 of 306, 84%) questionnaire returns occurred during the ﬁrst month following the initial mailing. Members from all publics did not speciﬁcally differentiate species of certain groups of mammals: squirrels (Sciurids), rabbits and hares (Lagomorphs), and weasels (Mustelids). For example, survey results indicated that red squirrels occurred in most counties of the state, however, based on the literature and habitat requirements for these species, they only occur in counties in the Black Hills region including nearby woodlands (Table 4). Survey data on 28 of the 42 mammals selected for study were useable; the squirrels, weasels, and lagomorphs were not included in analyses. In addition, the wolverine was not included in analyses because it was not reported in South Dakota by either Jones et al. (1983) or survey respondents. Number of counties inhabited by mammals based on Jones et al. (1983) differed (X2 = 308.441, df = 27, P < 0.001) from those generated from the mail survey. Of the 28 mammal species, 7 had expanded distributions relative to species distribution boundaries reported by Jones et al. (1983), 11 had similar distributions, and 13 had reduced distributions (Table 5). 82 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 2. Chronological listing of South Dakota mammal collections by county or area based on data obtained from museums and institutions.

Date Collector Area/Cty Institution

1856 Dr. Hayden Gregory National Museum 1873 E. Coues Gregory Natural History 1887 V. Bailey Pennington (USNM) 1888 V. Bailey Roberts Stanley Custer 1894 W. H. Granger Custer Illinois Field Museum Natural History (FMNH) 1898 H. Behrens Pennington Minnilusa Pioneer to 1924 Museum (MHM)* 1903 M. Cary Gregory National Museum 1910 A. H. Howell Central Black Natural History N. Dearborn Hills (USNM) 1916 T. C. Beach Custer 1929 P. Moulthrop Central Black Cleveland Museum G. W. Phillips Hills Natural History (CMNH) 1934 V. H. Cahalane Custer Wind Cave National Park (WCNP)* 1934 A. M. Stebler Jackson University Michigan 1935 L. R. Dice Pennington Museum Zoology (UNMZ) 1943 W. H. Osgood Custer Illinois Field Museum Natural History (FMNH) 1945 J. A. King Pennington University Michigan 1946 J. A. King Pennington Museum Zoology (UNMZ) 1946-87 (Table 5) Throughout University Kansas Museum South Dakota Natural History (KU) 1954 G. Barnes Custer National Museum Natural History

1973 L. Klaudt Aurora World Wildl. Adventures* 1975 S. L. Williams Stanley Texas Tech S. H. Genoways University 1975 R. C. Dowler Hughes Museum (TTU) M. De La Fuente 1977 R. B. Wilhelm Bennett Fort Hays State University 1978 Museums (MHP) & (FHSM NH) 1981 W. W. Goodpaster Charles Mix University Illinois Museum Hughes Natural History (UIMNH) 1982 Gregory Charles Mix 1985 Stanley Day *South Dakota Institutions Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 83

Table 3. Mail survey return rates by different respondents for the mammals of South Dakota study, 1993.

Publics # Sent # Returned % Returned

South Dakota Department 192 103 53.6 Game, Fish and Parks personnel

South Dakota Trappers 394 156 39.5

South Dakota Cooperative 60 21 35.0 Extension Service personnel

Others 188 29 15.4

Total 834 309 37.1

Trappers: Mailing # Received 1 111 245

Table 4. Questionable reports of occurrence (- response) for 3 squirrel species in South Dakota from a 1993 questionnaire survey.

Species Total Reports + Response - Response

Red squirrel 151 63 88

Gray squirrel 33 22 11

Eastern fox squirrel 99 97 2

+ Response = identiﬁed correctly according to known distribution

- Response = identiﬁed incorrectly according to known distribution 84 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 5. Relative distribution changes of mammal species in 66 counties in South Dakota, between that reported by Jones et al. (1983) and this study, 1993.

Species Number of Number of % change counties of counties of occurrence in occurrence in [A–B] x 100 Jones et al. this study, 66 (1983) [A] 1993 [B]

1. Virginia Opossum 58 55 -4.5 2. Desert Cottontail 20 7 -19.6 3. Eastern Cottontail 65 65 0.0 4. Mountain Cottontail 5 9 +6.0 5. Snowshoe Hare 0 0 0.0 6. Black-tailed Jackrabbit 27 42 +22.7 7. White-tailed Jackrabbit 66 66 0.0 8. Eastern Gray Squirrel 0 -- -- 9. Eastern Fox Squirrel 65 -- -- 10. Red Squirrel 5 -- -- 11. Northern Flying Squirrel 5 6 +1.5 12. American Beaver 66 66 0.0 13. Common Muskrat 66 66 0.0 14. Coyote 66 66 0.0 15. Gray Wolf 0 11 +16.7 16. Swift Fox 55 14 -62.1 17. Red Fox 66 65 -1.5 18. Common Gray Fox 59 29 -45.4 19. Black Bear -- 2 -- 20. Common Raccoon 66 66 0.0 21. American Marten 0 8 +12.1 22. Ermine (short-tailed weasel) 5 34 +43.9 23. Least Weasel 62 35 -40.9 24. Long-tailed Weasel 66 59 -10.6 25. Black-footed Ferret 38 6 -48.4 26. Mink 66 66 0.0 27. Wolverine -- 0 -- 28. American Badger 66 66 0.0 29. Spotted Skunk (civet) 65 54 -16.6 30. Striped Skunk 66 66 0.0 31. Northern River Otter 66 7 -89.3 32. Mountain Lion -- 23 -- 33. Lynx 66 8 -87.8 34. Bobcat 66 37 -43.9 35. Elk -- 30 -- 36. Mule Deer 66 64 -3.0 37. White-tailed Deer 66 66 0.0 38. Moose -- 25 -- 39. Pronghorn (antelope) 19 47 +42.4 40. Bison (buffalo) -- 31 -- 41. Mountain Goat -- 2 -- 42. Bighorn Sheep -- 8 -- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 85

DISCUSSION

Literature Search

Although published accounts about South Dakota mammals are available from the mid-1800's to the present, information voids exist relative to certain species and geographical areas of the state. These voids also varied with respect to time, making it difficult to assess general temporal trends in species distribution patterns. Voids in mammal collection areas were apparent for most counties in the state (n = 44), but particularly for northern, central, and eastern South Dakota (Figs. 2, 3). Furthermore, there is about a two-year lag between data collection and publication during which mammal populations may change. Even with nearly 600 references, present distribution of mammal species within counties could not be determined without the use of Hall and Kelson (1959), Jones et al. (1983) and Chapman and Feldhamer (1982), which indicated a need for periodic review and synthesis of recent published literature. Published literature on South Dakota mammals also lacked information concerning where and when past surveys and museum collections were conducted. This was most often found with historical documents where reporting detail was insufficient to determine the actual collection site, or whether specimens were preserved. Several documents stated that mammals were collected, but did not mention the species. Often only common names or scientific nomenclature were included in documents and either may have changed. "Observations" of mammals, rather than the collection of specimens, were common in many documents.

Museum Collections

Museum records provided important information on past collections of South Dakota mammals. Most museums only collected small mammals (e.g., Microtus) which were not included in our study. Because only collections containing medium to large mammals were considered, major collections of all mammal species collected in South Dakota were likely underestimated.

Mail Survey

Mail surveys are one of the least understood techniques for data collection because many have not been related to scientific theory and no experimental studies exist to determine their effectiveness (Kanuk and Berenson 1975). Al- though our return rate (37%) was adequate for this study, clear empirical evidence on response rates does not exist (Babbie 1979; Filion 1980). We used follow-up methods as suggested by Kanuk and Berenson (1975), Linsky (1975), Babbie (1979) and Filion (1980). However, response rates might have been higher if three mailings had been sent to all the publics sampled. Furthermore, 86 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) higher return rates from trappers and others may have resulted if a pre-paid postage and pre-addressed return envelope had been provided. Most mammalian distribution maps are generated from museum specimen collections. Specimens are used for developing maps because this method is assumed to be objective (i.e., it substantiates the occurrence of mammals by the use of their skins and skeletal remains). Specimens provide positive proof of species identification and the exact date and place of collections (Hazard 1982:8-9). Specimens also enable researchers to investigate physical variations within and among populations (sex, age, and seasonal difference) and to roughly document geographical range expansion and reduction (Hazard 1982:8-9). Specimens also may include information on reproduction, anatomy, pelage, molting, microevolution, ectoparasitism, dental pathology, and environmental chemistry (Jones et al. 1983:6). Therefore, museum specimens are a vital tool for developing distribution maps. Problems arise, however, when specimens are the sole information source for developing distribution maps. Disadvantages of only using museum specimens are: many species are poorly represented and areas exist where species have been infrequently collected (Hazard 1982:16). A lack of information may indicate rare occurrence of widespread species or the misrepresentation of rare species (Hazard 1982:16). Furthermore, Yates et al. (1987) stated that changes in institution management and administration of collections in repositories may modify access to specimens. Additional sources of information may indicate the presence of mammals in an area. These sources may include reported sightings of live mammals, road kills (Hazard 1982:16), skeletal remains in owl and raptor pellets, feces, and tracks (Jones et al. 1983:5). Reported sightings are likely the most abundant source of information, yet there have been few attempts to standardize this method. Although there is a lack of published research on wildlife mail survey techniques, Filion (1978) and Benson (1988) suggested that social science techniques can be applied to wildlife survey methods. Mail surveys can be efficient, fast, and inexpensive when information on a large geographical area must be obtained (Sommer and Sommer 1991). Surveys are efficient because they can be objective, standardized, and quickly sent to potential respondents. The cost of mailing surveys is minimal compared to the cost of travel and time it would take to acquire this type of information from personal interviews (Sommer and Sommer 1991, Filion 1978) or field collections. Surveys also are useful when conducting descriptive studies of large populations (Babbie 1979). They can empower traditional wildlife field survey techniques by providing additional information from people who know wildlife species and where they occur. Although survey use is not a precise technique, it can indicate trends from the acquired data (Hoinville and Jowell 1978). Other researchers have used mail surveys to determine wildlife distributions, but results have been inconclusive (Boshoff 1980, Bigale and Bateman 1962). However, those studies sampled farmers and landowners in South Africa, many of which may have had limited knowledge of mammalian species. Jobman and Anderson (1981) used a mail survey targeted to natural Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 87 resource managers to assess distribution of the black-footed ferret in the Unit- ed States. Questionnaires were sent to 62 offices within the historic range of the black-footed ferret; responses were received from 60 of those offices. That study, however, did not discuss the total number of questionnaires sent and received from individual offices, nor the utility of using a survey to determine the current distribution of the black-footed ferret. The increase in the distribution of gray wolves in South Dakota indicated by our mail survey is supported by 10 mortalities that were reported from 1981 to 1992 (Light and Fritts 1994). Most (n = 8) of these individuals were ≤ 2 years of age, suggesting they were dispersers. Furthermore, the majority of these wolves were believed to have dispersed from Minnesota. The increase in the distribution of moose may have occurred for similar (i.e., dispersal) reasons as gray wolf. Jones et al. (1983) and Hall and Kelson (1959) noted that moose were rare or extinct in South Dakota. American marten were reintroduced in the Black Hills in 1980 and 1981 (Ashton and Dowd 1991) and reintroductions of bighorn sheep and elk have occurred in western South Dakota since the early 1900's (Turner 1974). Hence, increases in these species are likely related to those activities. Other increases (mountain lion, pronghorn) may have resulted from dispersal of individuals from surrounding states or for other unknown reasons. Declines in species distributions were most common for mustelids (striped skunk, least weasel, black-footed ferret, river otter) and small canids (swift fox, gray fox) and felids (lynx and bobcat). These reductions are possibly related to changes in land use patterns (e.g., swift fox), prey availability/disease (e.g., black-footed ferret) or other reasons. Mail surveys were a useful, non-destructive technique for obtaining information concerning a select group of 42, commonly known, medium- to large- sized mammals in South Dakota. Modifications that would enhance surveys should include the addition of the descriptions of species or photos to differentiate similar species of a genus (e.g., three species of weasels), use of collo- quial (local) names, and better advertisement of the mail survey prior to mailing. Our survey was limited to 42 species that were assumed to be commonly known and easily identifiable. A similar survey for small mammals that are less frequently known or seen (e.g., mice, voles, gophers, shrews, and bats) might be less effective. However, with innovative modifications to survey procedures (e.g., use of photos) and distinct selection of survey publics, mail surveys could be useful in assessing distributions of all mammal species.

ACKNOWLEDGMENTS

Financial support was provided by the South Dakota Department of Game, Fish and Parks (Federal Aid in Wildlife Restoration, P-R Project W-75-R, Study Number 7557) through the South Dakota Cooperative Fish and Wildlife Re- search Unit, U.S. Fish and Wildlife Service, U.S. Geological Survey, South Dako- ta Department of Game, Fish and Parks, South Dakota State University, and the Wildlife Management Institute, cooperating. We thank individuals and from the following organizations who participated in the mail survey: South Dakota 88 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Trappers Association, Western South Dakota Fur Harvesters, South Dakota De- partment of Game, Fish and Parks, U.S. Forest Service, Badlands National Park, Wind Cave National Park, and the South Dakota Cooperative Extension Service. We also thank national and state museums and parks as well as private individuals who devoted much time to providing information on their mammal collections. A list of cooperators is presented in Blumberg (1993). The following individuals contributed many hours of assistance to this project: J. Higgins, M. Engel, T. Symens, K. Johnson, K. Cieminski, and R. Hansen.

LITERATURE CITED

Ashton, D. E., and E. M. Dowd. 1991. Fragile legacy. S.D. Dep. Game, Fish and Parks Report No. 91.04. 55pp. Armstrong, D.M., J.R. Chaote, and J.K. Jones, Jr. 1986. Distributional patterns of mammals in plains states. Texas Tech University Press, Museum Occa- sional Paper No. 65. 27pp. Babbie, E. R. 1979. The practice of social research. Second ed. Wadsworth Publishing Company, Inc. Belmont, California 596pp. Benson, D. E. 1988. Successful methods for human dimension surveys: 1985 RSA game farmer survey example. South Africa Journal of Wildlife Re- search 18:61-64. Bigale, R. C. and J. A. Bateman. 1962. On the status and distribution of ungulate mammals in the Cape Province, South Africa. Annals of the Cape Providence Museum 2:85-109. Boshoff, A. F. 1980. Some socio-economic aspects of a bird of prey questionnaire survey. South Africa Journal of Wildlife Research 10:71-81. Chapman, J. A. and G. A. Feldhamer, editors. 1982. Wild mammals of North America. Johns Hopkins University Press, Baltimore, Md. 1147pp. Filion, F. L. 1975. Estimating bias due to nonresponse in mail surveys. Pub- lic Opinion Quarterly 39(4):482-492. Filion, F. L. 1978. Increasing the effectiveness of mail surveys. Wildlife Soci- ety Bulletin 6:135-141. Filion, F. L. 1980. Human surveys in wildlife management. Pp. 441-453 in Wildlife management techniques manual (S. D. Schemnitz, ed.). The Wildlife Society, Washington D.C. 686pp. Hall, E. R. and K. R. Kelson. 1959. The mammals of North America. Ronald Press Company, New York, Vols. I and II. 1083pp. Hazard, E. B. 1982. The mammals of Minnesota. University of Minnesota Press, Minneapolis, Minn. 280pp. Hoinville, G. and R. Jowell. 1978. Survey research practice. Heinemann Ed- ucational Books Ltd., London. 228pp. Jobman, W. G., and M. E. Anderson. 1981. Curent black-footed ferret range as indicated by questionaire survey. U.S. Fish and Wildlife Service Unpubl. Rept. 63pp. Jones, J. K., Jr., D. M. Armstrong, R. S. Hoffmann, and C. Jones. 1983. Mam- mals of the northern Great Plains. University of Nebraska Press, Lincoln. 379pp. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 89

Jones, J. K., Jr., D. M. Armstrong, and J. R. Choate. 1985. Guide to mammals of the plains states. University of Nebraska Press, Lincoln. 371pp. Kanuk, L. and C. Berenson. 1975. Mail surveys and response rates: a literature review. Journal of Marketing Research 12:440-451. Light, D. S., and S. H. Fritts. 1994. Gray wolf (Canis lupus) occurrences in the Dakotas. American Midland Naturalist 132:74-81. Linsky, A. 1975. Stimulation responses to mailed questionnaires: a review. Public Opinion Quarterly 39(1):82-101. Sneve, V. D. H. 1973. South Dakota geographic names. Brevet Press, South Dakota 639pp. Sommer, B. and R. Sommer. 1991. A practical guide to behavioral research. Third edition. Oxford University Press, New York, N.Y. 362pp. Turner, R. W. 1974. Mammals of the Black Hills of South Dakota and Wyoming. University of Kansas Miscellaneous Publications No.60. 177pp. Wilhelm, R. B., J. R. Choate, and J. K. Jones, Jr. 1981. Mammals of Lacreek National Wildlife Refuge, South Dakota. Special Publications Museum of Texas Tech University No. 17. 39pp. Yates, T. L., W. R. Barber and D. M. Armstrong. 1987. Survey of North Amer- ican collections of recent mammals. Journal of Mammalogy 68:1-76.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 91

A SURVEY OF NATURAL RESOURCE PROFESSIONALS PARTICIPATING IN WATERFOWL HUNTING IN SOUTH DAKOTA

Jeffrey S. Gleason and Jonathan A. Jenks Department of Wildlife and Fisheries Sciences South Dakota State University Brookings, SD 57007

ABSTRACT

A 3-page waterfowl questionnaire and harvest diary was sent to natural resource agency professionals (n = 357) within South Dakota prior to the 1995 - 96 waterfowl season. Our primary objectives were to determine species composition in the bag and crippling losses by natural resource agency professionals. In addition, we wanted to determine hunting activity, demographics, experience and preferences, as well as assess attitudes concerning current waterfowl management and hunting issues. One hundred and forty-one (40%) usable questionnaires were returned. Ninety-six (67%) respondents hunted ducks and/or geese during the 1995 - 96 South Dakota waterfowl season. Species composition of waterfowl bagged by respondents indicated that mallards (n = 409) and Canada geese (n = 693) sustained the highest harvest levels of those species available. A total of 318 cripples (shot but unretrieved waterfowl) was reported by respondents resulting in a crippling rate of 10.3%. A total of 732 hunter days was documented for those respondents participating in the 1995 - 96 waterfowl season. Approximately 2,307 hours were spent pur- suing waterfowl. About 50% of all respondents indicated they had ≥ 16 years of waterfowl hunting experience. Results of this questionnaire and diary will provide waterfowl managers and administrators with a better understanding of hunter activities and demographics as well as attitudes and preferences of a segment of South Dakota’s waterfowl hunters.

INTRODUCTION

Waterfowl are considered the most economically important group of migratory birds (Anonymous, 1986). Millions of consumptive and nonconsumptive users annually take to the marshes generating annual expenditures in excess of several billion dollars (Anonymous, 1986; Gray and Kaminski, 1994). In the United States, declines have occurred in duck stamp sales and the number of active adult waterfowl hunters over the last 30 years (Sharp, 1996). During the same period in South Dakota, hunter numbers have ranged from an estimated low of 21,944 (1962) to a all-time high of 40,675 (1971) (Sharp, 1996). Of the Central Flyway states, South Dakota ranks ﬁfth based on the number of active adult waterfowl hunters (x = 29,786, 1961–93) and contributes substantially to the Central Flyway’s duck and goose harvest (Sharp, 1996). A slight 92 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) increase in hunter numbers has occurred over the last 4 years and may be attributed to an overall increase in waterfowl numbers (Caithamer and Dubovsky, 1996). Surveys targeting waterfowl hunters include efforts to determine demographic characteristics and hunter preferences in certain geographic regions (Filion and Parker, 1984; Tori et al., 1988; Humburg et al., 1997), hunters’ attitudes concerning speciﬁc waterfowl regulations (Anderson and Williamson, 1990), and estimating the number of illegal waterfowl hunters and illegal activities (Gray and Kaminski, 1993; 1994). Currently, South Dakota is one of 22 states cooperating in the Harvest Information Program (HIP) with all 50 states to be included by the 1998–99 hunting season. HIP is designed to provide more reliable harvest estimates of nonwaterfowl migratory birds as well as waterfowl harvested throughout the United States (USFWS unpubl. report). Species currently monitored by HIP in addition to waterfowl include coot (Fuli- ca americana), mourning dove (Zenaida macroura), woodcock (Scolopax mi- nor), snipe (Gallinago gallinago), sandhill crane (Grus canadensis), band- tailed pigeon (Columba fasciata), gallinule (Porphyrula martinica), and rails (Family Rallidae). Despite a recent increase in the number of waterfowl hunter surveys (Scherff and Ringleman, 1994; Gigliotti, 1995), relatively limited information is available regarding waterfowl hunters in South Dakota. To address this lack of information, we surveyed natural resource agency professionals in South Dako- ta to (1) determine species composition of the harvest and crippling losses, (2) determine waterfowl hunting activity, demographics, experience, and preferences, and (3) assess attitudes concerning waterfowl management and hunting issues.

METHODS

A statewide survey of 357 natural resource agency personnel within South Dakota were polled prior to the 1995–96 waterfowl season. These individuals were employed by several agencies within the state including South Dakota Department of Game, Fish and Parks, U.S. Fish and Wildlife Service, U.S. Army Corps of Engineers, U.S. Forest Service, Natural Resource and Conservation Ser- vice, as well as faculty, staff, and graduate students in the Department of Wildlife and Fisheries Sciences at South Dakota State University (SDSU). An initial list of names and addresses was generated from a database maintained through the Department of Wildlife and Fisheries Sciences at SDSU. Natural resource agency professionals (i.e., those individuals not at SDSU) were mailed a harvest survey in the form of a daily diary and cover letter (Fil- ion, 1978) on Department of Wildlife and Fisheries Sciences letterhead (Roe- her, 1963) 1 week prior to the 1995 - 96 waterfowl hunting season. Attached to the diary was a 3-page questionnaire and postage-paid, pre-addressed envelope. Individuals at SDSU were hand delivered the same survey and questionnaire during this period. The harvest survey requested information such as date hunted, group size, hours hunted, number of waterfowl harvested by species, area hunted, and number of cripples by species. Approximately 2 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 93 weeks prior to the end of the hunting season, a sample of surveyed individuals was contacted in writing or by phone. From survey information we derived estimates for species composition of the harvest, species composition for cripples, crippling rates, and hunter effort. In addition, the self-administered mail questionnaire provided a series of Lik- ert-type statements (Tittle and Hill, 1976) and closed-response questions (Tacha, 1976) in which personnel were asked about places available to hunt, factors contributing to the historic declines in waterfowl populations, the use of different media sources regarding waterfowl populations and the upcoming hunting season, factors influencing whether or not they hunted, importance of a limit in defining the success of a hunt, favorite hunting method, and demographics. Based on career affiliation, we assumed that attitudes and responses of questionnaires returned were representative of the sample population (Filion, 1975; 1976a), therefore nonresponse bias was not calculated. Similarly, Ringle- man (1997) following results from Pierce et al. (1996) noted that "response bias in my survey, if one exists, is likely to be small." Chi-square analyses (Wilkin- son 1990) were used to determine significance of associations between respondents’opinions concerning current waterfowl management issues (Dietz, 1990; Higgins et al., 1994). We tested for differences in attitudes between respondents that participated in the 1995–96 waterfowl season (i.e., hunters) and those that did not (i.e., non-waterfowl hunters). An alpha level of 0.05 was used in all statistical comparisons.

RESULTS

One hundred and forty-two (40%) usable questionnaires were returned. Ninety-six (67%) respondents hunted ducks and/or geese during the 1995 - 96 South Dakota waterfowl season. Based on answers to the ﬁrst 2 questions regarding participation in past (i.e., 1994) and upcoming (i.e., 1995) waterfowl seasons as well as respondents not participating in the harvest survey portion of the questionnaire, we were able to separate the returned questionnaires in- to 2 groups; individuals that participated in the 1995-96 waterfowl hunting season and those that did not. Signiﬁcant differences (P ≤ 0.05) occurred for only 20.6% (n = 7) of the 34 comparisons tested between 1995-96 waterfowl hunting season participants and nonparticipants, therefore we pooled all respondents in our analyses. We used a Bonferonni correction factor to maintain our alpha level. The adjusted alpha level for 34 comparisons was 0.002.

Hunting Activity of Natural Resource Professionals

Ninety-ﬁve individuals (67.4%) indicated that they had hunted waterfowl the previous year. Approximately 75 percent of respondents (n = 106) planned on participating in the upcoming waterfowl season. We derived estimates of hunting activity (732 hunter days [x = 7.63 days/hunter]) from 96 returns. Hunter effort was greatest in October (53.1% of hunter days) (Fig. 1) but declined throughout the season. About 2,307 hours (x = 3.15 hours/trip) were 94 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 1. Chronology of hunter effort calculated for the 1995–96 South Dakota waterfowl season. spent in the ﬁeld and approximately 56% of all hunting trips were > 3 hours in duration. Group size averaged 3 hunters but ranged from 1 individual to 14. An estimated 2,195 ducks and 901 geese were bagged by surveyed groups. An estimated 1,051 ducks, 776 geese, and 4 coots were bagged by survey respondents (Table 1). Average annual harvest per respondent during the season was 10.9 ducks and 8.1 geese. Average daily success rate for respondents was 1.4 ducks and 1.1 geese. Species composition of respondent’s bags indicated that mallards (Anas platyrhynchos) (n = 409) and Canada geese (Branta canadensis) (n = 693) sustained the highest harvest levels of the available duck and goose species (Table 1). Similarly, mallards (n = 31) and Canada geese (n = 19) sustained the highest crippling losses (Table 2). However, a large number of ducks (n = 79), geese (n = 24), and unknown cripples (n = 138) were reported. Crippling rates were estimated at 10.3% with an average of 0.4 cripples/day/hunter. Nearly 71% of all hunting trips resulted in no reported cripples.

Hunter Characteristics of Natural Resource Professionals

A total of 130 males (92.2%) and 9 females (6.5%) responded to the survey and of these 91 males (94.8%) and 4 females (44.4%) participated during the 1995–96 waterfowl season. Overall, 78.7% of respondents (n = 111) were ≥ 31 years of age with 19.7% (n = 28) falling into the class interval 21–30 years of age. Forty-seven percent (n = 66) of all individuals surveyed indicated they had ≥ 16 years of waterfowl hunting experience. Residency of individual respondents was similar with nearly equal numbers falling into the categories of Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 95

Table 1. Species composition of harvested waterfowl (n = 1831) by natural resource professionals (n = 96) that participated during the 1995–96 waterfowl season.

Species Number

Ducks Mallard (Anas platyrhynchos) 409 Gadwall (Anas strepera) 178 American wigeon (Anas americana) 84 American green-winged teal (Anas crecca) 78 Blue-winged teal (Anas discors) 72 Bufﬂehead (Bucephala albeola) 68 Northern shoveler (Anas clypeata) 45 Northern pintail (Anas acuta) 24 Wood duck (Aix sponsa) 20 Redhead (Aythya americana) 18 Ring-necked duck (Aythya collaris) 18 Canvasback (Aythya valisineria) 17 Ruddy duck (Oxyura jamaicensis) 9 Common (Mergus merganser) and hooded (Lophodytes cucullatus) merganser 5 Common goldeneye (Bucephala clangula) 5 American black duck (Anas rubripes) 1

Geese Canada goose (Branta canadensis) 693 Snow goose—white phase (Chen caerulescens) 54 Snow goose—blue phase 21 Greater white-fronted goose (Anser albifrons) 6 Ross’ goose (Chen rossii) 2

Other American coot (Fulica americana) 4 towns (≤ 10,000) (33.8%, n = 48), cities (> 10,000) (35.2%, n = 50), or rural (28.9%, n = 41) areas.

Hunter Attitudes of Natural Resource Professionals

Survey participants indicated that a parent (36.6%, n = 52) or friend (28.9%, n = 41) was primarily responsible for their initiation to waterfowl hunting. Thirty respondents (21.1%) indicated they started waterfowl hunting on their own. Twelve respondents indicated a relative other than a parent introduced them to waterfowl hunting. Nearly 75% (n = 106) of respondents thought there 96 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 2. Species composition of crippled waterfowl (n = 318) reported by natural resource professionals (n = 96) that participated during the 1995 - 96 waterfowl season.

Species Number

Ducks Mallard 31 Gadwall 6 American green-winged teal 6 American wigeon 5 Blue-winged teal 2 Bufﬂehead 2 Northern shoveler 1 Common merganser 1 Unknown teal species 3 Unknown duck species 79

Geese Canada goose 19 Greater white-fronted goose 1 Unknown goose species 24

Unknown species 138

were enough places to hunt waterfowl. Public land (56.3%, n = 80) was hunted more often than private land (14.8%, n = 21). When asked to rank the factors contributing to the historic decline in waterfowl populations, respondents indicated that loss of habitat, drought, and predators were most important (Table 3). Poor management and hunting ranked fairly high with regard to waterfowl population declines. About 84% of respondents ranked state wildlife personnel (n = 120) and state/federal waterfowl reports (n = 120) as important sources of information regarding waterfowl populations and the upcoming waterfowl season (Table 4). Federal wildlife personnel (79.6%, n = 113) and Ducks Unlimited (60.6%, n = 86) ranked somewhat lower, and television ranked the lowest (34.5%, n = 49). The 4 most important factors that determined whether or not respondents participated in waterfowl hunting were time constraints, other interests, duck abundance, and cost (Table 5). The use of steel shot was the lowest of these factors. When asked whether or not a limit was important in deﬁning a successful hunt, only 7% (n = 10) responded positively while most participants (88.7%, n = 126) indicated the opposite. Nearly 70% (n = 99) of respondents used decoys as their favorite method of hunting waterfowl, whereas 17% (n = 24) favored pass shooting and jump shooting (6%, n = 8). Attitudes differed re- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 97

Table 3. Ranking of factors that may contribute to the historic decline in waterfowl populations for natural resource professionals (n = 141).

Factors Strongly Disagree Neutral Agree Strongly No Disagree Agree Response

Loss of habitat 1.4% 0.7% 2.8% 16.9% 72.5% 5.6% (n = 2) (n = 1) (n = 4) (n = 24) (n = 103) (n = 8)

Drought 2.8% 9.9% 16.9% 39.4% 22.5% 8.5% (n = 4) (n = 14) (n = 24) (n = 56) (n = 32) (n = 12)

Predators 8.5% 16.2% 26.8% 31.0% 10.6% 7.0% (n = 12) (n = 23) (n = 38) (n = 44) (n = 15) (n = 10)

Poor management 9.9% 33.8% 35.9% 10.6% 3.5% 6.3% (n = 9)

Hunting 21.8% 38.0% 22.6% 9.2% 1.4% 7.0% (n = 31) (n = 54) (n =32) (n = 13) (n = 2) (n = 10)

X2 = 421.6 d.f. = 20 P ≤ 0.0001 garding the importance of a limit in deﬁning a successful hunt (X2 = 98.0, d.f. = 4, P < 0.0001) and one’s favorite hunting method (X2 = 108.1, d.f. = 2, P ≤ 0.0001).

DISCUSSION

Response rates and sample sizes, though somewhat lower for this study (39.8% and 142, respectively) compared to other waterfowl survey studies (60%, Smith and Roberts, 1976; 77%, Gray and Kaminski, 1994; 79%, Gigliotti, 1995; 71.9%, Humburg et al., 1997) was considered adequate and representative of the sampled population. In a survey of North American waterfowl hunters, Ringleman (1997) (NAWHS) had an overall response rate of 54.5% and a duck hunter response rate of 45.0%. He also stated that “response bias in my survey, if one exists, is likely to be small with respect to questions relating to participation rates and hunter success”. Numerous participants contacted after the waterfowl season indicated they had ﬁlled out the survey but had mis- placed it. Results from our comparison of 1995-96 waterfowl hunting season participants and nonparticipants indicated few statistical differences in attitudes. Hence, we believe that results from this survey portray the attitudes of natural resource professionals in South Dakota. 98 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 4. Ranking of sources of information regarding waterfowl populations and the upcoming waterfowl season for natural resource professionals (n = 141).

Source of Never Probably Maybe Probably Deﬁnitely No Information Not Response

Waterfowl Reports 6.3% 2.1% 14.1% 30.3% 40.1% 7.0% (n = 9) (n = 3) (n = 20) (n = 43) (n = 57) (n = 10)

State Wildlife Personnel 4.2% 5.6% 8.5% 32.4% 43.7% 5.6% (n = 6) (n = 8) (n = 12) (n = 46) (n = 62) (n = 8)

Federal Wildlife Personnel 6.3% 7.8% 13.4% 33.1% 33.1% 6.3% (n = 9) (n = 11) (n = 19) (n = 47) (n = 47) (n = 9)

Ducks Unlimited 9.9% 23.2% 22.5% 26.8% 11.3% 6.3% (n = 14) (n = 33) (n = 32) (n = 28) (n = 16) (n = 9)

Newspapers 13.4% 22.5% 27.5% 23.2% 5.6% 7.8% (n = 19) (n = 32) (n = 39) (n = 433) (n = 8) (n = 11)

Other Hunters 12.7% 24.7% 24.7% 19.0% 13.4% 5.6% (n = 18) (n = 35) (n = 35) (n = 27) (n = 19) (n = 8)

Magazines 16.2% 24.7% 26.8% 20.4% 4.9% 7.0% (n = 23) (n = 35) (n = 38) (n = 29) (n = 7) (n = 10)

Local Farmers 19.0% 31.7% 16.9% 19.7% 7.0% 5.6% (n = 27) (n = 45) (n = 24) (n = 28) (n = 10) (n = 8)

Television 24.7% 33.8% 21.1% 9.2% 4.2% 7.0% (n = 38 (n = 48) (n = 30) (n = 13) (n = 6) (n = 10)

X2 = 350.9 d.f. = 40 P ≤ 0.0001

Hunting Activity of Natural Resource Professionals

Surveyed agency professionals hunted a total of 732 days (x = 7.63 days/hunter) or 2,307 hours (x = 3.15 hours/trip). Approximately 56% of all hunting trips exceeded 3 hours in duration. These results are comparable to a hunter satisfaction survey conducted in South Dakota (Gigliotti, 1995) where the number of days hunted for ducks and geese averaged 8.6 and 7.0. Simi- larly, Humburg et al. (1997) documented an average of 10.2 and 9.6 days hunted for ducks and geese in Missouri, respectively. Hunter effort during our study showed a declining trend through time with over 50% of the effort oc- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 99

Table 5. Ranking of factors that were important in determining whether or not natural resource professionals hunted waterfowl (n = 141).

Factor Strongly Disagree Neutral Agree Strongly No Disagree Agree Response

Other interests 5.6% 16.2% 23.9% 23.9% 19.0% 11.3% (n = 8) (n = 23) (n = 34) (n = 34) (n = 27) (n = 16)

Time constraints 8.5% 14.1% 21.8% 27.5% 19.0% 9.2% (n = 12) (n = 20) (n = 31) (n = 39) (n = 27) (n = 13)

Duck abundance 12.0% 21.8% 25.4% 23.9% 6.3% 10.6% (n = 17) (n = 31) (n = 36) (n = 34) (n = 9) (n = 15)

Cost 14.8% 26.8% 26.1% 14.8% 6.3% 11.3% (n = 8) (n = 23) (n = 34) (n = 34) (n = 27) (n = 16)

No place to hunt 20.4% 26.1% 21.1% 12.7% 7.8% 12.0% (n = 29) (n = 37) (n = 30) (n = 18) (n = 11) (n = 17)

No one to hunt with 32.4% 24.7% 17.6% 11.3% 3.5% 10.6% (n = 46) (n = 35) (n = 25) (n = 16) (n = 5) (n = 15)

Regulations too restrictive 24.7% 33.8% 19.7% 5.6% 4.9% 11.3% (n = 35) (n = 48) (n = 28) (n = 8) (n = 7) (n = 16)

Regulations too complex 22.5% 35.9% 14.8% 7.8% 8.5% 10.6% (n = 32) (n = 51) (n = 21) (n = 11) (n = 12) (n = 15)

Have to use steel shot 43.0% 26.8% 12.7% 4.9% 2.1% 10.6% (n = 61) (n = 38) (n = 18) (n = 7) (n = 3) (n = 15)

X2 = 219.4 d.f. = 40 P ≤ 0.0001 curring during the month of October (Fig. 1). Conversely, Gigliotti (1995) determined that in general, of those hunters with a preference, twice as many preferred the latter part of the season. Similar to our results, Martin and Car- ney (1977) documented that a large proportion of waterfowl harvested occurred during the early part of the hunting season. This high “early” season harvest may be attributed to relatively mild weather and large hunter numbers (Martin and Carney, 1977). Our average annual harvest per respondent of 10.9 ducks and 8.1 geese is somewhat higher than comparable studies. Gigliotti (1995) documented an average annual harvest per hunter at 8.5 ducks and 5.0 geese. Humburg et al. (1997) estimated the average annual harvest for Missouri waterfowl hunters at 100 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) less than 5 ducks and geese. However, our estimates of average daily success rate of 1.4 ducks and 1.1 geese compare favorably to the national average of 1 to 2 ducks (Smith and Roberts, 1976). Gigliotti (1995) determined that waterfowl hunters averaged 1.2 ducks/day during the 1994–95 South Dakota waterfowl season, which was similar to our results. Species composition of respondent’s bags generally followed Central Flyway harvest estimates (Sharp, 1996) and species composition reported by Nieman et al. (1987) with mallards and Canada geese sustaining the highest harvest levels (Table 1). Mallards and Canada geese were cited as the preferred waterfowl species by personnel participating in this survey; however, hunters in general are considered oppor- tunistic (Nieman et al., 1987) desiring shooting activity rather than being selective (Hochbaum and Walters, 1984). Crippling losses for waterfowl have been estimated at 4–5 million birds annually (Martin and Carney, 1977) with an average of 1 cripple lost for every 4 ducks bagged (Anderson and Burnham, 1976). Some authors (Hochbaum, 1947; Anderson, 1994) have argued that crippling losses may actually approach 1 cripple for every bird bagged. In most studies, crippling estimates range from 16–41% (Bellrose, 1953; Halladay, 1969; Geis and Crissey, 1973; Hochbaum and Walters, 1984). Our crippling rate estimates (10.3% [0.4 cripples/day]) may represent “actual” losses based on our sampled population or the estimates may be biased low due to survey design, a reluctance to report actual losses, or simply that respondents underestimated the number of cripples. Using "spy blind" techniques in Canada, Nieman et al. (1987) determined that hunters reported crippling 6 - 18% of waterfowl shot at while observed losses were approximately 45%. Van Dyke (1980) noted that the percentage of cripples increased as bag size increased. Species composition for cripples (Table 2) in our study generally mirrored the species composition of the harvest (Table 1). Accurate- ly assessing crippling losses remains an important component of North Amer- ican waterfowl harvest management.

Hunter Attitudes of Natural Resource Agency Professionals

Similarities existed between our survey of natural resource agency professionals and a random sample of South Dakota waterfowl hunters (n = 676) (NAWHS) with regards to respondent’s gender (94.8% versus 97.9% male), age (78% versus 57.4% ≥ 31 years old), and hunting experience (50% versus 49.2% ≥ 16 years of waterfowl hunting experience). The nationwide NAWHS results indicated that 98% of respondents were male, mean age was 41, and hunting experience averaged 20 years (Ringleman, 1997). Our results generally support similar studies (Smith and Roberts, 1976; Enck et al., 1993; Humburg et al., 1997) regarding hunter characteristics, demographics, and residency. Our results also demonstrate that a gender disparity exists within natural resource agencies in South Dakota. Our ﬁndings, though similar to other studies, indicate that an aging waterfowl hunter population will likely reduce the potential for recruiting new hunters to the sport (Enck et al., 1993). Our ﬁndings also corroborate Smith and Roberts (1976) who documented that half the respondents stated that a family member or older friend was responsible for their participation in waterfowl hunting. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 101

Seventy-five percent of personnel surveyed during this study reported that access to waterfowl hunting areas was not a limiting factor to their participation. Conversely, Enck et al. (1993) demonstrated that "access issues" were important to waterfowl hunters in the state of New York. Public land was the most often hunted area (56.3%) for respondents in South Dakota and was selected more often than private land (14.8%). In Missouri, Humburg et al. (1997) determined that 44.2% of the days hunted for waterfowl were on public land while 38.3% of the days hunted were spent on private land. These same authors demonstrated that 53% of the waterfowl hunters surveyed indicated that "few places to hunt" was an important factor affecting their participation. Ac- cess to public lands is vitally important for waterfowl hunters; however, availability of these areas is highly variable geographically. Smith and Roberts (1976) asked 6,000 waterfowl hunters systematically sampled from Federal duck stamp purchases to rank 6 management activities in order of importance; providing more land for public hunting rated third with protecting nesting grounds and providing refuges for waterfowl ranking slightly higher. Respondents from our survey ranked loss of habitat, drought, predators, poor management, and hunting (in order of importance) as factors believed to have contributed to the historic declines in waterfowl populations (Table 3). Enck et al. (1993) documented that 74% of New York waterfowl hunters indicated that their interest in the sport had declined due to "management activities." Such activities included confusing regulations, season dates and lengths, bag limits, and a dislike for the mandatory steel shot regulations. Results from the NAWHS indicated respondent’s attitudes regarding regulations were the opposite (Ringleman, 1997). In fact, Ringleman (1997) demonstrated that in general, hunters were more accepting of "special" regulations. Gigliotti (1995) documented that resident and non-resident waterfowl hunters that participated in the 1994–95 South Dakota waterfowl season were generally satisfied with the current regulations. However, one suggestion made by this sample of hunters was a request for a longer season. Waterfowl hunters in Missouri generally did not cite regulations as important in their participation in waterfowl hunting (Humburg et al., 1997). The 4 most important reasons affecting participation cited by respondents of our survey were; other interests, time constraints, duck abundance, and cost (Table 5). These results support findings of both Enck et al. (1993) and Ringle- man (1997) who reported that respondents were "too busy" and that duck abundance had a greater effect on hunter participation rates than either bag size or season length, respectively. General comments provided by surveyed personnel during this study indicated that "costs" associated with the sport (i.e., equipment, stamps, shells, and gas) and poor quality table fare are other reasons that may determine whether or not respondents participated in waterfowl hunting in South Dakota. Achievement-oriented factors (i.e., a limit of ducks/geese) ranked low (7%) for respondents of our survey and were similar to findings of Enck et al. (1993) and Ringleman (1997). Similarly, Missouri waterfowl hunters ranked "seeing waterfowl", "hunting with family and friends", and "hunting with decoys" higher than "bagging ducks or geese" as factors that added to their enjoyment 102 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

(Humburg et al., 1997). Our results demonstrated that hunting with decoys, pass shooting, and jump shooting were the favored methods of natural resource professionals in South Dakota. Analogous to these results, Humburg et al. (1997) documented decoying birds over water, jump shooting, pass shooting, and decoying birds over land were preferred hunting methods in Missouri. Pass shooting ranks high as a hunting method in South Dakota, which may be due to this method being preferred by a large segment of the goose hunting population along the Missouri River. Natural resource agency professionals surveyed during this study indicated that they used other state wildlife personnel, state/federal waterfowl reports, other federal wildlife personnel, Ducks Unlimited, and other hunters as their 5 most important sources of information concerning waterfowl populations and the upcoming waterfowl season (Table 4). Newspapers, magazines, and television ranked relatively low (≤ 56%) as a source of information. In contrast, Smith and Roberts (1976) found that newspapers ranked highest followed by "talking to friends" and magazines. Results for South Dakota hunters who participated in NAWHS indicated that personal observation and contact with other hunters followed by state/federal waterfowl reports were considered important sources of information about duck abundance (Ringleman, 1997). Recently waterfowl management and more specifically waterfowl harvest management has become increasingly complex with the number of hunting zones, season splits, and species specific regulations nearly doubling over the last 20 years (Ringleman, 1997). Results of the NAWHS indicate that, in general, managers poorly predicted hunter attitudes regarding harvest regulations and influences of regulations on participation rates. Results of our questionnaire and diary will aid waterfowl managers and administrators by providing a better understanding of waterfowl species composition, crippling losses, hunter activities, and characteristics of a segment of South Dakota’s waterfowl hunters.

ACKNOWLEDGMENTS

This project was supported by Federal Aid to Wildlife Restoration P - R Pro- ject W-75-R, (Study No. 7569) through the South Dakota Department of Game, Fish and Parks, South Dakota State University, and the South Dakota Cooper- ative Fish and Wildlife Research Unit in cooperation with the U.S. Fish and Wildlife Service, U.S. Geological Survey, and the Wildlife Management Institute. We wish to thank K. Higgins, L. Gigliotti, K. McPhillips, D. Ode, and S. Vaa for providing suggestions and reviewing drafts of the survey. We also acknowledge J. Ringleman, Ducks Unlimited, Inc., Bismarck, ND and D. Humburg, Mis- souri Department of Conservation, Columbia, MO for providing additional literature. J. Ringleman, L. Gigliotti, S. Vaa and K. Higgins reviewed earlier drafts of this manuscript.

LITERATURE CITED

Anderson, D. R., and K. P. Burnham. 1976. Population of the mallard. VI. The effect of exploitation on survival. U.S. Fish Wildl. Serv. Resour. Publ. 128. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 103

Anderson, W. L., and R. A. Williamson. 1990. Illinois waterfowl harvest, hunter activity, and attitudes toward shooting hen mallards, legal shooting time, point system, and season dates in 1988. Ill. Dept. Conserv., Periodic Rep. No. 64. Anderson, I. 1994. A bird in the hand, a bird in the bush. New Scientist 4. Anonymous. 1986. North American waterfowl management plan. U.S. Dep. Inter. and Environ. Canada. Bellrose, F. C. 1953. A preliminary evaluation of cripple losses in waterfowl. Trans. N. Am. Wildl. Nat. Resour. Conf. 18:337-360. Caithamer, D. F., and J. A. Dubovsky. 1996. Waterfowl population status, 1996. U.S. Fish Wildl. Serv., Off. Migr. Bird Manage., Laurel, MD. Dietz, N. J. 1990. Surveys of citizens attitudes towards hunting, hunters and wildlife in South Dakota. M.S. Thesis, South Dakota State University, Brookings. Enck, J. W., B. L. Swift, and D. J. Decker. 1993. Reasons for decline in duck hunting: insights from New York. Wildl. Soc. Bull. 21:10-21. Filion, F. L. 1975. Estimating bias due to nonresponse in mail surveys. Pub- lic. Opin. Q. 39:482-492. Filion, F. L. 1976a. Exploring and correcting for nonresponse bias using follow-ups of nonrespondents. Pac. Sociol. Rev. 19:401-408. Filion, F. L. 1978. Increasing the effectiveness of mail surveys. Wildl. Soc. Bull. 6:135-141. Filion, F. L., and S. A. D. Parker. 1984. Human dimensions of migratory game- bird hunting in Canada. Can. Wildl. Serv., Occas. Paper No. 51. Geis, A. D., and W. F. Crissey. 1973. 1970 test of the point system for regu- lating duck harvests. Wildl. Soc. Bull. 1:1-21. Gigliotti, L. M. 1995. Waterfowl hunter satisfaction survey-1994. South Dako- ta Dept. Game, Fish and Parks, Admin. Rep., Pierre, SD. Gray, B. T., and R. M. Kaminski. 1993. Assessing a mail survey to estimate illegal waterfowl hunting. Wildl. Soc. Bull. 21:188-193. Gray, B. T., and R. M. Kaminski. 1994. Illegal waterfowl hunting in the Mis- sissippi Flyway and recommendations for alleviation. Wildl. Monogr. 127. Halladay, D. R. 1969. An attempt to evaluate waterfowl hunting crippling loss on the west side of Netley Marsh in Manitoba-1969. Can. Wildl. Serv. Un- publ. Rep. Higgins, K. F., N. J. Dietz, and R. D. Mendelsohn. 1994. Land operator problems associated with wildlife and hunters in South Dakota. Proc. S.D. Acad. Sci. 73:109-120. Hochbaum, H. A. 1947. The effect of concentrated hunting pressure on waterfowl breeding stock. Trans. N. Am. Wildl. Nat. Resour. Conf. 12:53-62. Hochbaum, G. S., and C. J. Walters. 1984. Components of hunting mortality in ducks. Can. Wildl. Serv., Occas. Paper No. 52. Humburg, D. D., G. S. Olson, and S. L. Sheriff. 1997. Attitudes and preferences of Missouri waterfowl hunters. Proc. Canada Goose Symp., Madison, WI. In Press. 104 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Martin, E. M., and S. M. Carney. 1977. Population ecology of the mallard. IV. A review of duck hunting regulations, activity, and success, with special reference to the mallard. U.S. Fish Wildl. Serv. Resour. Publ. 30. Nieman, D. J., G. S. Hochbaum, F. D. Caswell, and B. C. Turner. 1987. Mon- itoring hunter performance in prairie Canada. Trans. N. Am. Wildl. Nat. Resour. Conf. 52:233-245. Pierce, C. L., J. K. Ringleman, M. R. Szymczak, and M. J. Manfredo. 1996. An investigation of factors affecting waterfowl hunting participation in Col- orado. Colo. State Univ., Human Dimensions in Nat. Resour. Unit Rep. 10. Ringleman, J. K. 1997. Effects of regulations and duck abundance on duck hunter participation and satisfaction. Trans. N. Am. Wildl. Nat. Resour. Conf. 62:361-376. Roeher, G. A. 1963. Effective techniques in increasing response to mailed questionnaires. Public. Opin. Q. 27:299-302. Scherff, K. L., and J. K. Ringleman. 1994. Preferences, attitudes, and behavior of North American waterfowl hunters: a compendium of references. Colo. Div. Wildl., Unpubl. Rep. Sharp, D. E. 1996. Waterfowl harvest and population survey data book- Cen- tral Flyway. Occas. Rep., Off. Migr. Bird Manage., Denver, CO. Smith, R. I., and R. J. Roberts. 1976. The waterfowl hunters’ perceptions of the waterfowl resource. Trans. N. Am. Wildl. Nat. Resour. Conf. 51:188- 193. Tacha, T. C. 1976. Analysis of aerial surveys and tolerance of landowners for a Canada goose ﬂock in northeastern South Dakota. M.S. Thesis, South Dakota State University, Brookings. Tittle, C., and R. J. Hill. 1967. Attitude measurement and prediction of behavior: an evaluation of conditions and measurement techniques. So- ciometry 30:199-213. Tori, G. M., J. L. Weeks, and S. J. Miller. 1988. Ohio waterfowl hunter attitude survey results, 1984. Ohio Dept. Nat. Resour. Fed. Aid Wildl. Restor. Proj. W-126-R Final Rep. 50pp. Van Dyke, F. 1980. Hunter attitudes and exploitation on crippled waterfowl. Wildl. Soc. Bull. 8:150-152. Wilkinson, L. 1990. SYSTAT: The system for statistics. SYSTAT, Inc., Evanston, IL. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 105

THE IMPORTANCE OF CONSERVATION RESERVE PROGRAM FIELDS TO BREEDING GRASSLAND BIRDS AT BUFFALO RIDGE, MINNESOTA

Krecia L. Leddy Department of Wildlife and Fisheries Sciences, and Kenneth F. Higgins South Dakota Cooperative Fish and Wildlife Research Unit USGS-BRD, South Dakota State University Brookings, SD 57007

David E. Naugle College of Natural Resources University of Wisconsin—Stevens Point Stevens Point, WI 54481

ABSTRACT

Nongame birds were surveyed during summer 1995 at Buffalo Ridge in southwestern Minnesota, to evaluate the importance of Conservation Reserve Program (CRP) grasslands to local avifauna. Bird abundance and composition were compared among 3 habitat types (CRP grasslands, pasturelands, and croplands) using an index to breeding bird density (i.e., number of singing males/transect area), percent species composition, and total species richness. Vertical height and density of vegetation were measured early in the growing season (mid-May) and during the peak of the growing season (mid-June) to determine whether vegetative structure was related to bird use of vegetation. Conservation Reserve Program ﬁelds had higher vegetation measurements and supported higher bird densities and species richness than pasturelands and croplands. Mean bird density (birds/100 ha) in CRP grasslands was 312.5 compared to 166.7 in pasturelands and only 75.0 in croplands. Ten bird species were present in CRP grasslands compared to 8 in pasturelands and 9 in croplands. The presence of 3 native bird species (sedge wren, dickcissel, and clay- colored sparrow) in CRP grasslands that were not found in pasturelands or croplands indicated that CRP grasslands were an important habitat type for maintaining avian diversity at Buffalo Ridge.

INTRODUCTION

Agricultural production has destroyed most grasslands that historically provided habitat for upland nesting grassland birds. Excessive grazing and inva- sive woody species have degraded many of the remaining grasslands. In response to grassland losses, several prairie grassland bird species have declined in abundance (Johnson and Schwartz 1993). The Conservation Reserve Pro- gram (CRP), Title XII of the Federal Food Security Act of 1985 (Public Law 99- 106 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

198), encourages land operators to restore croplands to perennial grassland cover through 10-year contracts with the U.S. Department of Agriculture. Al- though the CRP was primarily designed to limit crop surpluses and conserve soil and water resources (Young and Osborn 1990), the CRP also provides upland habitat for grassland nesting birds (Johnson and Schwartz 1993; Kennedy,1994; Igl and Johnson 1995; Johnson and Igl 1995; King and Savidge 1995). A wealth of research was conducted in the early 1990's to evaluate the importance of CRP grasslands to upland nesting birds before 10-year contracts with land operators began to expire in 1996. Research indicated that CRP grasslands supported diverse bird communities and that vegetative structure of grasslands inﬂuenced bird species richness and abundance (Johnson and Schwartz 1993; King and Savidge 1995). Despite the usefulness of this information, previous studies have not compared bird use of CRP grasslands to other herbaceous habitat types available to breeding birds in the northern Great Plains to determine the relative importance of CRP grasslands. Objectives of this study were to (1) assess bird species richness, composition, and density in 3 major habitat types, (2) to characterize the vegetative structure within each habitat type, and (3) to relate vegetative characters to bird habitat use. We hypothesized that CRP grasslands support a community of native bird species unlike that of surrounding habitat types.

STUDY AREA

Buffalo Ridge in southwestern Minnesota is a 100 km segment of the Be- mis Moraine that begins 3 km northeast of Holland, Minn., and extends 10 km northwest of Lake Benton, Minn. Elevation is 546-610 m above mean sea level. Habitats on Buffalo Ridge consist primarily of corn (Zea mays), soybeans (Glycine max), small grains, pasture, hay, and CRP grasslands. The majority of the CRP ﬁelds on Buffalo Ridge were planted to a mixture of smooth brome (Bromus inermis)/alfalfa (Medicago sativa) or to switchgrass (Panicum virgatum). Scattered deciduous woodlands exist near farmsteads and in ravines. A variety of grassland bird species reside on Buffalo Ridge during the summer. Climate is temperate continental with cold winters and warm summers. Dur- ing 1995, precipitation was above average (87.9 cm) and temperature was normal (6.2 C) (MN Dep. Nat. Resour., pers. comm., 1996).

METHODS

Bird surveys were conducted 15 May to 1 July 1995 in 3 habitat types (pastureland, cropland, and CRP grasslands) using 40-m fixed-width transects (Wake- ley 1987). Fluorescent flagging was used to delineate transect boundaries. In- consistencies in surveys attributable to periodic bird inactivity (Skirvin 1981; Verner and Ritter 1986) were minimized by conducting surveys between sunrise and 1000 hrs. Surveys were not conducted during heavy rain or high winds (≥ 20 km/hr) (Mikol 1980; Ralph et al. 1993). We recorded all birds seen or heard while walking transects at 1.0-1.5 km/hr (Mikol 1980; Wakeley 1987). Flushed birds seen leaving transects were recorded (Burnham et al. 1980) while birds seen entering transects or only flying overhead were excluded from surveys. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 107

Each habitat type was represented by 3 fields ranging in size from 19 to 43 ha. One transect was established in each field. Transects that varied in length according to field size were placed ≥30 m from field borders and wetlands to avoid bias associated with edges (Arnold and Higgins 1986; Reese and Ratti 1988). An index to breeding bird density was calculated by dividing the number of perched and/or singing males by transect area. Percent species composition was calculated by dividing the number of a particular bird species by the total number of birds for that habitat type. Species richness was defined as the number of species in a given area (Koford et al. 1994). Vegetation measurements were collected within habitat types early in the growing season (mid-May) and during the peak of the growing season (mid- June). A modified Robel pole was used to measure vertical density (visual obstruction) of vegetation (Robel et al. 1970; Higgins and Barker 1982). Visual obstruction readings (VOR) were taken at the lowest point where vegetation restricted 100% visibility of the pole from a sighting height of 1 m and at a distance of 4 m (Robel et al. 1970). Forty VORs per field (1 for each cardinal direction/10 sampling stations) were recorded to the nearest 0.25 dm. Vegeta- tion height was estimated by measuring the tallest plant within a 30 cm radius of the pole to the nearest 0.25 dm (Higgins and Barker 1982). One height measurement was taken per sampling station, totaling 10 measurements per field. One-way Analysis of Variance was used to evaluate relationships between bird densities and the 3 habitat types.

RESULTS

Bird species richness and mean density were greater in CRP grasslands than in pasturelands or croplands (P<0.01) (Table 1). Ten species with a mean density of 312.5 individuals/100 ha were recorded in CRP grasslands (Tables 1,2). Three species comprised 80.0% of the bird species composition in CRP grasslands (Table 2). Eight species with a mean density of 166.7 individuals/100 ha were recorded in pasturelands. Two species comprised 76.3% of the bird species composition in pasturelands (Table 2). Nine species with a

Table 1. Mean density of male birds per 100 ha by habitat type during May- June 1995 at Buffalo Ridge, Minnesota.

Breeding Males

Habitat N Species N Mean SE Richness Densitya,b

Cropland 3 9 36 75.0 A 33.1 Pastureland 3 8 80 166.7 A 44.1 CRP grassland 3 10 150 312.5 B 15.7 a LSD(.05) = 110. b Means denoted by the same letter did not differ, (P ≤ 0.05). 108 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 2. Percent (%) composition of perched and/or singing male passerine birds surveyed in Conservation Reserve Program grasslands, pastureland, and cropland during May-June 1995 at Buffalo Ridge, Minnesota. Common and scientiﬁc names follow Banks et al. (1987).

Species Grassland Pastureland Cropland

Total Number 150 80 36 Total species 10 8 9 Common yellowthroat Geothlypis trichas 4.7 American robin Turdus migratorius 1.3 5.6 Sedge wren Cistothorus platensis 26.0 Savannah sparrow Passerculus sandwichensis 22.0 37.5 Red-winged blackbird Agelaius phoeniceus 4.7 1.3 13.8 Brown-headed cowbird Molothrus ater 4.0 Grasshopper sparrow Ammodramus savannarum 2.7 38.8 Dickcissel Spiza americana 1.3 Clay-colored sparrow Spizella pallida 0.7 Western meadowlark Sturnella neglecta 0.7 8.8 Killdeer Charadrius vociferus 1.3 16.7 Le Conte's sparrow Ammodramus leconteii 1.3 Vesper sparrow Pooecetes gramineus 16.7 Song sparrow Melospiza melodia 11.1 Western kingbird Tyrannus verticalis 11.1 Horned lark Eremophila alpestris 8.3 Eastern kingbird Tyrannus tyrannus 5.6 Unknown 1.3 7.5 5.6 mean density of 75.0 individuals/100 ha occurred in croplands. Six species constituted 77.7% of the cropland bird species (Table 2). Mean VOR and vegetation height measurements differed by habitat type (P<0.01) (Table 3). A signiﬁcant habitat by time interaction for VOR (F= 11.82; 3, 16 df; P<0.01) and vegetation height (F = 9.96; 3, 16 df; P<0.01) indicated that vegetation differed between mid-May and mid- June for each habitat type (Table 3). Visual obstruction readings in CRP grasslands and croplands were 3-6 times higher in June compared to May (Table 4). Visual obstruction readings in pasturelands were only 2 times higher in June compared to May (Table 4).

DISCUSSION

Bird surveys in the 3 habitat types used by breeding birds indicated that CRP grasslands supported a community of native bird species unlike that of surrounding habitat types. Differences in vegetative structure of herbaceous cover in pasturelands, croplands, and CRP ﬁelds played a key role in determining bird species composition. Vegetation in CRP grasslands typically was tall and dense compared to that in pasturelands and croplands. Conservation Reserve Program ﬁelds had higher vegetation measurements and supported a higher density and diversity (i.e., species richness) of birds than pasturelands and croplands. More importantly, CRP grasslands provided habitat for bird Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 109

Table 3. Mean visual obstruction readings (VOR) and Vegetation height (VH) measurements (dm) during May-June 1995 at the Buffalo Ridge, Minnesota.

Mean VORa,c Mean VHb,c

Habitat N N (dm) SE N (dm) SE

Cropland 3 240 0.2 A 0.02 60 0.6 A 0.09

Pastureland 3 240 0.6 A 0.03 60 1.7 A 0.21

CRP grassland 3 240 3.7 B 0.19 60 5.0 B 0.42 a Visual obstruction reading LSD(0.05) = 1.0253 b Vegetation height LSD(0.05) = 1.165 c Means denoted by the same letter did not differ, (P ≤ 0.05).

Table 4. Mean vegetation height (dm) in habitat types (cropland, pastureland, Conservation Reserve Program grasslands) in mid-May and mid-June at Buffa- lo Ridge, Minnesota, 1995.

Mean Vegetation Height(dm)

Habitat N mid-Maya SD mid-Junea SD

Cropland 30 0.15 0.17 0.96 0.80

Pastureland 30 1.15 1.00 2.26 1.87

CRP 30 2.16 0.56 7.88 1.99 a Vegetation height differed (LSD (0.05) = 0.8238) between mid-May and mid-June. species that required idled grasslands with tall dense herbaceous growth. Sedge wrens, dickcissels, and clay-colored sparrows were recorded only in CRP grasslands and bobolinks densities were 12 times higher in CRP ﬁelds than in pasturelands. Sedge wrens and clay-colored sparrows used dense stands of switchgrass while dickcissels and bobolinks usually used stands of smooth brome and alfalfa. Vegetative height and VOR measurements indicated that season-long grazing reduced the height and density of vegetation in pasturelands throughout the growing season. Croplands on Buffalo Ridge also were characterized by low levels of ground cover and short vegetation despite abundant weed growth that occurred before crops were planted. Grasshopper and savannah sparrows, the 2 bird species most abundant in pasturelands in our study, are generalist species that typically inhabit sparsely vegetated areas (Wiens 1969). 110 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

We recorded additional generalist species such as killdeer and vesper sparrows that typically nest in sparse vegetation with low ground cover (Johnsgard 1980; Best and Rodenhouse 1984). Although pasturelands and croplands supported numerous generalist bird species, the absence of sedge wrens, dickcissels, and clay-colored sparrows from habitats other than CRP ﬁelds indicated that CRP grasslands were an important habitat type for maintaining avian diversity at Buffalo Ridge.

ACKNOWLEDGMENTS

We thank G. Arnold, L. Flake, and C. Gritzner for early reviews of our manuscript. P. Evenson provided statistical support. V. Swier and R. Osborn assisted with ﬁeld work. We also thank landowners that granted access to private property. The project was funded by Kenetech Windpower, Inc., and South Dakota Department of Game, Fish and Parks. Additional support was provided by the Natural Resources Conservation Service and the South Dako- ta Cooperative Fish and Wildlife Research Unit, in cooperation with the Na- tional Biological Survey, Wildlife Management Institute, U.S. Fish and Wildlife Service, U.S. Geological Survey/BRD, and South Dakota State University.

LITERATURE CITED

Arnold, T. W., and K. F. Higgins. 1986. Effects of shrub coverages on birds of North Dakota mixed-grass prairie. Can. Field-Nat. 100:10-14. Banks, R. C., R. W. McDiarmid, and A. L. Gardner. 1987. Checklist of vertebrates of the United States, the U.S. Territories, and Canada. U.S. Fish and Wildl. Serv. Resour. Pub. No. 166, Washington, D.C. 79pp. Best, L. B., and N. L. Rodenhouse. 1984. Territory preference of vesper sparrows in cropland. Wilson Bull. 96:72-82. Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. No. 72. Higgins, K. F., and W. T. Barker. 1982. Changes in vegetation structure in seeded nesting cover in the Prairie Pothole Region. U.S. Fish and Wildl. Serv. Spec. Sci. Rep. Wildl. No. 242. 26pp. Igl, L. D., and D. H. Johnson. 1995. Dramatic increase of Le Conte's sparrow in Conservation Reserve Program ﬁelds in the Northern Great Plains. Prairie Nat. 27:89-94. Johnsgard, P. A. 1980. A revised list of the birds of Nebraska and adjacent plains states. Occas. Pap. Nebr. Ornithol. Union No. 6, Lincoln. 170pp. Johnson, D. H., and L. D. Igl. 1995. Contributions of the Conservation Reserve Program to populations of breeding birds in North Dakota. Wilson Bull. 107:709-718. Johnson, D. H., and M. D. Schwartz. 1993. The Conservation Reserve Pro- gram: habitat for grassland birds. Great Plains Res. 3:273-295. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 111

Kennedy, C. L. 1994. Effects of grazing on nongame breeding birds, insects, and vegetation in Conservation Reserve Program grasslands in North Dako- ta. M.S. thesis, South Dakota State Univ., Brookings. 65pp. King, J. W., and J. A. Savidge. 1995. Effects of the Conservation Reserve Pro- gram on wildlife in southeast Nebraska. Wildl. Soc. Bull. 23:377-385. Koford, R. R., J.B. Dunning, Jr., C. A. Ribic, and D. M. Finch. 1994. A glossary for avian conservation biology. Wilson Bull. 106:121-137. Mikol, S. A. 1980. Field guidelines for using transects to sample nongame bird populations. U.S. Fish and Wildl. Serv. Bio. Serv. Prog. OBS-80-58. 26pp. Ralph, C. J., G. R. Geupel, P. Pyle, T. E. Martin, and D. F. DeSante. 1993. Hand- book of ﬁeld methods for monitoring landbirds. Gen. Tech. Rep. PSW- GTR-144. Albany, Calif: Pac. Southwest Res. Stn., For. Serv. U.S. Dep. Agric. 41pp. Reese, K. P., and J. T. Ratti. 1988. Edge effect: a concept under scrutiny. Trans. N. Am. Wildl. Nat. Resour. Conf. 53:127-136. Robel, R. J., J. N. Briggs, A. D. Dayton, and L. C. Hulbert. 1970. Relationships between visual obstruction measurements and weight of grassland vegetation. J. Range Manage. 23:295-297. Skirvin, A. A. 1981. Effect of time of day and time of season on number of observations and density estimates of breeding birds. Pages 271-274 in C. Ralph and J. Scott, eds. Estimating numbers of terrestrial birds. Stud. Avian Biol. 6. Verner, J., and L. V. Ritter. 1986. Hourly variation in morning point counts of birds. Auk 103:117-124. Wakeley, J. S. 1987. Avian line-transect methods. Section 6.3.2, U.S. Army Corps Eng. Wildl. Resour. Manage. Manual. Tech. Rep. EL-87-5. 21pp. Wiens, J. A. 1969. An approach to the study of ecological relationships among grassland birds. Ornithol. Monogr. No. 8. 93pp. Young, E. C., and C. T. Osborn. 1990. Costs and beneﬁts of the Conservation Reserve Program. J. Soil Water Conserv. 45:370-373.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 113

EFFECTS OF WIND TURBINES ON NESTING RAPTORS AT BUFFALO RIDGE IN SOUTHWESTERN MINNESOTA

Robert E. Usgaard, David E. Naugle and Robert G. Osborn Department of Wildlife and Fisheries Sciences, and Kenneth F. Higgins South Dakota Cooperative Fish and Wildlife Research Unit, USGS-BRD, South Dakota State University Brookings, SD 57007

ABSTRACT

Raptor surveys were conducted at the Buffalo Ridge Wind Resource Area (BRWRA) in southwestern Minnesota during the summers of 1994-95 to determine whether the nesting habits of locally breeding raptors were inﬂuenced by the presence of wind turbines. Red-tailed and Swainson's hawks, American kestrels, and northern harriers were the primary species recorded during surveys. Raptor abundance (raptors/10 km) was consistent between years of our study (2.08 in 1994, 2.07 in 1995) and was comparable to abundances reported for similar habitats in the region. Average raptor nest density (nests/100 km2) for the 2 years on lands surrounding the windplant was 5.94. All raptor nests (n=31) within the 293 km2 BRWRA were found on lands where wind turbines were absent (261 km2) despite the presence of abundant treebelt and riparian habitat within the 32 km2 windplant facility. We recommend that areas providing extensive raptor nesting habitat be avoided in future selections of windplant construction sites.

INTRODUCTION

Recent technological advances in equipment designed to harness windpower has made the cost of electricity from windpower generating plants competitive with that of electricity derived from fossil fuels. In response to decreased production costs, the windpower industry is expanding from Califor- nia into the eastern United States and Canada, as well as Latin America (Nel- son and Curry 1995). Although renewable energy resources such as windpower have received strong public support, impacts of wind turbines on avian communities have not been adequately researched. Avian mortality from collisions with wind turbines varies greatly from little or no mortality (Byrne 1983, Winkelman 1985 and 1990, Higgins et al. 1995) to substantial mortality (Mc- Crary et al. 1983, Orloff and Flannery 1992). In addition to direct mortality from collisions, research also has indicated that bird densities surrounding turbines were lower compared to densities in similar habitats outside the vicinity of turbines (Leddy et al. 1999). The objective of this study was to determine 114 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) whether the nesting habits of breeding raptor populations in southwestern Minnesota were inﬂuenced by the presence of wind turbines. We hypothesized that raptor nest densities within a newly constructed windplant site would not differ from those in the surrounding area.

STUDY AREA AND METHODS

The Buffalo Ridge Wind Resource Area (BRWRA) in southwestern Min- nesota is located along a 100 km segment of the Bemis Moraine that begins 3 km northeast of Holland, Minn., and extends 10 km northwest of Lake Ben- ton, Minn. Elevation is 546-610 m. The 293 km2 BRWRA is comprised of a 32 km2 windplant and 261 km2 of lands surrounding the windplant. The windplant within the BRWRA contains 73 wind turbines. The additional lands (261 km2) within the BRWRA have been leased as future wind turbine development sites. Raptor breeding habitat within the BRWRA consists of deciduous tree- belts and riparian woodlands in ravines comprising 3% of the landscape. Habi- tats other than woodlands include agricultural crops (i.e., corn (Zea mays), soybeans (Glycine max), and small grains), hay and pasturelands, and Con- servation Reserve Program grasslands. Raptor surveys were conducted biweekly (10 June to 14 September) or weekly (15 September to 26 October) (n=13) in 1994 and weekly (22 March to 26 October) (n=30) in 1995 along a survey route (Higgins et al. 1995) within the BRWRA to determine species composition and relative abundance of raptors. We recorded the number and species of raptors seen while driving at 40 kph along a 69 km survey route that bisected the BRWRA. Surveys began 2-3 hrs after sunrise. The north-south starting points for surveys were alternated to reduce biases associated with diurnal raptor activity patterns. We calculated relative abundance of raptors/10 km of survey route by dividing the number of raptors seen during all surveys within each year by 6.9. Raptor nest searches also were conducted during April-May 1994 and 1995 to determine the location and species composition of breeding raptors on the BRWRA. We scanned all shelterbelts and woodlands for raptor nesting activity from all available roads within the BRWRA using spotting scopes and binoculars. Potential raptor nests were monitored weekly until nests became occupied or tree foliage obstructed our vision of the site. Location and species were recorded for all active raptor nests.

RESULTS

A total of 187 raptors representing 6 species were recorded during 13 road surveys in 1994 (Table 1). In 1995, 428 raptors representing 7 species were recorded during 30 road surveys (Table 1). Red-tailed hawks, northern harriers, and American kestrels constituted 86% of all raptors seen in 1994 and 77.1% of all raptors seen in 1995. The relative abundance of raptors per 10 km along the survey route was 2.08 in 1994 and 2.07 in 1995 (Table 1). Fifteen occupied raptor nests representing 4 species were found on the BRWRA in 1994 (Table 2). In 1995, 16 occupied raptor nests of 3 species were found Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 115

Table 1. Total number and relative abundance (sightings/10 km) of raptors recorded during summer 1994 (n=13 surveys) and 1995 (n=30 surveys) along a 69 km road survey route at the Buffalo Ridge Wind Resource Area near Lake Benton, Minnesota.

Raptor Abundance Total Numbers (sightings/10 km)

Species 1994 1995 1994 1995

All Species Combined 187 428 2.08 2.07

Red-tailed hawk (Buteo jamaicensis) 81 164 0.90 0.79

Northern Harrier (Circus cyaneus) 49 73 0.55 0.35

American Kestrel (Falco sparverius) 30 93 0.33 0.45

Swainson's hawk (Buteo swainsoni) 19 37 0.21 0.18

Great horned owl (Bubo virginianus) 3 4 0.03 0.02

Ferruginous hawk (Buteo regalis) 2 0 0.02 ----

Rough-legged hawk (Buteo lagopus) 0 5 ----- 0.02

Cooper's hawk (Accipiter cooperii) 0 1 ----- 0.01

Unknown hawks 2 51 0.02 0.25

(Table 2). Overall raptor nest density (nests/100 km2) was 5.94 on lands surrounding the windplant facility (261 km2) where wind turbines were absent (Table 2). No raptor nests were found on the 32 km2 windplant facility where turbines were present.

DISCUSSION

Species composition and relative abundance of raptors were consistent between years during our surveys on the BRWRA and were comparable to those reported for similar habitats in eastern South Dakota (Norelius 1984). Nests of red-tailed and Swainson's hawks, the most common above-ground nesting raptors on the BRWRA, were usually found in treebelt habitat that provided secure nesting cover. During our surveys on the BRWRA, we found no evidence of raptor nesting activity within the 32 km2 windplant facility despite the presence of treebelt and riparian habitat that was comparable to the lands (261 km2) surrounding the windplant where raptor nests were found. Research at the BR- 116 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 2. Number and density (nests/100 km2) of occupied raptor nests found within the 261 km2 area surrounding the windplant facility on the Buffalo Ridge Wind Resource Area near Lake Benton, Minnesota, 1994-95.

Raptor Nests

1994 1995

Species # Density # Density

Swainson's hawk (Buteo swainsoni) 9 3.46 6 2.30

Red-tailed hawk (Buteo jamaicensis) 4 1.53 8 3.06

Ferruginous hawk (Buteo regalis) 1 0.38 0 -----

Great horned owl (Strix nebulosa) 1 0.38 2 0.77

Totals 15 16

Overall Nest Density/year 5.74 6.13

Nest Density/years combined 5.94

WRA has indicated that no raptor mortalities have occurred due to collisions with wind turbines (Higgins et al. 1995, Nelson and Curry 1995, Osborn et al. 2000). Although wind turbines have not caused direct mortality, the presence of wind turbines may be indirectly affecting local raptor populations by decreasing the use of suitable nesting sites. Recent studies in the Netherlands have indicated that the mere presence of wind turbines has prevented waterfowl and wading bird species from using otherwise suitable habitat near wind turbines (Winkelman 1990, Pedersen and Poulsen 1991). As the demand for wind generated power increases in the future, so will the demand for information concerning the effects of wind turbines on nesting and migrating birds. Potential windplant construction sites that provide extensive habitat for nesting raptors should be avoided until additional information is available.

ACKNOWLEDGMENTS

We thank C. D. Dieter for assisting with ﬁeld work. K. K. Bakker reviewed an earlier draft of the manuscript. This project was funded by Kenetech Wind- power, Inc., and the South Dakota Department of Game, Fish and Parks. Ad- ditional support was provided by the South Dakota Cooperative Fish and Wildlife Research Unit, in cooperation with the National Biological Survey, Wildlife Management Institute, U.S. Fish and Wildlife Service, U.S. Geological Survey/BRD, and South Dakota State University. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 117

LITERATURE CITED

Byrne, S. 1983. Bird movements and collision mortality at a large horizontal axis wind turbine. Cal-Neva Wildl. Trans. 1983:76-83. Higgins, K. F., R. E. Usgaard, and C. D. Dieter. 1995. Monitoring seasonal bird activity and mortality at the Buffalo Ridge Windplant, MN. South Dakota State Univ. 64pp. Leddy, K.L., K.F. Higgins, and D.E. Naugle. 1999. Effects of wind turbines on upland nesting birds in Conservation Reserve Program grasslands. Wilson Bull. 111(1): 100-104. McCrary, M. D., R. L. McKernan, R. E. Landry, W. D. Wagner, and R. W. Schreiber. 1983. Nocturnal avian migration assessment of the San Gor- gonio wind resource study area, spring 1982. Unpubl. summary for South- ern California Edison Co., Rosemead. Nelson, H. K., and R. C. Curry. 1995. Assessing avian interactions with windplant development and operation. Trans. North Am. Wildl. Natur. Re- sour. Conf. 60:266-277. Norelius, S. E. 1984. Use of eastern South Dakota shelterbelts by nesting birds of prey. M. S. Thesis, South Dakota State Univ., Brookings. 48pp. Orloff, S. G., and A. W. Flannery. 1992. Wind turbine effects on avian activity, habitat use, and mortality in the Altamont Pass and Solano County Wind Resource Areas, 1989-1991. Final Rep. to Alameda, Contra Costa and Solano Counties and the California Energy Comm. by Biosystems Analysis, Inc., Tiburon, Calif. 14pp. Osborn, R.G., K.F. Higgins, R.E. Usgaard, C.D. Dieter, and R.D. Neiger. 2000. Bird mortality associated with wind turbines at the Buffalo Ridge Wind Re- source Area, Minnesota. Am. Midl. Nat. 143(1):41-52. Pedersen, A. M. B., and O. E. Poulsen. 1991. Avian responses to the implementation of the Tjaereborg Wind Turbine at the Danish Wadden Sea. Danske Vildtundersogelser 47:5-43 (with English summary). Winkelman, J. E. 1985. Bird impact by middle-sized wind turbines-on ﬂight behaviour, victims, and disturbance. Limosa 58:117-121 (with English summary). Winkelman, J. E.. 1990. Impact of the wind park near Urk, Netherlands, on birds: bird collision victims and disturbance of wintering waterfowl. Pages 402-403 in the supplement to the Twentieth Int. Ornithol. Congr. Christchurch, New Zealand.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 119

THE ORIGIN OF WATERFALLS IN THE BLACK HILLS, SOUTH DAKOTA

Charles Michael Ray and Perry H. Rahn South Dakota School of Mines and Technology

ABSTRACT

We studied the largest waterfalls in the Black Hills to better understand the geologic and hydrologic reasons for their existence. The waterfalls studied include: an unnamed falls on the Cheyenne River above Angostura Reservoir, “Big Falls” on Battle Creek below Keystone, “Bridal Veil Falls” on Rubicon Gulch along Spearfish Canyon, “Spearfish Falls” on Little Spearfish Creek near Savoy, “Roughlock Falls” on Little Spearfish Creek above Savoy, the waterfalls of Fall River below Hot Springs, and “Cascade Falls” on Cascade Creek. The geology of each waterfall was determined, their heights were measured, and cross sections were constructed. The discharge over each waterfall was obtained from USGS records; where no records were available the discharge was estimated. The waterfalls form in primarily two different ways. First, where more resistant rock units are exposed, this resistant ledge of rocks forms rapids and/or waterfalls. Thus a “knickpoint” is formed in the longitudinal stream profile. Second, some waterfalls form where calcareous-tufa is deposited in the stream bed. This usually happens below springs draining the Madison Limestone. The calc-tufa deposits typically accumulate within a mile below the spring, in a reach that already contains rapids. The rapids further encourage the outgassing of carbon dioxide, leading to the deposition of calcite and/or aragonite. Data compiled in this research may be used to help understand the geologic processes that created these waterfalls and help preserve them as aes- thetic resources for future generations.

INTRODUCTION

The Black Hills contains several significant waterfalls. They are located in various geologic environments (Fig. 1). The seven waterfalls studied in this investigation are believed to be the largest. Listed alphabetically they are: (1) Big Falls on Battle Creek , (2) Bridal Veil Falls on Rubicon Creek, (3) Cascade Falls on Cascade Creek, (4) Cheyenne River Falls on the Cheyenne River, (5) Fall River Falls on the Fall River, (6) Roughlock Falls on Little Spearfish Creek, and (7) Spearfish Falls on Little Spearfish Creek. The Black Hills is a Laramide uplift that consists of Paleozoic and Meso- zoic rocks unconformably overlying a core of Precambrian metamorphic and igneous rocks. Waterfalls and rapids are fairly common in the Black Hills, particularly in the Precambrian rocks. Many are on small ephemeral streams. For the pur- 120 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 1. Generalized geologic map of the Black Hills and its relationship to stream ﬂow, showing the location of waterfalls (map taken from Rahn et al, 1981). Numbers 1 to 7 refer to waterfalls shown in Table 1. pose of this research, only the seven largest known waterfalls were studied. Each waterfall was located on a 7.5 minute topographic map, and the waterfalls measured for height. Geologic cross sections were constructed. Dis- charge data over each waterfall was obtained from USGS records where available (USGS, 1995); where no records were available the discharge was estimated. The origin of each waterfall was then interpreted.

ORIGIN OF WATERFALLS

Big Falls

Big Falls is located on Battle Creek approximately 4 miles below Keystone (Sec. 11, T 2 S, R 5 E, in the Rockerville 7.5 minute quadrangle). The average discharge over the falls is about 8 cfs (Table 1), based on data taken at Keystone. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 121

Table 1. Waterfall data

Name Height (ft) Average discharge (cfs)

1. Big Falls 10 8 2. Bridal Veil Falls 80 0.6 3. Cascade Falls 6 20 4. Cheyenne River falls 3 112 5. Fall River falls 70 total 22 6. Roughlock Falls (upper) 14 13 7. Spearﬁsh Falls 60 13

Big Falls is about 10 ft high (Fig. 2), and is caused by a dike of pegmatite in the Precambrian schist and metagraywake (Rahn, 1987). As Battle Creek eroded into Precambrian rocks it encountered a pegmatite dike which is more resistant to erosion than the surrounding metagraywacke. The pegmatite forms a ledge over which the water falls (Fig. 3). There are other waterfalls in the Black Hills that are similar to Big Falls. The most noteworthy are the series waterfalls on upper Spring Creek below Sylvan Lake, “Little Falls” on Battle Creek about 1 mile below Big Falls, and the falls on Grizzly Creek about 2 miles above Keystone. All of these waterfalls are due to Precambrian granite outcrops in the stream bed. Rapid Creek in “Dark Canyon” about 5 miles west of Rapid City has drop of approximately 8 ft over a ledge of Precambrian quartzite.

Bridal Veil Falls

Bridal Veil Falls is located where Rubicon Gulch meets Spearﬁsh Canyon (Sec. 9, T 5 N, R 2 E, in the Maurice quadrangle). The height is 80 feet.

Figure 2. Geologic cross-section of Big Falls on Battle Creek. PCpeg is the pegmatite, PCgw is the Metagraywacke. 122 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 3. Photograph of Big Falls.

Discharge data for Rubicon Creek is not available, although some information is published regarding the headwaters at the Richmond Hill gold mine reclamation area. The average discharge of Rubicon Gulch is estimated at Bridal Veil Falls to be 0.60 cfs, based on the 1.36 square mile drainage area, using an value of 2.5 inches runoff from the northern Black Hills (Rahn and Davis, 1993). Bridal Veil Falls is formed on a portion of a Tertiary nephelene syenite lac- colith within the Deadwood formation (Lisenbee, 1995). As the Black Hills uplift and erosion carved out Spearﬁsh canyon the stream encountered the igneous rocks. The falls is formed over the resistant outcrop (or “knickpoint”) where Rubicon Gulch meets Spearﬁsh Canyon (Figs. 4 and 5). [Squaw Creek, about 2 miles to the south, has a series of small waterfalls and potholes in the Deadwood Formation. One big pothole is known as the “bathtub”.]

Cascade Falls

Cascade Falls is located on Cascade Creek below Cascade Springs (Sec. 30, T 8 S, R 5 E in the Cascade Springs quadrangle). Based on USGS data, the discharge over Cascade Falls is about 20 cfs. The height is 6 ft. Cascade Creek originates at Cascade Spring, about 3 miles above the falls. The spring is an outlet for ground water in the Madison Limestone and Min- nelusa Formation that originates from recharge over a vast area of the southern Black Hills (Rahn and Gries, 1973). As Cascade Creek cut into the sedi- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 123

Figure 4. Geologic cross-sectional sketch of Bridal Veil Falls.

mentary rocks on its way to the Cheyenne River it encountered a slightly more resistant layer of sandstone, most likely the Newcastle sandstone (Figs. 6 and 7). Rapids formed over the more resistant sandstone. The calcium carbonate rich water deposited calc-tufa at the falls, adding height to the rapids.

Cheyenne River Falls

Cheyenne River Falls is located in the Black Hills Wild Horse Sanctuary on the Cheyenne River above Angostura Reservoir (Sec. 35, T 8 S, R 4 E, in the Cas- cade Springs quadrangle). The discharge of the Fall River at Edgemont, about 30 miles just above the falls, is approximately 112 cfs (Table 1). Figure 5. Photograph of Bridal Veil falls, tak- The falls is about 3 ft en in Febuary, 1997. A large ice buildup can high, and is located where be seen. 124 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 6. Geologic cross section of Cascade Falls. Qct is the calc-tufa, Kns is

Newcastle Sandstone, and Ks is Skull Creek shale.

Figure 7. Photograph of Cascade Falls taken in 1970.

a resistant limestone bed crops out in the streambed. As the Cheyenne River cut through the sedimentary rocks ﬂanking the southern Black Hills it encountered the southeasterly dipping Lakota Formation, which includes the Fu- son shale member (Kfs), the Minnewasta Limestone (Kmw), a more resistant member, and deeper units of shale and sandstone shown on Figure 8 as the

Lakota formation (Kl). The more resistant Minnewasta limestone forms the falls as a 3 foot high ledge across the Cheyenne River (Fig. 9). Units of shale above and below the sandstone each erode faster than the limestone, leaving it exposed as a waterfall, or “knickpoint” in the stream proﬁle. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 125

Figure 8. Geologic cross section of Cheyenne River falls. Kfs is the Fuson Shale,

Kmw is the Minnewasta Limestone, and Kl is the Lakota formation.

Figure 9. Photograph of Cheyenne River falls.

Fall River Falls

Fall River falls is located on the Fall River about 5 miles east of Hot Springs (Sec. 33, T 7 S, R 4 E in the Hot Springs quadrangle). Based on the USGS gage at Hot Springs, the discharge over Fall River falls is about 22 cfs. The waterfalls are actually a serves of falls and rapids, with a total drop of 70 ft within a reach of about 300 ft. As the Fall River exits the Black Hills, it encounters the easterly-dipping Fall River Sandstone. The river eroded the sandstone until the gradient approximately matched the dip of the sandstone, and the water began to cascade and fall over the sandstone, forming several channels that were carved into the bedrock (Figs. 10 and 11). [Three boys 126 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 10. Geologic cross section of the Fall River falls. Qct is the calc-tufa, Kfr is the Fall River Formation. drowned there in 1995, tes- timent to the powerful current in some of the potholes.] Because of the high calcite content of the water from the spring us- tream at Hot Springs, calc- tufa was deposited on the bedrock falls, and built up in different locations where the water was already falling. Thus the falls is primarily caused by a resistant bedrock units, but is enhanced by the depostion of calc-tufa.

Roughlock Falls

Roughlock Falls is located on Little Spearﬁsh Creek about 4 miles upstream from Savoy (Sec.36, T 5 N, R 1 E) in the Savoy quadrangle. The water Figure 11. Photograph of Fall River falls. originates from a large spring discharging from the Madison Limestone about 3 miles above the falls. The discharge of Little Spearﬁsh Creek measured at a USGS gaging station below the falls, and is approximately 13 cfs. Roughlock Falls is formed as a calc-tufa deposit over the Whitewood For- mation, a white, buff, dolomitic limestone (Figs. 12 and 13). There are actu- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 127

Figure 12. Geologic cross section of Roughlock Falls. Qct is the calc-tufa, Ow is the Whitewood Dolomite. ally two waterfalls at this location, the upper one being about 14 ft high. Pre- sumably a ledge in the dolomite initially began the waterfall, encouraging calcite to precipitate, forming a calc-tufa, further building up the waterfalls.

Spearﬁsh Falls

Spearfish Falls is located on Little Spearfish Creek at the intersection of Little Spearfish Creek and Spearfish Creek (Sec. 31, T 5 N, R 2 E, in the Savoy quadrangle). It is about 1 mile below Roughlock Falls. The discharge over Spearfish Falls is regulated by Homestake Mining Company who for over 100 Figure 13. Photograph of Roughlock Falls. years have diverted prac- tially all of the water for their mining operations. In 1996 water flowed over the falls for the first time in almost 100 years (Figs. 14 and 15). The discharge above the falls, before the water is removed by Homestake, is about 13 cfs. Spearfish Falls is formed by a calcareous tufa deposited over the Dead- wood formation. Calcareous-tufa is porous calcite and/or aragonite that typically forms downstream from a spring discharging from a carbonate aquifer 128 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

(Back et al., 1983). Little Spearﬁsh Creek carries an abundance of dissolved calcite because it was discharged from a carbonate aquifer. As the water encounters the falls the calcite is deposited, over the bedrock ledge, and on top of the alluvium and plants at the rapids.

Conclusion

There are many waterfalls in the Black Hills. We restricted our study to the Figure 14. Geologic cross section of Spearfish larger ones. From Table 1 Falls. Qct is the calc-tufa, Cd is the Deadwood the greatest drop (70 ft) is Formation. at the Fall River falls. This drop takes place over a reach of 300 ft, and conse- quentally is not as impres- sive as others. The largest single drop (60 ft) is Spearfish Falls on Little Spearfish Creek. In terms of discharge, the Cheyenne River Falls (112 cfs) ranks as the largest. The greatest waterfall in terms of energy release (height times discharge) is the Falls River falls. Aesthetically, they are all beautiful in their own way and deserve to be protected for the enjoyment of everyone. There are two different types of waterfalls: (1) where a more resistant rock unit is exposed, forming a ledge or knickpoint for water to flow over, and (2) where calc-tufa is deposited below a limestone spring. The waterfalls form Figure 15. Photograph of Spearfish Falls. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 129 within a few miles of the spring on some reach already possessing rapids or waterfalls, thereby further enhancing the waterfalls. The geomorphology literature is full of references to the formation of knickpoints supposedly due to the headward erosion from multiple baselevels (Ritter et al., 1995) or “peneplains”. This concept has been largely discredit- ed (for example see Rahn, 1971, and references contained therein). The present reseach indicates that the knickponts are due to geological controls and have no relevance to knickpoint recession from baselevel episodes. While this research is primarily descriptive geomorphology, it also illustrates some fundamental scientific facts, and opens up several questions that could be examined in future research. This research illustrates that a knickpoint or abrupt steepening of the longitudinal profile of stream is formed by a resistant rock unit, and that calcareous-tufa deposits form waterfalls within a few miles downstream from a spring discharged from a carbonate aquifer. Fu- ture research could examine the true thickness of calc-tufa deposits on waterfalls and determine the time it takes them for them to form.

REFERENCES CITED

Back, W., B. Hanshaw, L., Plummer, and P. H. Rahn. 1983. Process and rate of dedolomitization mass transfer and 14C dating in a regional carbonate aquifer. Geological Society of America Bulletin. 94: 1415-1429. Gregory, K. J. and D. E. Walling. 1973. Drainage basin form and process. Ed- ward Arnold Publishers, Ltd. Leopold L. B. 1964. Fluvial processes in geomorphology. San Francisco, W.H. Freeman. Lisenbee, A. L., J. G. Kirchner., and C. J. Paterson. 1996. Geologic setting, Black Hills of South Dakota. Guidebook to the Geology of the Black Hills. South Dakota School of Mines and Technology Bull No. 19: 126 Plumb. 1993. A scale for comparing the visual magnitude of waterfalls. Earth Science Reviews. 34: 261-270 Rahn, P. H. 1971. The weathering of tombstones and its relationship to the topography of New England. Jour. Geol Ed. 19 (3):112-118. Rahn, P. H., and J. P. Gries. 1973. Large springs in the Black Hills, South Dako- ta and Wyoming. South Dakota Geological Survey, Rept. of Invest. 107. Rahn, P. H. 1974. Ground water geochemistry of the Pahasapa Limestone: Non Published. Completion Report for the Department of Interior project number A-046-SDAK Rahn, P. H., V. L. Bump, and F. V. Steece. 1981. Engineering geology of the Central Black Hills, South Dakota: in Rich , F. J. 1985. Geology of the Black Hills, South Dakota and Wyoming, 2nd Edition: AGI Field Trip Guidebook, Geological Society of America., Rocky Mountain Section Meeting, Rapid City, South Dakota, p. 152 Rahn, P. H. 1987. Geologic map and measured stratigraphic section for the Rockerville quadrangle, Pennington County, South Dakota. Geol. Soc. Am. Map and Chart Series MCH062 130 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Rahn, P. H., and A. D. Davis. 1993. Stream runoff from Black Hills watersheds. Proc. South Dakota Acad. Science 72: 161-175. Rahn, P. H., and A. D. Davis. 1996. Engineering geology of the Central and Northern Black Hills: Road log, Field Trip #3. in Paterson, C. J., and J. G. Kirchner, eds. Guidebook to the geology of the Black Hills, South Dakota South Dakota School of Mines and Technology Bull. 19, p. 19-29. Redden, J. A., and A. L. Lisenbee. 1996. Geologic setting, Black Hills, South Dakota. Guidebook to the Geology of the Black Hills, South Dakota. South Dakota School of Mines and Technology. Bull No. 19, p. 1-8 Ritter, D. F., R. C. Kochel, and J. R. Miller. 1995. Process geomorphology. Wm. C. Brown, Publ., Dubuque, IA. U.S.G.S., 1995, Water Resources Data for South Dakota: U.S. Department of the Interior, U. S. Geological Survey. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 131

STABLE CARBON ISOTOPES IN CELTIS (ULMACEAE) FROM THE BEAVER CREEK SHELTER, WIND CAVE NATIONAL PARK, SOUTH DAKOTA AND PALEOCLIMATIC INTERPRETATIONS

Mark L. Gabel and Marlina Cowan Biology Department Black Hills State University Spearﬁsh, SD 57799

Larry L. Tieszen Biology Department Augustana College Sioux Falls, SD 57197

ABSTRACT

The Beaver Creek Shelter contains nearly a complete record of sediments from the Holocene. A number of the strata contain endocarps of Celtis (Ul- maceae) fruits. The endocarps were analyzed for ratios of stable carbon isotopes. Delta 13C values from hackberry endocarps at the Beaver Creek site range from 24.01 to 26.18so . The values indicate that there were two periods of lesser water stress in horizons 11 to 14 (3870 to 4710 years before present) and again at horizon eight (>2300 ybp). Comparisons to regional Holocene sites indicates some agreement with data from previous studies and contradic- tions with others.

INTRODUCTION

The Beaver Creek Shelter in Wind Cave National Park contains strata from nearly the entire Holocene, with radiocarbon dates from 1750 ± 60 years before the present (ybp) to 9380 ± 300 ybp (Martin et al., 1993). The shelter has produced remains or artifacts from plants, invertebrates and vertebrates including humans from 22 horizons (Alex, 1991). The vertebrates have been treated by Abbott (1989), Benton (1991) and Martin et al. (1993). The discovery of the site has allowed previously impossible analyses of paleoclimates in the Black Hills have been largely ignored. Grimm (1994) described a dry climate from 8000 to 6000 ybp followed by a wetter climate. An examination of mammals of Lamar Cave from 3200 ybp to 1100 ybp in Yellowstone National Park by Hadley (1996) showed more mesic conditions than at the present. At about 1200 ybp, mammal remains indicate that the climate became warmer and drier. High incision rates between 4500 and 3700 ybp were revealed in a study of the Republican River (Nebraska) by Martin (1992). He correlated to these 132 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) more mesic conditions, and reported more xeric conditions from 1800 to 1100 ybp. Albanese (1990) and Laird et al. (1996) reported similar conditions for the period. Stable isotopes have been used for ecological purposes in numerous studies (Tieszen, 1991) and for archaeological investigations (DeNiro, 1987; 12 13 Tieszen, 1994). The difference in amounts of C and C and C3 plants is determined by the relative signiﬁcance of stomatal conductance (Ci/Ca=internal leaf CO2 concentration/air CO2 concentration) or the activity of the enzyme ribulose-1, 5-biophosphate carboxylase. Tieszen (1991) has shown that numerous ecological factors including temperature and water stress can change ratios of carbon isotopes. Other factors including change in elevation, declining irradiance, and other factors can probably be ignored at the Beaver Creek site since the Celtis were probably not moved over long distances. The C3 plants readily discriminate against 13C in mesic conditions, but to a lesser extent in hotter and drier conditions. Celtis is a genus of the Ulmaceae which consists of large to small trees. The genus currently is found throughout most of the United States, with the most widespread species, C. occidentalis ranging from the east coast to the Rocky Mountains. The species occurs in a variety of habitats from mesic forest to the margins of open grasslands. The endocarps of the fruits are highly mineralized, with calcium carbonate and silica (opal) present in unusual concentrations (Cowan et al., 1997). The purpose of this study is to determine stable carbon isotope values of Celtis endocarps from the Beaver Creek Shelter. The isotope ratios can be used for paleoecological interpretations and for comparison to other interpretations to better understand the environment of the Black Hills during the Holocene.

MATERIALS AND METHODS

Celtis endocarps were obtained from the Beaver Creek Shelter. Samples from eight levels, which were excavated before stratigraphy was completely understood, and from seven horizons were ground to powder. In order to determine δ13C values, one to three pulverized endocarps from each stratum were decarbonated through treatment with 1 N HC1 under vacuum for two hours. The samples were then rinsed twice in distilled water and centrifuged before drying at 60 C for 24 hours. Thirteen of the samples were weighed and loaded into a Carlo Erba elemental analyzer for the determination of relative levels of carbon and nitrogen. After combustion, the resultant gas was separated and introduced to a VG Iso- gas isotope ratio mass spectrometer. Delta13C values were calculated with the following formula:

13C/12Csample – 13C/12Cstandard δ13C = X 1000 13C/12C standard Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 133

RESULTS

The stable carbon istope ratios from the Celtis endocarps were plotted against the horizons described by Alex (1991) in Figure 1. Since not all horizons were dated, and not all horizons contained Celtis in expendable quantities, we plotted radiocarbon dates vs. horizon (Fig. Figure 1. A plot of horizons (Alex, 1991) vs. 2). This nearly linear rela- δ13C for hackberry endocarps from the Beaver tionship then allowed us to Creek Shelter. plot δ13C vs. the actual or extrapolated radiocarbon date (Fig. 3). The values indicate that water stress or some other factor reducing the Ci/Ca ratio was apparent during three periods. The ﬁrst, during the period from about 6000 to 9380 ybp, which includes the Altither- mal period. A second interval of increased water stress occurred at about 3800 ybp, and again in the Figure 2. A plot of horizons (Alex, 1991) vs. latest period represented radiocarbon dates (Martin et al.,) from by the specimens. Con- Beaver Creek Shelter. versely, periods of less water stress (26.0so or less) are indicated from about 4000 to 5000 ybp, and again from about 2000 to 3000 ybp.

DISCUSSION

One possible explanation for the results presented above might be disturbance of the strata, possibly by rodents or other organ- Figure 3. A plot of radiocarbon dates vs. δ13C isms. Agreement between values for hackberry endocarps from Beaver our results and interpreta- Creek Shelter. 134 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) tions from other sites suggests that disturbance is not the cause of the data distribution. Additional literature below support the episode of aridity at 3800 ybp. The results presented above are compared to other studies in Table 1. Grimm's (1994) study from northeast of the Beaver Creek shelter reported climate data, based on pollen, plants, and diatoms from the period from 10,000 ybp to about 4000 ybp. He showed fluctuations, with a tendency to move from more arid more mesic conditions at 4000 to 5000 ybp. Hadley (1996) reported mesic conditions declining to more arid conditions during the last 3000 years based upon mammals from Yellowstone Park, Wyoming. A study of sediments (Martin, 1992) in the valley of the Republican River, Nebraska indicated that incision between 4500 and 3700 is due to a mesic climate, with arid conditions present from 1800 to 1000 ybp. Using similar techniques, Albanese (1990) reported a gradual drop in the water table of the eastern Powder River Basin (Wyoming) in the period from 1600 to 1000 ybp. Laird et al. (1996) in a precise paleoclimate study based upon diatoms from Moon Lake, North Dakota reported four periods of fluctuation, including drying by 7300 ybp, a dry climate from 7300 to 4700 ybp, a “transitional period of high salinity” from 4700 to 2200 ybp, and fluctuating levels of moisture from 2200 ybp. Semken and Falk (1987) described seven Holocene sites in Iowa, and one each in South Dakota and North Dakota. Unfortunately, the Iowa sites were usually limited in time represented, not allowing for broad interpretations of climate. Two sites in southwest Iowa which are in close proximity, the Garrett Farm site (dated at 3600 and 3590 ybp) and the Pleasant Ridge site (1450 ybp) showed a change from grasslands at the Garrett Farm site, which were present in a climate more arid than present (compare to our Fig. 3) at 3600 ybp to more

Table 1. A comparison of various studies in North America indicating mesic (m), intermediate (i), or arid (a) paleoclimates. Years before present (ybp) are rounded to the nearest 1000 years. Note that interpretations have been sim- pliﬁed. ybp TS G94 H96 M92 A90 L96 BB96 F96 B91 A89

9,000 a i m a m a 8,000 a a i a a a 7,000 a a a i a a 6,000 a a a m a a 5,000 i m a m m i 4,000 m m m a m m m 3,000 i m i a m i m 2,000 m m a a a a a 1,000 a a a a a m a a

A89=Abbott, 1989; A90=Albanese, 1990; B91=Benton, 1991; BB96=Bryson and Bryson, 1996; F96=Fredlund, 1996; G94=Grimm, 1994; H=Hadley, 1996; L96=Laird et al., 1996; M92=Martin, 1992; TS=this study Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 135 mesic vegetation at the Pleasant Ridge site. The Modrian site in McKenzie Co., North Dakota (Semken and Falk, 1987) showed that grasslands were dominant in that area from about 4500 ybp to the present, with a slight trend to cooler and moister climates. The Oakwood Lakes site in Brooking Co., South Dakota (16501350 ybp) again represented grassland, indicating a cooler climate than at present and possibly greater climate extremes. In Montana, Graham et al. (1987) reported four sites, two of which, False Cougar Cave and Shield Trap Cave, are located in the Pryor Mountains (south- central to southwestern Montana). These sites included sediments which were dated to time intervals similar to the Beaver Creek site (14,500 to 1690 ybp and 9230 to 620 ybp respectively). The general tendency reported from these caves was toward a warmer climate with more grassland. The other two sites covering abbreviated time periods from the Holmes Terrace (3320 to 1800 ybp) and the Drake site (2610 to 1190 ybp) showed little difference in climate from today. Walker (1987) reported a general drying and warming during the Holocene in sediments from Wyoming. Bryson and Bryson (1996) modeled precipitation from several sites including the Mill Iron site of Ekalaka, Montana. For the period from 14,000 ybp to the present, the model indicated a generally increasing annual precipitation, with a notable decrease in moisture at 11,800, and smaller decreases at about 3800 ybp and 1800 byp. Fredlund (1996) used the geomorphology of Highland Creek in Wind Cave National Park to interpret past climates. He found the period from 14,000 to 9,000 ybp to be a time of geomorphic stability and mesic conditions. Holocene sediments and soils were removed fom 8,000 to 4500 ybp, correlating with the altithermal period. From 4500 to 3600 ybp was a time of landscape stability and soil development, again indicating more mesic conditions. This was followed by a period of landscape instability. Fredlund and Tieszen (1996) in a study of phytolith types and stable carbon isotopes from phytoliths showed the rise of the grassland from 11,000 to 9000 ybp. Their data indicate erosion during the Altithermal (about 8000 to 4500 ybp), and stability from 4500 to 3600 ybp. Their most negative b13C values were before 9000 ybp with another decrease just after 3600 ybp. Benton (1991) and Abbott (1989) produced detailed and comprehensive studies of vertebrate bones from the Beaver Creek shelter site. Benton (1991) reported that in levels 15 through 10 (about 9380 to about 6220), species were more grassland (arid) adapted, while levels 9 to 5 (about 6220 to about 3870 ybp) were dominated by more riparian vertebrates. Abbott (1989) studied vertebrates from the top 11 horizons (about 3870 ybp and younger). Interpretations of the types of vertebrates include ﬂuctua- tions from riparian species to somewhat drier grassland species toward the present. It is interesting to note that Abbott (1989) concluded that there were not major changes in the vertebrates from the Beaver Creek site during the last 3800 years. This is in apparent contradiction to our paleoclimate interpretation based upon data from Celtis isotopes. Our data indicate that in the 3500 to 3800 ybp time period, the Celtis fruits indicate a relatively dry period. 136 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

The Black Hills are indeed an island in the plains, and if the climate did deteriorate at 3800 ybp, mesic adapted species that were in the Black Hills would have probably become trapped if the climate deteriorated. Those that survived, must have, out of necessity, tolerated the more arid conditions. This view is supported by the nearly universal aridity in the last 2000 years, throughout which many of the same species have survived in the Black Hills.

ACKNOWLEDGEMENTS

We gratefully acknowledge the cooperation of the U.S. National Park Ser- vice, J. Martin of the South Dakota School of Mines, and Jim Donohue of the South Dakota Archaeological Research Center.

LITERATURE CITED

Abbott, J.P. 1989. The paleoecology of the late and Post Archaic section of the Beaver Creek Shelter, Wind Cave National Park. M.S. thesis, South Dakota School of Mines. Albanese, J. 1990. The geoarchaeology of the eastern Powder River Basin. State Historic Preservation Ofﬁce, Wyoming Archives Museums, Cheyenne, Wyoming. Alex, L.M. 1991. The archaeology of the Beaver Creek Shelter (39CU779): a preliminary statement. Selections from the Division of Cultural Resources, Rocky Mountain Region, National Park Service, No. 3. Benton, R. 1990. Paleoecology of the Early and Middle Archaic section of the Beaver Creek Rock Shelter, Wind Cave National Park. M.S. thesis, South Dakota School of Mines. Bryson, R.A. and R.U. Bryson. 1996. Sitespeciﬁc highresolution archaeoclimat- ic modeling for the Great Plains. 54th Annual Plains Anthropological Con- ference (in press). Cowan, M. C., M. L. Gabel, A.H. Jahren, and L.L. Tieszen. 1997. Growth and biomineralization of Celtis occidentalis (Ulmaceae) pericarps. American Midland Naturalist (accepted July, 1996). DeNiro, M.J. 1987. Stable Isotopy and Arhaeology. American Scientist 75:182191. Fredlund, G.G. 1996. Late Quaternary geomporphic history of lower Highland Creek, Wind Cave National Park, South Dakota. Physical Geography 17:446464. Fredlund, G.G. and L.L. Tieszen. 1996. Phytolith and carbon isotope evidence for Quaternary vegetation and climate change in the southern Black Hills, South Dakota, USA. Quaternary Research (in press). Graham, M.A., M.C. Wilson, and R.W. Graham. 1987. Paleoenvironments and mammalian faunas of Montana, Southern Alberta and Southern Saskatchewan. pp. 410459 In Graham, M.A., Semken and R.W. Graham (eds.). Late Quaternary mammalian biogeography and environments of the Great Plains, Illinois State Museum 24: 1491. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 137

Grimm, E.C. 1994. LateGlacial and Holocene vegetation change in the northern Great Plains. Island in the Plains Symposium 3, p. 8 (abstract). Hadley, E.A. 1996. Inﬂuence of LateHolocene climate on northern Rocky Mountain mammals. Quaternary Research 46:298310. Laird, K.R., S.C. Fritz, K.A. Maasch, and B.F. Cumming. 1996. Greater drought intensity and frequency before AD 1200 in the northern Great Plains, USA. Nature 384:552554. Martin, C.W. 1992. Late Holocene alluvial chronology and climate change in the central Great Plains. Quaternary Research 37:315322. Martin, J.E., R.A. Alex, and R.C. Benton. 1988. Chronology of the Beaver Creek Shelter, Wind Cave National Park, South Dakota. Proceedings of the North Dakota Academy of Sciences 46:16. Martin, J.E., R.A. Alex, L.M.Alex, J.P. Abbott, R.C. Benton, and L.F. Miller. 1993. The Beaver Creek shelter (39CU779): A Holocene succession in the Black Hills of South Dakota. Plains Anthropologist, vol. 38, no. 145, Memoir 27, pp. 1736. Semken, and Falk. 1987. Late Pleistocene/Holocene mammalian faunas and environmental changes on the northern plains of the United States. pp 176313 In Graham, M.A., Semken and R.W. Graham (eds.). Late Quaternary mammalian biogeography and environments of the Great Plains, Illinois State Museum 24: 1491. Tieszen, L.L. 1991. Natural variations in the carbon isotope values of plants: implications for archaeology, ecology and paleoecology. Journal of Archaeo- logical Science 18:227248. Tieszen, L.L. 1994. Stable iosotpes on the plains: Vegetation analyses and diet determinations. pp 261282 In Skeletal biology in the Great Plains. Owsely, D. W. and R.L Jantz (eds.) Smithsonian Institution Press, Washington, D.C. Walker, D. 1987. Late Pleistocene/Holocene environmental changes in Wyoming. pp. 334397 In Graham, M.A., Semken and R.W. Graham (eds.). Late Quaternary mammalian biogeography and environments of the Great Plains, Illinois State Museum 24: 1491.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 139

COMPARISON OF IMPLEMENTS FOR IMPROVING SODIC RANGELAND SOILS

John Bechtold Natural Resources Conservation Service Santa Maria, CA 93454

F. R. Gartner West River Agricultural Research and Extension Center South Dakota State University Rapid City, SD 57701-3097

E. M. White Plant Science Department South Dakota State University Brookings, SD 57007-2207A

Keywords

slickspots, panspots, natric soils, claypans, water inﬁltration.

ABSTRACT

Unproductive sodic claypan soils were disrupted by ripping, chiseling, ripper-furrowing and furrowing treatments which disturbed 7, 11, 29, and 41% of treated plot surface areas, respectively. Six years after treatment, soils were sampled from untreated plots and from the furrows and interfurrows of mechanically treated plots. Salt content of the 15- to 30-cm soil layer, as measured by electrical conductivities (EC) , decreased up to 5 and 2 dSm-1 in furrows formed, respectively, in sparsely vegetated and vegetated areas. The ripper- furrower and furrower were more effective in causing changes in EC than the chisel or ripper. Nearly all sparsely vegetated areas were vegetated six years after treatment. Mechanical treatments increased both salt leaching and plant densities in sparsely vegetated areas. The sparsely vegetated slickspot areas have not reformed 18 years after they were treated mechanically.

INTRODUCTION

Mechanical treatments increase water inﬁltration (Wight and Siddoway, 1972; Soiseth et al., 1974; Neff and Wight, 1977) and leaching (White et al., 1981) in sodic claypan soils. The purpose of this study was to evaluate the effect of ripping, furrowing, and furrowing-ripping implements on salt distribution in the soil. If water inﬁltration is improved more by implement than the others, the soluble salt content also should decrease more in upper soil layers. 140 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

METHODS AND MATERIALS

Study Areas

Mechanical treatments were applied at two sites in Meade County, South Dakota, on sodic claypan soils with barren areas that have been called slickspots or panspots. Slickspots are barren, shallow, 1- to 5-m-wide depres- sions that pond water because a sodium-dispersed layer limits water infiltration. The Harrington Ranch site (8.1 km north of Rapid City) had Arvada- slickspot soils (Ollila, 1978), and the Kovarik Ranch site (45.1 km east-northeast of Sturgis) had Hisle-slickspot soils (Ollila, 1978). Average annual precipitation is approximately 43 cm and 38 cm, respectively at the Harrington and Kovarik sites. Seventy-eight percent falls from April to September. Both study sites are in the xeric portion of the Northern Mixed Prairie and have western wheatgrass (Agropyron smithii Rydb.) as a component. Arvada and Hisle soils are fine, montmorillonitic, mesic Ustollic Natrargids. Arvada soils have thin silt loam A horizons, clay to clay loam B horizons, and shale at a depth of 100 cm or more. Hisle soils have a thin loam A horizon, clay B horizon, and shaly clay at about 50 cm. Soil pH is neutral in the A horizon, and strongly alkaline in the B and C horizons. The exchangeable sodium percentage is greater than 15% in the B horizon. The Harrington site was on a long, gentle slope where the soil parent materials were uniform across the area. In contrast, the Kovarik site was on a former floodplain where a small intermittent stream meandered and deposited inextensive subsoil layers with pockets of sandy gravel.

Mechanical Treatments

Three mechanical treatments and one untreated control were replicated four times in a randomized complete block design at the Harrington site. Four mechanical treatments and one untreated control were replicated three times in a randomized complete block design at the Kovarik site. Treatments were applied on the contour at both sites in October 1978. Each of the 16 plots at the Harrington site was 9.8 X 23.2 m, and at the Kovarik site the 15 plots were 9.8 X 30.5 m. Ripping: Ripper openings were 5- to 15-cm wide, about 45-cm deep, and 1.5-m apart. Chiseling: Irregular shaped openings were 5- to 15-cm wide, 30-cm deep, and 91 cm apart. This treatment was requested by the owner at the Kovarik site. Furrowing: Frank Sparks, Plevna, Montana, welded left and right plow moldboards together to form a lister-type plow. It formed nearly ﬂat-bottomed furrows 61-cm wide, 8- to 13-cm deep, and spaced on center 1.5-m apart. The furrower was raised brieﬂy every 6 m to create furrow dams to store water. Ripper-furrower: The lister-type furrows were 30- to 40-cm wide, 25-cm deep, and spaced on center 1.2-m apart. The ripper shank penetrated 25- to 30-cm deep. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 141

Initially, the furrows and disturbed soils comprised 7, 11, 29, and 41% of the areas treated with the ripper, chisel, ripper-furrower, and Sparks furrower, respectively. Two furrows in border alleys between treated plots were designed to prevent water from draining from the alleys into treated plots. Like- wise, two furrows were placed upslope of the experimental site to prevent surface water from draining onto treated plots.

Vegetation

Pretreatment aerial photographs, taken from an altitude of 457 m, were used to identify the slickspot and vegetated areas to be sampled. Thus soil sampling in each plot was stratiﬁed into the two categories which reduces the variability within treatments (Steel and Torrie, 1960).

Soil sampling

A hydraulic coring machine was used to extract 8, 100-cm-long, 5.1-cm-diameter soil cores from each mechanically treated plot in 1984. Four were from vegetated and four were from pre-treatment slickspot areas. Two of each four were from furrows and two were from interfurrows. Two cores each were collected randomly from the vegetated and the slickspot areas of the untreated plots. Thus, 112 and 108 cores were extracted at the Harrington and Kovarik sites, respectively. Each soil core was placed in a tray in the ﬁeld and subdi- vided into increments of 0-15, 15-30, 30-45, 45-60, 60-80, and 80-100 cm. Sam- pling depths within furrows were determined relative to the original (untreated) soil surface of adjacent interfurrow areas. Furrow depths in the ripper-furrower treatment were about 15 cm when soils were sampled. Furrows were deeper in a few locations; thus, both the 0-15 and 15-30 cm sample increments were incomplete. The surface dark-colored sediment that had washed into furrows was removed from the upper core increment. In the laboratory, soil samples were air dried, ground (< 2mm), and electrical conductivity was determined (Soil Survey Staff, 1972). Since 0-15 cm samples from furrows were incomplete, they were not included in statistical analyses. Analysis of variance (Steel and Torrie, 1960) was used to determine if the soil electrical conductivities from the different treatments, including controls, were signiﬁcantly different. In this analysis at the Harrington site, six sets of data from furrows and interfurrows of the three mechanical treatments and the data from untreated control plots were compared. At the Kovarik site, eight sets of data and the control plot data were compared. Data in these statistical analyses were from untreated, furrow, and interfurrow positions (P). In a second analysis, only the mechanically treated plots were included so that the effect of the implements could be compared more accurately. The data was analyzed in a factorial arrangement to determine if the mechanical treatments caused different amounts of salt movement. The F test was used to evaluate differences within the main effects of depth (D), vegetation cover (C), i.e., vegetated and sparsely vegetated slickspots, location (L) in furrows or interfurrows, mechanical treatments (T) or their interactions. A Waller-Duncan 142 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

K-ratio T-test (SAS Institute Inc., 1985) was used to compare means within main effects. Statistical signiﬁcance was recognized at P=0.05.

RESULTS AND DISCUSSION

Slickspots at both sites were vegetated six years after the treatments were applied. The western wheatgrass weight percentage of the total vegetation on the treated slickspot and vegetated areas of the plots, respectively, were 30 and 80 at Harringtons and 19 and 34 at Kovariks. Vegetation responses were previously reported (Gartner et al, 1983; Gartner 1984; Gartner and White, 1984). The soil EC values at the Harrington site were significantly lower in furrow locations of the ripper-furrower and furrower than in the interfurrows, ripper treatment, or control plots (Table 1). The ANOVA of soil EC data from treated and untreated control plots indicated significant differences for treatment (T), soil depth (D), and the interactions TXD, CXD, and TXCXD. The interaction can be visualized in Figure 1 where the natural increase in EC with depth is obvious. The mechanical treatments decreased the EC’s in upper layers at different rates in the furrows and interfurrows and in the vegetated areas and un- vegetated slickspots to cause the TXD, CXD, and TXCXD interactions. If only the mechanical treatments (T) are compared at the Harrington site, the ripper-furrower and Sparks furrower treatments have decreased soil EC’s significantly more than the ripper treatment (Table 2, Fig. 2). Significant differences also occurred for D and L and for the interactions TXD, TXL, CXD, and DXL for the same reason as when the untreated control plots are included. The EC values for the soil profiles at the Kovarik site are variable, particularly in the subsoil where the parent material contains thin discontinuous layers of sandy gravel. The EC values for the different treatments (T) were not significantly different, but the soil depth (D) and TXD and TXCXD interactions were significantly different. The TXD interaction is apparent in Figure 3 where

Table 1. Comparison of effects of mechanical treatment (including untreated control) on the electrical conductivity of rangeland claypans, Harrington site.a

EC N Treatment Location (dsm-1)

8.6A 80 Ripper Interfurrow 8.5A 80 Control 8.4A 80 Sparks Furrower Interfurrow 8.3A 80 Ripper Furrow 7.6A 80 Ripper-Furrower Interfurrow 6.4B 80 Ripper-Furrower Furrow 6.2B 80 Sparks Furrower Furrow aMinimum signiﬁcant difference = 1.1 Means with same letter are not signiﬁcantly different (P ≤ 0.05). Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 143

Figure 1. Mean soil EC of proﬁles from furrow and interfurrow locations with vegetated and sparsely vegetated cover at the Harrington site (n = 8).

Table 2. Comparison of effects of mechanical treatment (excluding control) on the electrical conductivity of rangeland claypans, Harrington site.a

EC N Treatment (dsm-1)

8.4A 160 Ripper 7.3B 160 Sparks Furrower 7.0B 160 Ripper-Furrower aMinimum signiﬁcant difference = 0.806 Means with same letter are not signiﬁcantly different (P ≤ 0.05).

the ripper-furrower and Sparks furrower reduced the EC’s in the upper part of the soil more than the other treatments. The soil EC’s for the treated plots, without including the untreated controls in the analysis, were signiﬁcantly different for D, TXD, TXL, and DXL. If the untreated control data is ignored in Figure 3, the differences between treatments and depths or locations in the furrow or interfurrow are apparent. When the EC’s for furrows and interfurrows of each treatment are combined (Fig. 4), the overall effect in reducing EC’s is largest for the furrowing implements in the upper soil layers, the same as they were for the Harrington site. 144 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Effects mechanical treatments have on the soil.

Mechanical treatments create wind-exposed interfurrows and wind-protected furrows which can collect snow in the winter and runoff water in the summer (Neff and Wight, 1977; White et al., 1981). Fur- rows have a more humid Figure 2. Mean soil EC of profiles, irrespec- microclimate than the inter- tive of location and cover, at the Harrington furrows. Thus, desiccation site (n = 22). cracks will form in a moist furrow as plant roots remove water from the adjacent subsoil of interfurrows. This is because moist soil can be pulled apart more easily than the drier soil which has a higher tensile strength. Loose granular peds in the interfurrow soil surface generally prevent water from flowing directly into subsoil cracks that extend to the surface. This granular surface layer has been dis- placed from the furrows so that a desiccation crack is more likely to be open at the furrow surface. Thus, water may drain directly in- to the subsoil (White, 1986). The EC,s of upper soil layers of mechanically treated areas decrease in Figure 3. Mean soil EC of profiles from fur- approximately the same or- row and interfurrow locations, irrespective of der as the furrow width in- cover, at the Kovarik site (n = 12). creases, i.e., ripper < chisel, < ripper-furrower, < Sparks furrower. Cracks that form in soils of furrows may duplicate the natural cracks in the vegetation transition zone surrounding barren slickspots (Hopkins et al., 1991). Western wheatgrass has a deeper root system than blue grama or buffalograss Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 145

(Weaver and Darland, 1949). Thus, the increase in western wheatgrass density in furrowed plots likely creates more subsoil channels (Baver, 1948 p. 184) for water drainage and salt leaching (Hopkins et al., 1991). Large natural cracks around prisms at a depth of 50 to 75 cm are about 50 Figure 4. Mean soil EC of proﬁles, irrespec- cm apart. Thus, furrow tive of location and cover, at the Kovarik site width of about 61 cm and (n = 24). 35 cm, and interfurrow widths of 85 to 90 cm, appear to be optimal for subsoil cracks to form (White, 1986). Cracks may form in wider furrows, but likely the ripper-furrow would disrupt the subsoil sufﬁciently that the crack would continue to reform as long as the furrow microclimate is more humid than that of the interfurrows.

CONCLUDING COMMENTS

Mechanical treatments improved barren slickspot areas associated with vegetated sodic claypan soils. Rhizomatous western wheatgrass can invade rapidly into the slickspot areas, create root channels, and increase water inﬁl- tration. Soil EC’s are reduced, which in turn reduces the possible reformation of the slickspots. Slickspots had not reformed when soils were sampled six years after the mechanical treatments were applied and are not reforming after 18 years. Unreplicated demonstrations of mechanical treatments of slickspot areas caused improvements which have persisted some 30 years. Furrowing mechanical treatments of slickspot areas can be recommended with considerable conﬁdence that they will cause long-lived improvement.

REFERENCES CITED

Baver, L.D. 1948. Soil physics. John Wiley and Sons, Inc. New York. Gartner, F.R. 1984. Progress in non-structural range improvement in the North- ern Great Plains—future needs. Proc. Vegetative Rehabilitation & Equip- ment Workshop. 38th Ann. Rep., USDA, Forest Serv. 2200-Range. pp. 57- 60. Gartner, F.R. and E.M. White. 1984. Range renovation. South Dakota State Univ., Cow-Calf Day. pp. 25-34. Gartner, F.R., E.M. White and R.I. Butterﬁeld. 1983. Improving forage production on claypan soils. South Dakota State Univ., Animal & Range Sci. Dep., Field Day Rep.-Antelope. 6 p. 146 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Hopkins, D.G., M.D. Sweeney, D.R. Kirby, and J.L. Richardson. 1991. Effects of revegetation on surficial soil salinity in panspot soils. J. Range Manage. 44:215-220. Neff, E.L. and J.R. Wight. 1977. Overwinter soil water recharge and herbage production as influenced by contour furrowing on eastern Montana rangelands. J. Range Manage. 30: 193-195. Ollila, T.J. 1978. Soil survey of Meade County, South Dakota, southern part. USDA, Soil Conserv. Serv. and Forest Serv. in cooperation with So. Dak. Agri. Exp. Sta. SAS Institute, Inc. 1985. SAS User's Guide: Statistics. Version 5 Ed. SAS Insti- tute, Inc., Cary, NC. Soil Survey Staff. 1972. Soil survey laboratory methods and procedures for collecting soil samples. Soil Survey Investigations Rep. No. 1. USDA, Soil Conserv. Serv., Revised April 1972. 63 p. Soiseth, R.J., J.R. Wight and J.K. Aase. 1974. Improvement of panspot (solonet- zic) range sites by contour furrowing. J. Range Manage. 27:107-110. Steel, R.G.D. and J.H. Torrie. 1960. Principles and procedures of statistics. Mc- Graw-Hill Book Co., Inc. New York, NY. Weaver, J.E. and R.W. Darland. 1949. Soil-root relationships of certain native grasses in various soil types. Ecological Monographs 19:303-338. White, E.M. 1986. Longevity and effect of tillage-formed soil surface cracks on water infiltration. J. Soil Water Conserv. 41:344-347. White, E.M., F.R. Gartner, and R. Butterfield. 1981. Range claypan soil improvement: Response from furrowing and ripping in northwestern South Dakota. J. Range Manage. 34:119-125. Wight, J.R. and F.H. Siddoway. 1972. Improving precipitation use efficiency on rangeland by surface modification. J. Soil Water Conserv. 27:170-174. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 147

ATRAZINE AND ALACHLOR ADSORPTION CHARACTERISTICS TO BENCHMARK SOIL SERIES IN EASTERN SOUTH DAKOTA

Z. Liu, S.A. Clay, J. Gaffney, and D. Malo Plant Science Department South Dakota State University, Brookings, SD 57007

Abstract

Corn, grain sorghum, and soybean are grown on about six million acres in eastern South Dakota each year. Two herbicides used routinely for weed control are atrazine(6-chloro-N-ethyl-N’-(1-methylethyl)-1,3,5-triazine-2,4-diamine) in corn and grain sorghum and alachlor (2-chloro-N-(2,6-di- ethylphenyl)-N-(methoxymethyl)acetamide) in all three crops. Six benchmark soil series that include a majority of the cropped acres treated with these herbicides are the Egan, Moody, Nora, and Brandt silty clay loams, and Clarno and Enet loams. Batch adsorption studies determined atrazine and alachlor binding characteristics to these soils and aids in assessing the amount of herbicide available for movement. These data also provide a basis for future use and management decisions for these and other related herbicides in similar soils. Soils from three horizons (A, B, and C) for each soil type were treated with atrazine or alachlor at four herbicide concentrations. Atrazine and alachlor sorption partition coefficients differed most in A horizon soils and ranged from 2.16 to 5.35 µmol1-1/n L1/nKg-1 for atrazine and 1.95 to 5.78 µmol1-1/n L1/nKg-1 for alachlor. Atrazine binding to A horizon soils ranked as Brandt >Egan = Moody > Enet = Clarno > Nora. Alachlor binding to A horizon soils ranked as Brandt >Moody > Nora > Enet > Clarno. B and C horizon soils had lower binding for both herbicides; the sorption partition coefficient for atrazine ranged from 0.12 to 1.9 µmol1-1/n L1/nKg-1 while alachlor ranged from 0.43 to 1.64 µmol1-1/n L1/nKg-1. These data indicate that some soil types would be more susceptible to herbicide leaching than others. Once the herbicide moves through the A horizon, it may move rapidly through the lower soil profile (because of the decrease in binding capacity), and therefore, increase the vulnerability of the aquifer to contamination. Best management practices for these herbicides are being investigated to limit their movement through soil.

INTRODUCTION

Since the late 1940’s, agricultural production in the United States has de- pended on the use of chemicals ( e.g. herbicide and insecticides) to control weeds and other pests. From 1992 to 1994, about 97 % of corn and soybean acreage in the U.S. were treated with one or more herbicide applications (Agricultural Statistics, 1995-1996). Short term benefits of herbicide ap- 148 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) plication are weed control and increased crop yields. However, long term effects of herbicides on the environment have largely been ignored in crop production. Corn, grain sorghum, and soybeans are grown on about 6 million acres in eastern South Dakota each year. Atrazine (6-chloro-N-ethyl-N’-(1- methylethyl)-1,3,5-triazine-2,4-diamine) and alachlor (2-chloro-N-(2,6-di- ethylphenyl)-N-(methoxymethyl)acetamide) are two herbicides which are relatively inexpensive and used routinely for weed control for corn, grain sorghum, or soybean in eastern South Dakota. In a survey of 16 eastern South Dakota counties conducted in 1990 and 1991, about 510,000 acres of corn were treated with atrazine, and 630,000 acres of corn and soybean were treated with alachlor (Table 1). Undesirable environmental side effects such as detectable concentrations of herbicides in groundwater have become major national social and scientific concern. For example, US Environmental Protection Agency set primary drinking water standards at maximum contaminant levels of 3 ppb for atrazine and 2 ppb for alachlor (Code of Federal Regulations, 1993). In 1990, the National Survey of Pesticides in Drinking Water Wells found that about 10.4% of the community water supply wells and 4.2% of the rural domestic wells contain pesticides or pesticide degradates at or above the minimum reported limit used in the survey (Cohen, 1992). Both atrazine and alachlor have been detected in well water surveys conducted in Iowa, Min- nesota, Nebraska and South Dakota (Hallberg, 1989 ). Factors influencing herbicide detection in groundwater include point source problems, such as spills and back siphoning into wells, and nonpoint source contributions. Nonpoint source contamination may be from herbicides applied at normal use rates with low level herbicide movement through soil profiles and into aquifers. Herbicide movement into aquifers may be influenced by: 1) herbicide physical and chemical properties; 2) soil physical and chemical properties such as soil texture, soil permeability, soil hydraulic properties, depth to the aquifer, soil pH and organic carbon content; 3) environmental parameters such as rainfall; and, 4) crop management such as tillage. Herbicide movement through soil is affected by its binding capacity to soil. Soil properties ( e.g., organic matter, pH, and clay content) have a significant effect on herbicide adsorption to soil (Koskinen and Harper, 1990). In general, as herbicide binding capacity to soil increases, herbicide movement through the soil decreases. Kladivko et al. (1991) reported that the concentration of herbicides in subsurface drainflow corresponded to the rank-order of batch adsorption coefficients measured in the laboratory. The objective of this study was to investigate atrazine and alachlor adsorption characteristics for benchmark soils in eastern South Dakota. The data obtained will aid in assessing the amount of herbicide available for movement and, ultimately, the vulnerability of groundwater to contamination by these herbicides. These data also provide a basis for future use and management decisions for these and other related herbicides in similar soils. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 149

MATERIALS AND METHODS

Adsorption

Six benchmark soil series used in this study were collected from five eastern counties of South Dakota. Four silty clay loam (scl) and two loam (l) soils were chosen and included Egan scl (fine-silty, mixed, mesic Udic Hap- lustolls) and Nora scl (fine-silty, mixed, mesic Udic Haplustolls) from Lincoln county, Moody scl (fine-silty, mixed, mesic Udic Haplustolls) from Minnehaha county, Brandt scl (fine-silty, mixed, mesic Udic Haplustolls) from Brookings county, Clarno l (fine-loamy, mixed mesic Udic Haplustolls) from Hutchinson county and Enet l (fine-loamy over sandy or sandy-skeletal, mixed, Mesic Pachic Haplustolls) from Clay county. The soil was collected by horizon, air- dried, and passed through a 2-mm sieve. Herbicides were dissolved in methanol(< 1 mL) and then diluted to final concentration of 2.5, 5.0, 10.0, or µ -1 20.0 mol L of herbicide in 0.01 M CaCl2. Each solution then was spiked with uniformly-ring-labeled-14C-atrazine/alachlor in the range of 0.1 to 1.2 kilobecquerel (KBq). Ten g of soil were placed in a centrifuge tube and 5 14 mL of 0.01 M CaCl2 solution and 5 mL of C- herbicide solution were added. After herbicide addition, a one-day batch equilibration study was conducted by: (i) mechanically shaking the treated soil slurry for 24 h at 25ºC; (ii) centrifuging the slurry at 8,000 r.p.m (approx. 8,000 g) for 20 minutes to separate supernatant and soil; (iii) pipetting out the supernatant, and (iv) adding scintillation cocktail and counting 14C in solution using a Packard 1600TR Liquid Scintillation Analyzer. The amount of herbicide adsorbed to soil was determined by the difference between the amount radioactivity in the initial solution and the radioactivity in the solution after soil equilibration. The linear form of the Freundlich equation was used to describe herbicide sorption to soil (Koskinen and Harper, 1990) . The linear form of the Fre- undlich equation is :

log [Cs] = Log Kf + 1/n Log [Ce]; µ where, Cs is the mol herbicide adsorbed to per kg soil;

Ce is the umol herbicide in per L of supernatant solution after equilibration; Kf and 1/n are empirical constants.

Data Analysis

Freundlich sorption isotherm coefﬁcients were calculated by the least squares technique using the log-transformed equilibrium data. Statistical evaluation included comparison of the slopes and intercepts of the regression lines, and calculation of the 95% conﬁdence intervals for the intercept (log Kf ) and slope (1/n). 150 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

RESULTS AND DISCUSSION

Atrazine and Alachlor usage. The acres of corn and soybean in 16 counties of eastern South Dakota treated with atrazine and alachlor in 1990 and 1991 are listed in Table 1. The percentage of corn ﬁelds treated with atrazine ranged from 5% to 75%. Lincoln (75%), Union (60%), Moody (50%) and Minnehaha (45%) counties had the highest percentages of corn treated with atrazine. The percentage of corn and soybean ﬁelds treated with alachlor ranged from 5% to 50%, with the highest percentage of use in Mc- Cook and Clay (50%), Hutchinson (40%), Lake and Spink (33%), and Moody (30%) counties. The combined atrazine and alachlor usage was over 1.1 million acres, and accounted for 42% of total corn and soybean acres. The survey also estimated that the usage of atrazine and alachlor from 1991 to 1996 would remain stable or even increase in Moody, Minnehaha, Lincoln, and Hutchinson counties. Soil Characteristics. The six soil series from eastern South Dakota were chosen because they represent a large portion of the acreage treated with alachlor and/or atrazine (Table 2). The physical characteristics of these soils are reported Table 3. The soil pH in A horizon soil ranged from 5.50 to 7.21

Table 1. 1990–1991 survey results of atrazine and alachlor usage in 16 eastern South Dakota counties1.

County Acres % of corn acres Acres of Corn % of acres of corn of Corn treated with + Soybean and soybean treated Atrazine with Atrazine

Lincoln 133,000 75 268,000 5 Union 107,000 60 195,000 20 Moody 118,000 50 118,000 30 Minnehaha 150,000 45 260,000 17.5 Spink 128,000 40 219,000 33 Lake 108,000 33 186,000 33 McCook 95,000 20 180,000 50 Hutchinson 151,000 15 260,000 40 Clay 84,000 7.5 170,000 50 Lyman 13,900 45 13900 17 Turner 130,000 20 262,000 20 Charles Mix 71,000 17 not determined not determined Brown 130,000 10 130,000 10 Yankton 228,000 5 228,000 5 Brookings 127,000 7.5 127,000 4.5 Tripp 33,000 15 36,000 0.5

1 Surveyed by Jim Gaffney, former graduate research assistant, Personal Communication with S.A. Clay Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 151

Table 2. The percent coverage of the six benchmark soil series chosen for this study across eastern South Dakota.

Soil Association Counties Covered % Coverage

Clarno/Crossplain-/Bonilla-/Prosper-/ loam McCook, Hutchinson 36.3

Egan/Wentworth-/Beadle-/silty clay loam Lake, Lincoln, Moody 14.6

Moody/Nora-/silty clay loam Minnehaha, Moody 11.1

Nora/Moody-/Crofton-/silty clay loam Minnehaha, Union 12.8

Enet loam Clay <1

Brandt/Estilline silty clay loam Brookings, Codington, 5 Deuel, Hamlin, Moody, Minnehaha

Table 3. Selected soil properties for six benchmark soils in eastern South Dakota.

Soil Series Sampled Horiz Depth pH OC Sand Silt Clay County (in) (1:1) % % % %

Moody silty Minnehaha Ap 0-7 6.39 2.47 9.65 56.9 34.6 clay loam Bw 7-17 7.17 0.55 21.0 54.6 24.6 Bk 30-42 7.96 0.27 25.7 56.6 17.8 Ck 42-56 8.20 0.22 7.45 75.1 17.5

Nora silty Lincoln Ap 0-7 7.21 2.02 9.2 62.0 28.9 clay loam Bw 7-18 7.40 0.51 11.1 62.3 26.7 Bk 18-30 7.76 0.24 10.5 65.3 25.3

Clarno loam Hutchinson Ap 0-9 6.56 1.51 30.4 40.6 29.1 Bw 9-16 7.01 1.10 31.9 38.4 29.6 Bk 16-36 7.84 0.54 36.7 38.7 24.6 C 36-60 7.96 0.05 37.0 39.0 24.2

Egan silty Lincoln Ap 0-8 5.81 3.25 7.55 55.9 36.6 clay loam Bw 8-25 6.61 2.22 4.25 61.2 34.6 Bk(2C) 25-30 7.93 0.84 9.9 53.4 36.8 2C(Bk) 30-60 7.91 0.22 30.3 37.3 32.5

Enet loam Clay Ap 0 -6 6.6 2.6 27.5 46.5 26.0 AB 6 - 12 6.0 1.7 30.4 41.4 28.2 B 12 -18 6.3 1.17 23.2 49.4 27.4

Brandt silty Brookings A 0 - 16 5.5 4.5 16.2 54.4 29.4 clay loam B 16 -51 6.2 1.0 11.8 56.5 31.7 C 51+ 7.9 0.2 17.2 52.9 29.5 152 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) and ranked as Nora scl>Enet l = Clarno l = Moody scl > Egan scl > Brandt scl. The soil pH ranged from 6.61 to 7.96 in B horizon soil and 6.20 to 8.25 in C horizon soil. The % carbon (%C) in A horizon soil ranged from 1.53 to 4.50 and ranked as Brandt scl > Egan scl > Enet l = Moody scl > Nora scl > Clarno l. The %C ranged from 0.24 to 2.84 in B horizon soil and 1.50 to 3.31 in C horizon soil. The % sand content in A horizon ranged from 7.55 to 30.4 and ranked as Clarno l > Enet l > Brandt scl > Moody scl = Nora scl > Egan scl. Atrazine sorption. Atrazine sorption characteristics for the six soils chosen for this study are reported in Table 4 and Figures 1 & 2. The Kf value is the amount of herbicide adsorbed to the soil when the amount in solution is 1 µmol L-1. It is an index of herbicide binding capacity to soil with lower Kf values indicating less herbicide sorbed to soil. The atrazine Kf values differed most in A horizon soils and ranged from 2.16 to 5.35. Atrazine bind-

Table 4. Atrazine Adsorption Isotherms for six benchmark soils in eastern South Dakota.

Soil type Sampled Horiz Depth Kf* l/n** r2 County (in)

Moody silty Minnehaha Ap 0 - 7 3.72 (3.47 -3.99) 0.86 (0.03) 0.99 clay loam Bw 7 - 17 0.92 (0.89 -0.94) 0.94 (0.01) 0.99 Bk 30 - 42 0.63 (0.56 -0.70) 0.88 (0.05) 0.98 Ck 42 - 56 0.56 (0.53 -0.60) 0.96 (0.03) 0.99

Nora silty Lincoln Ap 0 - 7 2.16 (2.06 - 2.26) 0.90 (0.02) 0.99 clay loam Bw 7 - 18 1.02 (0.996 - 1.06) 0.89 (0.01) 0.99 Bk 18 - 30 0.65 (0.60 -0.69) 0.94 (0.03) 0.99

Clarno loam Hutchinson Ap 0 - 9 2.80 (2.62 - 2.98) 0.85 (0.03) 0.99 Bw 9 - 16 1.57 (1.51 -1.63) 0.91 (0.02) 0.99 C 36 - 60 0.61 (0.56 -0.65) 0.91 (0.03) 0.99

Egan silty Lincoln Ap 0 - 8 4.13 (3.68 -4.64) 0.89 (0.05) 0.98 clay loam Bw 8 - 25 1.91 (1.78 -2.05) 0.88 (0.03) 0.99 Bk 25 - 30 0.95 (0.88 -1.04) 0.93 (0.04) 0.99

Enet loam Clay Ap 0 - 6 2.93 (2.81 - 3.06) 0.80 (0.01) 0.99 AB 6 - 12 2.29 (2.12 -2.47) 0.80 (0.01) 0.99 Bw 12 -18 0.81 (0.73 -0.90) 0.84 (0.01) 0.99

Brandt silty Brookings A 0 - 16 5.35 (5.16 -5.55) 0.82 (0.01) 0.99 clay loam B 16 - 51 1.06 (0.98 - 1.14) 0.87 (0.01) 0.99 C1 51 + 0.12 (0.08 - 0.18) 0.93 (0.05) 0.97

* Numbers in parentheses are the 95% conﬁdence interval (CI) for Kf. ** Numbers in parentheses are the standard error of 1/n. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 153 ing to A horizon soils ranked as Brandt scl > Egan scl = Moody scl > Enet l = Clarno l > Nora scl. These data indicate that about twice as much atrazine is in solution of the Nora scl and Clarno l than Egan and Brandt scl. The lower atrazine sorption to the Nora scl and Clarno l may be due to the higher soil pH and lower organic carbon than Brandt scl. A negative correlation between Kf and pH was calculated for A Figure 1. Atrazine sorption in A horizon soils. horizon soils (r = -0.96; p<0.05) (Fig. 3 and Table 5). These data indicate that as soil pH increases atrazine sorption decreases. Soil pH has been shown to affect atrazine binding to soil with more sorption to high than low pH soils (Clay et al., 1988; Liu et al.,1995). This is due to the fact that atrazine is slightly basic and at lower pH, the molecule has slightly positive charge, resulting in greater sorption. A positive correlation of Kf with organic carbon content (r = 0.91 p<0.05) was observed (Fig. 4 and Table 5) so that as %C increases so did atrazine sorption. The high sand content of the Clarno l also may have contributed to lower atrazine sorption than in Figure 2. Atrazine sorption in six benchmark other soils. soils. Atrazine binding to soil decreased as the soil 154 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) depth increased most likely due to the increase in soil pH and sand content. The atrazine Kf value ranged from 0.63 to 1.91 in B horizon soils and 0.12 to 0.61 in C horizon soils. The lower Kf values in B and C horizons indicate that once the herbicide moves through the A horizon soil, it is less sorbed, and perhaps more readily transported through the lower soil horizons. The 1/n value for all six soil series were less than 1 and ranged from Figure 3. Correlation of atrazine Kf with soil 0.79 to 0.94. No signiﬁ- pH. cant differences in 1/n values were noted among soils or soil horizons studied. The 1/n value is the slope of the herbicide sorption isotherm. When 1/n is less than 1, this indicates that the amount of herbicide sorbed to soil is dependent of initial herbicide concentration in solution with less herbicide sorbed to soil as the initial concentration increases. However, when 1/n = 1, the percent of herbicide sorbed to soil is independent of initial concentration in solution. Figure 4. Correlation of atrazine Kf with soil OC%.

Table 5. Correlation of Kf values with pH and organic carbon content for A horizon soils. Kf atrazine Kf Alachlor rr pH -0.96 -0.86 % Organic Carbon 0.91 0.93 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 155

Alachlor sorption: Alachlor sorption characteristics for ﬁve soil series (Egan scl not included) in eastern South Dakota are reported in Table 6 and Figures 5 & 6. The alachlor Kf values differed most in A horizon soils and ranged from 1.95 to 5.78. Alachlor binding to A horizon soils ranked as Brandt scl > Moody scl >Nora scl > Enet l > Clarno l. These data indicate that about three times as much alachlor is in solution in the Clarno l than Brandt scl. A positive correlation of Kf with organic carbon content for A horizon soils was calculated (r = 0.93, p<0.05) (Table 5 & Fig. 7) and a negative correlation of Kf with pH was observed (r=-0.86, p=0.09) (Table 5). The positive correlation with organic carbon agrees with Peter and Weber (1985) who reported that soil organic matter content was highly correlated with sorption of alachlor. Alachlor sorption was not as strongly affected by pH as atrazine because alachlor is a neutral molecule. Alachlor binding to soil decreased rapidly as the soil depth increased. The alachlor Kf value ranged from 1.06 to 1.64 in B horizon soils and 0.43 to 0.77 in C horizon soils. The lower Kf values in B and C horizons indicate that less alachlor was sorbed to B and C horizon soils and if alachlor moved

Table 6. Alachlor Adsorption Isotherms for six benchmark soils in eastern South Dakota.

Soil type Sampled Horiz Depth Kf * 1/n** r2 County (in)

Moody silty clay loam Minnehaha Ap 0 - 7 3.36 (3.04 -3.70) 0.72 (0.01) 0.99 Bw 7 - 17 1.63 (1.19 -2.24) 0.64 (0.05) 0.97 Bk 30 - 42 1.14 (0.78 -1.67) 0.71 (0.06) 0.96

Nora silty clay loam Lincoln Ap 0 - 7 2.72 (2.53 - 2.94) 0.76 (0.01) 0.99 Bw 7 - 18 1.55 (1.43 - 1.68) 0.73 (0.01) 0.93 Bk 18 - 30 1.18 (0.70 -2.00) 0.85 (0.09) 0.99

Clarno loam Hutchinson Ap 0 - 9 1.95 (1.68 - 2.27) 1.06 (0.07) 0.97 Bw 9 - 16 1.64 (1.50 -1.79) 0.91 (0.04) 0.99 C 36 - 60 0.77 (0.73 - 0.80) 0.98 (0.03) 0.99

Enet loam Clay Ap 0 - 6 2.39 (2.29 - 2.50) 0.75 (0.01) 0.99 AB 6 - 12 2.10 (1.86 -2.37) 0.73 (0.02) 0.99 Bw 12 -18 1.06 (0.86 -1.30) 0.76 (0.03) 0.98

Brandt silty clay loam Brookings A 0 - 16 5.78 (4.94 -6.75) 0.72 (0.01) 0.99 B 16 - 51 1.88 (1.69 - 2.10) 0.86 (0.01) 0.99 C1 51 + 0.43 (0.25 - 0.71) 0.72 (0.12) 0.90

* Numbers in parentheses are the 95% confidence interval (CI) for Kf. ** Numbers in parentheses are the standard error of 1/n. 156 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) out of the A horizon, it has the potential to move rapidly through the lower soil profile. The 1/n value for four of the five soil series were < 1. The exception was that Clarno l had an 1/n of 1.06, which indicated that the percent of alachlor adsorbed to soil is independent of initial alachlor concentration in solution.

CONCLUSIONS

The A horizon of the Enet l and Clarno l were Figure 5. Alachlor sorption in A horizon the two soil types that had soils. the lowest binding capacity for atrazine and alachlor while Brandt scl had the highest binding capacity for both atrazine and alachlor. These data indicate that the loam soils may be more susceptible to atrazine and alachlor movement through the A horizons than silty clay loam soil. Ranking of the soils for sorption of both herbicides was similar except the Nora scl that had less binding of atrazine than alachlor. The low sorption capacity of the Nora scl for atrazine is most likely due to the relatively high soil pH compared to the other soils. Due to the decrease of atrazine and alachlor binding capacity to B and C horizon soils, once the herbicides move through Figure 6. Alachlor sorption in ﬁve bench- the A horizon, they have mark soils. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 157 the potential to move rapidly through the B and C horizons. Herbicide degradation is slower (2 to 1000 times slower) as depth in the soil increases (Clay et al., 1997; Yen et al., 1994; Pothuluri et al., 1990). The low sorption and slow degradation properties increase the potential of these herbicides to contaminate shallow aquifers. Therefore, best management practices for these herbicides must be investigated to limit their movement through soil. Figure 7. Correlation of alachlor with soil OC%.

REFERENCE

Calvet, R. 1980. Adsorption-desorption phenomena. pp 1-25. In R.J. Hance (ed.) Interactions between herbicides and soil. Academic Press Inc. NY. Clay, S. A., T. B. Moorman, D.E. Clay, and K. A. Scholes. 1997. Sorption and degradation of alachlor in soil and aquifer material. J. Environ. Qual. 26:1348-1353. Clay, S.A., W.C. Koskinen, R.R. Allmaras, and R.H. Dowdy. 1988. Differ- ences in herbicide adsorption of soil using several soil pH modification techniques. J. Environ. Sci. Health. B23:559-573. Cohen, S. 1992. Results of the National Drinking Water Survey: Pesticides, nitrate and well characteristics. Water Well Journal. 35-38 Hallberg. G.R. 1989. Pesticide pollution of groundwater in the humid United States. Agric, Ecosyst and Environ, 26:299-367 Hassett, J.J., W.L. Banwart, and R.A. Griffin. 1983. Correlation of compound properties with sorption characteristics of nonpolar compounds by soils and sediments: Concepts and limitations. p161-178. In C.W. Francis and S.I. Auerbach (eds.) Environment and solid waste. Butterworths, Boston. Kladivko, E.J., G.E. Van Scoyoc, E.J. Monke, K.M. Oates, and W. Pask. 1991. Pesticides and nutrient movement into subsurface tile drains on a silt loam soil in Indiana. J. Environ. Qual. 20:264-270. Koskinen, W.C. and S.S, Harper. 1990. The retention process: mechanism. p51-78. In H.H Cheng (ed.) Pesticides in the Soil Environment. SSSA Book Ser. SSSA, Madison, WI. Liu, Z., S. A. Clay, D. E. Clay and S. S. Harper. 1995. Ammonia fertilizer influences atrazine adsorption-desorption characteristics. J. Agric. Food Chem. 43:815-819. 158 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Peter, C. J., and J. B. Weber. 1985. Adsorption, mobility and efficacy of alachlor and metolachlor as influenced by soil properties. Weed Sci. 33:874-881. Pothuluri, J. V., T. B. Mooman, D. E. Obenhuber, and R. D. Wauchope. 1990. Aerobic and anaerobic degradation of alachlor in samples from a surface- to-groundwater profile. J. Environ. Qual. 19:525-530. United States Department of Agriculture. Agricultural Statistics. 1995 – 1996. US EPA. Code of Federal regulations. 1993. 40 CFR, 141.50-51. P. 688. Yen, P. Y., W. C. Koskinen, and E. E. Schweizer. 1994. Dissipation of alachlor in four soils as influenced by degradation and sorption processes. Weed Sci. 42:223-240 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 159

ANALYSIS OF SPATIAL DISTRIBUTION OF CANADA THISTLE (CIRSIUM ARVENSE) IN NOTILL SOYBEAN (GLYCINE MAX)

B.L. Broulik, J. Lems, S.A. Clay, and D.E. Clay Plant Science Department South Dakota State University Brookings, SD 57007

M.M. Ellsbury USDAARS Northern Grain Insect Laboratory Brookings, SD 57006

ABSTRACT

The nonuniform spatial distribution of weeds across a field landscape com- plicates sampling and modeling, but allows site specific rather than broadcast management of weed populations. Where weeds are aggregated, densities measured at random locations are not independent, but rather spatially related or autocorrelated. Geostatistical methods were used to describe and map non- random distribution and variation of shoot density across ten well established patches of Canada thistle, a perennial weed, in a 65 hectare notillage soybean field in Moody county, South Dakota in 1996. Canada thistle densities were determined by counting the number of shoots present in a 20 by 50 cm (0.1m2) rectangle. Shoot densities were recorded at 3.04 m increments in 8 .directions from the center of each patch using adaptive sampling. The boundary of the thistle patch on each axis was arbitrarily defined as having 2 consecutive measurements of 0 shoots per 0.1 m2. Contour maps of weed densities were generated and overlaid on field topography maps. A contour map was generated to estimate the size and density of each thistle patch. Generally, the highest densities of Canada thistle appear in the center of the patches. Shoot density within the patches declined as the distance from the center of the patch increased. Near infrared images were generated with a digital camera and compared to weed maps produced with ground scouting.

INTRODUCTION

Environmental and social concerns about pesticide use has led to an interest in more sustainable farming systems with a reduced reliance on pesticides. Current research focusing on integrated pest management practices has demonstrated that a reduction in herbicide use must be accompanied by other control and management technologies to decrease weed infestation level and fitness (16). To date these approaches have been applied uniformly across fields. An alternative approach is to manage with rather than overcome spatial 160 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) heterogeneity (16). The importance of spatial distribution in sampling weed populations, modeling population dynamics, and for longterm weed management, has drawn attention to the need for reliable methods to describe and analyze spatial distribution of weeds in the field landscape. Recent interest in spatial distribution of weeds represents a new direction for weed science with the potential to enhance the understanding of how weed populations develop, persist, and change across the landscape (3). Our ability to precisely determine the location and density of weed patches in the field will provide farmers with technologically superior methods to topographically map weed populations and the ability for more accurate herbicide decisions, which will have both economic and environmental benefits as the herbicides are only applied where required as opposed to blanketing the entire field. Many existing models and management strategies assume weed populations are not spatially structured, but randomly distributed, despite the fact that most fields are managed homogeneously. However, many perennial species such as Canada thistle grow in patches (7). Canada thistle is a perennial weed with an extensive, spreading root system (4,7). Adventitious root buds arise from its roots to form new secondary shoots. This is the primary method of reproduction for Canada thistle after seedling establishment. The spread of horizontal creeping roots from one parent can result in dense round patches of shoots typically encountered in the field. Roots from one established plant can spread over a circular area of 1020 ft in diameter in one year (6). Many integrated control programs for Canada thistle have been proposed which likely would require 8 to 10 years of effort if implemented properly (8). Control measures need to be repeated in a time- ly manner to be effective. Increases in Canada thistle densities are greatest under minimum tillage cropping systems due to the fact that tillage disrupts root proliferation (17). There are few published studies of the effects of repeated annual herbicide treatment on Canada thistle patches in notill and how the perimeter and density of the patches change over time (12). Weed species distribution and density maps may also be used for making decisions on herbicide applications in individual fields. Characterizing spatial distribution of weeds within fields could lead to better herbicide decisions and reduced herbicide use (19). Excess or inappropriate herbicide use reduces profit and may cause unnecessary environmental or health risks as well as unnecessary pressure for the development of herbicide resistant weeds (20). Eco- nomic threshold weed densities can be calculated as the lower limit at which herbicide control is cost effective. Weed seedlings are not uniform, but spatially aggregated in the field (16). Aggregated populations are individual plants occurring in patches of varying density and size with few or no plants occurring between patches (16). Be- cause of the aggregate nature of Canada thistle populations, herbicide cost and environmental degradation may be reduced if herbicide is applied only where weeds are present or exceed some predetermined threshold. Crop yield loss from competition can be predicted from weed density estimates. These data in turn could be used to determine if herbicides are economically practical and to select the best herbicide application strategy when control is required. Crop Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 161 loss from weed competition depends on the composition of the weed population and also its spatial distribution since weeds compete with each other as well as with the crop. Theoretical aspects have been discussed extensively, while little attention has been given to practical aspects, such as determining weed densities (1). Weed scouting has relied on random sampling to calculate average population densities as the first step in determining control recommendations (3). Random sampling and estimating populations are appropriate if samples are independent and the variance is uniform throughout the field (3). However, where weeds are aggregated, densities measured at several random locations are not independent, but rather spatially related or autocorrelated (3). Samples obtained close to one another are more similar than samples taken a greater distance apart, meaning the variance between pairs of samples is not uniform. Methods of spatial statistics can be used to map weed distributions across landscapes and describe variables that are not uniformly dispersed across the landscape. Measured variables, including weed density, are assumed to have spatial relationship with one another. Geostatistics is one method of spatial statistics originally developed for spatial analysis of ore deposits in mining (5) and are now widely used in many earth sciences, including soil and weed science (7,16). Geostatistics allows spatial relationships of sampled values for a variable to be used for estimating a value of a nearby unsampled location in preparing contour maps (16). Contour maps of estimated points between measured variables are generated by a process called point kriging. Point kriging is a mathematical interpolation process which estimates the values of unknown points from the values of known sample points and their distance from one another. One method of geostatistics is remote sensing. Remote sensing has provided agricultural crop and soil condition information for over 50 years (14). Recently, aerial photography and global positioning system (GPS) technology have been integrated and shown to be useful tools to map weed infestations in cropped areas. The latitude and longitude data provided by the GPS can be entered into a geographic information system (GIS) to georeference weed problems. Everitt et al. (10,11) used airborne videography and GPS technology to detect weed and brush infestations on rangelands and entered the geo- references data into a GIS to map noxious plant populations over an extensive area. The merging of remote sensing and GIS technologies can be useful to assess the extent of weed infestations, develop management strategies, and evaluate control programs. Conventional-color or colorinfrared aerial photography can be used for digital image analysis and spectral plant identification (2). The cameras typically used produce high resolution digital images, 0.6 to 2 megapixels per image, using ChargeCoupled Device (CCD) photocells (2). These images are converted to video format and stored on a compact disk (cd) with up to 50 images per cd. These cameras can acquire up to 10 images per second, so the combination of speed and data storage makes the system flexible for field use. The objectives of this experiment were: (i) map Canada thistle shoot densities and locations of ten established Canada thistle patches; (ii) compare the 162 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) location and density of each Canada thistle patch with contour maps generated in the spring of 1996 using extensive coarse grid sampling data, and to maps generated from aerial photography; (iii) determine how accurately this particular sampling method estimates the size and density of the selected thistle patches.

OUTLINE OF RESEARCH METHODS

Ten Canada thistle patches were randomly selected in midSeptember, 1996, from a 65 hectare notill soybean field in Moody county, South Dakota. A center location for each patch was visually estimated and marked with a flag. Shoot density was determined in the center of each patch by counting the shoots present in a 20 by 50 cm (0.1 m2) rectangle. Shoot densities were recorded in 3.04 meter increments in 8 directions from the center of each patch. The outside boundary of the thistle patches on each axis was arbitrarily defined as having 2 consecutive axis measurements of 0 shoots per 0.1 m2. Latitude and longitude coordinates of the center location of each patch were determined using a differentially corrected GPS(Chervils MicroComputer Systems, Chervils, OR). GPS coordinates were converted into meters with the following equation: COS(minimum latitudeº/(180)*3.14159))*minimum longi- tudeº(96+37/60+((point longitudeº960000)3700)/3600))*60* 1855 (15). The location of each patch in latitude and longitude degrees was converted to coordinate values in meters using the latitude and longitude of the southwest corner of the field as the origin. Spatial dependence of weed populations was characterized by semivariograms (Fig. 1), calculated using Geoeas, a geostatistical software program. Spherical models were fitted to calculated semivariance values. Semivariograms plot lag (sampling) distances between sample pairs against semivariance statis- tic for samples at each lag distance. The computing formula for determining semivariance is:

Y(h) is the semivariance of the variable Z(xi) with a lag distance of h. N(h) is the number of pairs of points within the lag distance [h+∆(h)]. For ecological data, such as weed densities, the semivariance is expected to increase as the lag distance increases out to a distance where spatial dependence ceases to be detectable (9). Features of semivariograms that are important for interpretation of spatial data are shown in Figure 1. The nugget, (Co) is the distance on the y-axis from the origin to the y-intercept and is the variability due to experimental error and other random factors. The range is the lag distance at which samples are independent of one another. The value of the semivariance at this point is the sill and is equal to the combination of the nugget and random variability. The percentage of variability (C) explained by spatial dependence is estimated as % variability = C/(Co+C), where Co is the nugget erect and C is the variability due to spatial dependence as described by the distance from the nugget to the sill along the yaxis. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 163

Figure 1. Generalized semivariogam

Locations of the ten intensively sampled Canada thistle patches in relation to the entire ﬁeld which was sampled in a 15 by 30 m grid in the spring of 1996 are shown in Figure 2. A 17 ha portion containing the intensively sampled Canada thistle patches lies inside the rectangle (Fig. 2). The 17 ha fraction of the ﬁeld containing the 10 mapped thistle patches was enlarged and overlaid on previous coarse grid sampled data (Fig. 3). A contour map of the Cana- da thistle patches overlaid on a point map of individual sample points is shown in Figure 4. The area of each patch was also determined by estimating the perimeter. The patch sizes ranged from 90 to 1600 m2. The total area of the ten patches was approximately 7035 m2 (0.7 ha). Point maps were created from sampling point coordinates and transact sampling of the ten thistle patches.

The point and contour maps were generated using Surfer® (13), a contouring and 3-dimensional graphics software program which analyzes geostatistical data. The maps were converted either to feet or meters in relation to the southwest corner of the field. The minimum latitude-longitude coordinate is in the southwest of the field and was converted to 0-latitude and 0-longitude. Aerial imagery was acquired on October 1, 1996 using a Spectraview digital camera carried on a Piper PA34 equipped with a 15 cm belly hole. Imagery was acquired at a 486 m altitude. A near infrared (782-968 nm) spectral channel was used to determine the weeds present. The raw data for one of the photographs is shown in Figure 5. The image was acquired under sunny conditions around 2:30 p.m. solar time. The GPS latitude-longitude coordinates obtained from digital imagery were integrated with GIS technology to georeference weed populations on a fieldwide basis. The field lies in the center of the aerial image (Fig. 5). The road that borders the south and west edges of the field is light gray. The upper righthand corner of the field in this photo was used as the 0,0 coordinate as the photograph was superimposed on the topography map (Fig. 6.) The ten intensively sampled Canada thistle patches are displayed in the small squares. 164 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 2. 1300 grid sampled points and 10 intensively mapped Canada thistle patches.

Figure 3. Intensively sampled Canada thistle patches overlaid on 17 ha grid sampling. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 165

Figure 4. Contour map of 10 sampled Canada thistle patches overlaid on individual sampling points.

Figure 5. Near infrared photograph superimposed on a topography map. 166 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 6. Near infrared photograph superimposed on a topography map.

The latitude-longitude coordinates of the four corners of the ﬁeld were previously determined with a GPS receiver. The data in the aerial image was reduced by sorting the latitude-longitude points from the aerial image and eliminating the points outside the ﬁeld perimeter. Each of over 60,000 points in the digital image was designated a number from 0 to 256 based on re-

ﬂectance. Colors (red, yellow, etc.) were assigned to these points in Surfer®.

RESULTS AND DISCUSSION

Semivariogram functions relate the variance of each variable to the sampling distance intervals between the pairs of values at increasing distances from one another at sampling points (7). Graphs of semivariance verses lag distance are usually used to establish whether known variables have spatial dependence and at what lag distance the values become independent of one another (7). Variogram functions indicated that four of the ten patches exhibited strong spatial dependence at lag distances less than 20 meters. These lag distances are the "sills" of the variograms. The sill is the plateau value of the variogram function and is the lag distance between measurements at which one value is not spatially related to neighboring values (7). Lag distances beyond the sill value are not spatially related. Variogram functions showed that 6 of the 10 patches did not display spatial dependence at the lag distances modeled. This suggests that the samples were collected too far apart or that a different sampling method should be used to determine if spatial dependence is present. In the spring of 1996 the entire field was intensively sampled on a 15 by 30 m grid. A point map of Canada thistle density across the entire field according to coarse grid sampling is shown in figure 2. The large black dots point out where the ten random patches were sampled in the fall of 1996. It appears that only 6 of the 10 patches mapped in the fall overlay on a patch mapped Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 167 the previous spring because many of the patches fall between the 1336 grid sampled points. This indicated that to accurately map weed species present across an entire field, an adaptive sampling method should be implemented to avoid weed patches present between grid sample points. This adaptive process should sample only areas of the field where weeds exist to maximize sampling efficiency. Generally, the highest densities of Canada thistle shoots were in the centers of the patches (Figs. 3 & 4). Shoot density within the patches declined as the distance from the center of the patch increased. Many of the patches appeared to be larger in the north-south direction than in the east-west direction possibly due to the notill cropping practices. The field is combined in a north- south direction which distributed weed seeds in a similar pattern. Figures 5 and 6 show aerial digital images of weed infestations in the 65 ha field in the raw and topographically superimposed forms, respectively. The aerial image was acquired on October 1, 1996, which was about 1 week prior to the first killing frost. Most of the leaves on the soybean plants were brown and thus not transpiring or photosynthesizing. The near infrared wavelengths in the digital camera detect transpiration or water loss by plants. Since the soybeans had reached physiological maturity they were not carrying on transpiration and appeared a lighter color in the image. The weeds were still green and transpiring, showing up as a darker color on the images. The weeds have a conspicuous brown-red color that can easily be distinguished from soybean plants and bare soil. Integration of the latitudelongitude coordinates with the digital image was useful to georeference weed populations over the entire field. Geostatistics can be used to map spatial distribution of weeds in a field. Variograms provide the ability to test whether or not the variables exhibit spatial dependence. In this type of research, sample spacing is important and may not have been optimum for sampling the Canada thistle shoots in this experiment. Better kriged maps may have been developed if the patches had been intensively grid sampled instead of sampled in transects. By sampling in transects, the center of the patches are mapped very accurately because the transects come together at the centers of the patches. However, as the sampling points moved away from the center, the sampling points on the axis become farther away from each other which leaves more unsampled area between the transects. By grid sampling the entire patch there is less unsampled area between each sampling point. In this particular field, geostatistics can be used to analyze how weed control treatments change weed distribution and densities across landscapes over time. Site-specific weed management can improve our knowledge of the factors influencing the location and densities of weed patches across a field. An understanding of these factors will aid in selecting appropriate weed management strategies that decrease the presence of weeds. These management strategies may lead to reductions in herbicide use and increased profitability as well as reduce negative environmental impacts. 168 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

ACKNOWLEDGMENTS

This research was partially funded by USDA-CSREER Grant 94-3414-1136, North Central IPM Grant number 97-34103-4789, UMAC-a NASA funded multi- state consortium, South Dakota Soybean Research and Promotion Board, and South Dakota Corn Utilization Council.

LITERATURE CITATIONS

1. Brain, P. 1990. The Effect of weed distribution on predictions of yield loss. J. Appl. Ecol. 27:73 5742. 2. Brown, R. B., J.P.G.A. Steckler, and G.W. Anderson. 1994. Remote sensing for identification of weeds in notill corn. Trans. ASAE 37 (I) 297302. 3. Cardina, J., D.H. Sparrow, and E.L. McCoy. 1995. Analysis of spatial distribution of common lambsquarters(Chenopodium album) in notill soybean(Glycine max). Weed Sci. 43 :258268. 4. Carlson, S.J. and W.W. Donald. 1988. Glyphosate effects on Canada thistle(Cirsium arvense) roots, root buds, and shoots. Weed Res. 28:3745. 5. Clark, I. 1979. Practical geostatistics. Applied Science Publ., Ltd., London. 6. Doll, J. D. 1984. Controlling Canada Thistle. North Cent Reg Ext Publ. Coop Ext Serv, Michigan State University. 3p. 7. Donald, W.W. 1994. Geostatistics for mapping weeds, with a Canada thistle(Cirsium arvense) patch as a case study. Weed Sci. 42:648657. 8. Donald, W. W. and T. Prato. 1992. Efficacy and economics of herbicides for Canada thistle (Cirsium arvense) control in notill spring wheat(Triticum aestivum). Weed Sci. 40:233240. 9. Ellsbury, M.M., W.D. Woodson, S.A. Clay, and C.G. Carlson. Spatial characterization of corn rootworm populations in continuous and rotated corn. Pages 487494 In Robert, P.C., R.H. Rust, and WE. Larson (eds.) Presision Agriculture. American Society of Agronomy, Madison, WI. 10. Everitt, J.H., D.E. Escobar, M.A. Alaniz, R. Villarreal, and MR. Davis. 1993. Integration of airborne video, global positioning system, and geographic information system technologies for detecting and mapping two woody legumes on rangelands. Weed Technol. 7:981987. 11. Everitt, J.H., J.V. Richardson, D.E. Escobar, M.A. Alaniz, R. Villarreal, and MR. Davis. 1994. Light reflectance characteristics and remote sensing of Bid Bend loco (Astragalus mollissimus var. earlei) and Wooton loco (Astragalus wootonii). Weed Sci. 42:115122. 12. Glenn, S. and L.K. Heimer. 1994. Canada thistle(Cirsium arvense) control in notillage corn(Zea mays). Weed Technol. 8:134138. 13. Golden Software, Inc; Golden, CO, version 6. 14. Hansen, L.D., P.C. Robert, and M. Baur. 1995. Mapping wild oat infestations using digital imagery for sitespecific management. Pages 495503 In A.F. Wiese, ed. Weed Control In Limited Tillage Systems. Weed Science Society of America, Champaign, IL. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 169

15. Kennedy, Michael. 1996. The Global Positioning System and GIS: An Intro- duction. Ann Arbor Press, Inc. Chelsea, Michigan. 16. Mortensen, D.A., G.A. Johnson, D.Y. Wyse, and A.R. Martin. 1995. Manag- ing spatially variable weed populations. Pages 397415 In Robert, P.C., RE Rust, and W.E. Larson, eds. SiteSpeciﬁc Management for Agricultural Sys- tems. American Society of Agronomy, Madison, WI. 17. Staniforth, D.W. and A.F. Wiese. 1985. Weed biology and its relationship to weed control in limitedtillage systems. Pages 1525 In A.F. Wiese, ed. Weed Control In Limited Tillage Systems. Weed Science Society of America, Champaign, IL. 18. Trangmar, B.B., R.S. Yost, and G. Uhara. 1985. Application of geostatistics to spatial studies of soil properties. Adv. Agron. 38:4593. 19. Wiles, L.J., G.W. Oliver, A.C. York, H.J. Gold, and G.G. Wilkerson. 1992. Spatial distribution of broadleaf weeds in North Carolina soybean(Glycine max) ﬁelds. Weed Sci. 40:554557. 20. Wiles, L.J., G.G. Wilkerson, H.J. Gold, and H.D. Coble. 1992. Modeling Weed distribution for improved postemergence control decisions. Weed Sci. 40:546553.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 171

LEAFY SPURGE—A REVIEW

Sharon A. Clay and Chad M. Scholes Plant Science Department South Dakota State University Brookings, South Dakota 57007

ABSTRACT

Leafy spurge (Euphorbia esula L.) is a perennial herbaceous weed that infests millions of acres of range and pasture in the northern Great Plains. It out- competes grasses and lowers land productivity because cattle will not graze infested areas even if spurge makes up only 10% of the vegetative biomass. This presentation will cover the history, taxonomy, and phenology of leafy spurge. A discussion of chemical, mechanical, and biocontrol techniques that aid in leafy spurge management will also be included.

REVIEW

Leafy spurge (Euphorbia esula L.) is a highly variable plant or plant group that is found in several countries including Hungary, Austria, Italy, and Russia. It has been introduced several times and at different places in the United States and Canada (Dunn, 1985). Dunn (1985) speculated that leafy spurge was introduced ﬁrst in ship ballast with the earliest specimens collected in 1877 near Philadelphia. The North American introductions were probably from several separate sites in Europe. Probable sources of Midwestern introductions are im- migrant Mennonite seed, Russian seed grain shipments, and brome grass seed (Dunn, 1985). Thirty of the 48 contiguous United States are infested with leafy spurge (Dunn, 1979). The states where leafy spurge is considered a problem include: Colorado, Idaho, Minnesota, Montana, Nebraska, North Dakota, Oregon, South Dakota, Wisconsin, and Wyoming (Watson, 1985). In 1979, the total area of leafy spurge infestation in the United States and Canada was estimated to be 1,000,000 ha (Noble et al., 1979). A 1989 South Dakota Department of Agri- culture survey estimated 61,000 ha (151,000 acres) in South Dakota were infested with leafy spurge, a 9% increase from 1988 (Clarke, personal communication) , and a 250% increase from the 1979 estimate of 24,000 ha (60,000 acres) (Noble et al., 1979).

TAXONOMY

Croizat (1945) and Radcliffe-Smith (1985), among others, have discussed the taxonomic problem of classifying the E. esula complex. To date, a deﬁni- tive relationship has not been established between the North American leafy spurge and the Eurasian species. Ebke and McCarty (1983) did a nursery study and concluded that the majority of leafy spurge plants are actually E. x pseu- 172 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) dovirgata, a hybrid between E. esula and E. virgata. The growth of North American leafy spurge was highly variable with some populations having more E. esula characteristics, some having more E. virgata characteristics, but most populations possessed characteristics of both species. Stahevitch et al. (1988) examined leafy spurge populations for breeding barriers to determine if North American leafy spurge comprised a single breeding population. All E. esula samples were hexaploid, accessions from Europe and different regions in North America were successfully crossed, and pollen stainability was 100% for almost all samples. These results support the concept of a single, polymorphic species. A number of chemotaxonomic studies also have been done to resolve the leafy spurge taxonomic problems (Manners and Davis, 1984; Harvey et al., 1988; Stahevitch et al., 1988; and Valcarce et al., 1989). The results of these studies have either concurred that leafy spurge is one variable species or been inconclusive. Manners and Davis (1987) discovered differences in leafy spurge accessions using the distribution of di- and triterpenoids occurring in the latex of Euphorbia. They suggested that jatrophane diterpenes, secondary metabo- lites possibly unique to Euphorbia spp., have considerable taxonomic importance. Holden and Mahlberg (1992) studied qualitative and quantitative triter- penoid content in leafy spurge latex by gas-liquid chromatography. Relation- ships between 39 North American and 37 European accessions were examined with the goal of identifying European populations of leafy spurge that supported natural enemies most likely to accept North American leafy spurge as a host. The results showed two things. First, European leafy spurge was much more variable than North American leafy spurge, indicating North American leafy spurge may have arisen from relatively few introductions. The other noteworthy result was that eight leafy spurge accessions from Montana, North Dakota, and South Dakota were most closely related to an accession from near Pest, Hungary. With the same goal in mind as Holden and Mahlberg (1992) but using a different approach, Nissen et al. (1992) analyzed restriction fragment length polymorphisms (RFLPs) of chloroplast DNA (cpDNA) from two North Ameri- can and three European leafy spurge accessions to measure genetic variation and relatedness. Results from this study showed that accessions from Russia and Montana were identical to each other and were only slightly different from a Nebraska accession. An Austrian accession was most divergent. These data suggest that biocontrol organisms taken from where the Russian leafy spurge accession was found may be preadapted to the Montana and Nebraska accessions.

PHENOLOGY

The biological significance of leafy spurge is in part due to the phenology of the plant. Leafy spurge shoots developing from adventitious shoot buds on the root crown and lateral roots emerge early in the spring (early to late April depending on latitude). Seeds may germinate throughout the growing season Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 173 if adequate moisture is present. However, germination most commonly occurs in the months of April and May (Selleck et al. 1962). Initiation of the inflorescence occurs within 5-7 days of the shoot's emergence (Selleck et al., 1962) although flowering begins approximately one month after shoot emergence. Yellow bracts that subtend the inflorescence are highly visible during flowering in late May and June. Flowering at the terminal inflorescence lasts about one month but additional branches may continue to flower throughout the summer if conditions are favorable (Messersmith et al., 1985). Seed formation and maturation continues for about 30 days after the last flower appears (Best et al., 1980). Seeds are dispersed as capsules dry and dehisce. Senescence in the fall is characterized by the leaves and stem turning red or yellow with most leaves senescing before a killing frost (Messersmith et al., 1985).

PLANT BODY

A mature leafy spurge stem is herbaceous, but woody in texture (Bakke, 1936), and may grow to a height of 90 cm (Selleck et al., 1962). In cross section a majority of the stem is comprised of woody, secondary xylem with a relatively insignificant amount of vascular tissue that surrounds a central pith. To the outside of the vascular cambium is a tight layer of phloem, latex vessels, and periderm. The outer layer of the periderm, along with a heavy cuticle, protect the stem from water loss (Bakke, 1936). The waxy, linear leaves of leafy spurge have an alternate arrangement except for the first two pairs of leaves of a seedling, which are opposite (Raju, 1985). Leafy spurge leaves are adapted for xeric conditions with sunken stom- ata on the upper and lower leaf surfaces (Bakke, 1936) and a heavy crystal-like waxy cuticle to prevent water loss (Galitz and Davis, 1983). The vast, spreading root system of leafy spurge is often cited as the key to the "weediness" of leafy spurge (Best et al., 1980; Derscheid et al., 1985; Raju et al., 1963). Selleck et al. (1962) calculated the mass of leafy spurge roots in 0.4 ha, 1.2 m deep to be 3823 kg of plant material. Raju et al. (1963) described the morphology of the leafy spurge root system as heterorhizic, which is the presence of a dimorphic pattern of root growth. Long and short roots are present, although both types may be difficult to recognize in mature plants because the short root growth is determinate (Raju, 1985). The degree of terminal growth and branching, and the presence of secondary growth distinguish- es long from short roots. Only the long roots are important in vegetative propagation through bud formation (Raju et al., 1963). A cross-section of a young leafy spurge root would reveal a typical arrangement of tissue with a central stele surrounded by cortex and an epidermis (Bakshi and Coupland, 1959). As secondary xylem and phloem are produced, the pericycle becomes meristematic and produces a phellogen layer. Cork is produced on the periphery of the phellogen and phelloderm is produced inside this layer. With the secondary growth and production of cork, the outer cortex and epidermis die and slough off (Myers et al., 1964). Mature leafy spurge roots are protected from dry conditions by the cork (Selleck et al., 1962) that varies from 2 to 7 cells in thickness (Bakshi and Coupland, 1959). 174 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Secondary growth produces parenchyma that serves as storage tissue. Laticif- erous tissue is abundant throughout the secondary cortex of mature roots (Ra- ju, 1985). Leafy spurge plants produce axillary and adventitious shoot buds (Raju, 1985). Axillary buds increase branching of aboveground shoots. Adventitious buds found on the roots can be classiﬁed as reparative or additional buds (Ra- ju et al., 1963). Reparative buds are produced after an injury while additional buds are produced spontaneously. Bakke (1936) found buds on roots at a depth of 2 m. The number of buds present per unit length of root is greatest just beneath the soil surface and decreases with increasing depth (Coupland and Alex, 1955). Buds counted by depth averaged almost 3 per centimeter in the top 2.5 cm layer of soil. This gradually decreased to approximately 1 bud per 2 cm at the 30 cm layer. The number of buds per unit length of root was found to be positively correlated with the weight of root tissue. With respect to the soil proﬁle, 62% of the buds were found within the A-horizon (Coup- land and Alex, 1955). The ability of an adventitious bud to produce a new shoot is not related to size or placement on the root system. Root bud dormancy control is correlative (enforced by main shoot) rather than innate (Nissen and Foley, 1987). Buds present on the root crown are less susceptible to freezing than buds from root sections (Schimming and Messersmith, 1988). Root crown buds could be expected to endure colder temperatures during the winter than buds on root sections deeper in the soil. Therefore, the ability of leafy spurge buds to resist freezing appears to be correlated with the probability of experiencing the coldest temperatures during winter.

FLOWERING; SEED PRODUCTION AND DISPERSAL

Leafy spurge produces flowers in terminal umbels at the stem apex and on axillary branches (Raju, 1985). The cyathium, a type of flower unique to the genus Euphorbia, possesses a single pistil with a compound ovary which is surrounded by 12 to 20 staminate flowers. The pistillate and staminate flowers are surrounded by a cup-shaped involucre. The cyathium is enclosed by five yellow involucral bracts and four nectar glands (Messersmith et al., 1985). Wind may cause pollination through incidental contact of pistillate and staminate flowers but not through the transport of pollen (Selleck et al., 1962). The presence of nectaries attracts insects to leafy spurge flowers. A survey in Saskatchewan listed 60 insect species associated with leafy spurge during late June and early July of 1955 (Best et al., 1980). Insect feeding on nectar and pollen causes most leafy spurge pollination (Messersmith et al., 1985). The leafy spurge capsule develops from a superior, 3-celled ovary possessing a central placenta. The capsule, as described by Bakke (1936), has walls with three layers of tissue. The inner layer of contractile tissue is comprised of 3 layers of columnar cells arranged longitudinally. The middle layer is made up of a single layer of parenchymatous cells arranged in a radial fash- ion. Cells in the outer layer are arranged horizontally toward the lodicular su- ture. The mature capsule dehisces explosively, projecting seeds up to 4.5 m Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 175 when dried by high temperatures and low humidity (Bakke, 1936). A maximum of 426 seeds/shoot resulting in a seed yield of 3800 kg/ha has been observed (Selleck et al., 1962). Animals may serve as agents of leafy spurge seed dispersal. Bakke (1936) claimed that mourning doves (Zenaida macroura L.) fed almost exclusively on leafy spurge seed if abundant. Mourning doves have the potential to be an efficient seed dispersal mechanism if intact seeds pass through the digestive tract. Bakke (1936) examined 14 doves and found that no seed from the digestive tract was viable, although 17% of seed from the bird's crop was viable. Block- stein et al. (1987) also found no intact leafy spurge seed in the digestive tract of seven wild mourning doves. They examined fecal material from nests and found viable leafy spurge seed in some nests. Therefore, the primary way leafy spurge seed may be dispersed by mourning doves is by mature doves carrying seed to nestlings, which may not be able to digest the seed as effectively as mature birds. Ants also have been suggested to be leafy spurge seed dispersal agents. The fleshy caruncle of the leafy spurge seed has led to the suggestion that leafy spurge seed is myrmecochorous (Selleck et al., 1962). Myrmecochory is a type of mutualism between ants and plant species with seeds bearing caruncles or elaiosomes. In this relationship, the ants feed on the elaiosomes and discard the viable seed about the nest. The possible advantages of this relationship to the plant includes: escape of seed predation, competition avoidance, the availability of favorable substrate for germination, and escape from fire (Pemberton, 1988). While Selleck et al. (1962) found no suggestion of myrmecochory in two genera of ants (Formica sp. and Lasius sp.), Pemberton (1988) reported a myrmecochorous relationship between leafy spurge and Formica obscuripes. Leafy spurge seed may remain dormant for 5-8 years (Messersmith et al., 1985) with viability decreasing each year (Selleck et al., 1962). Temperature appears to be the most important factor influencing leafy spurge seed germination. Early spring germination following several days of 26-28º C maximum air temperature (Selleck et al., 1962) is the norm, with germination through mid-October dependent upon plentiful moisture (Selleck et al., 1962; Best et al., 1980; and Messersmith et al., 1985). Seedling emergence is possible from a depth of 15 cm despite the small seed size (Selleck et al., 1962). In competition with other plants, leafy spurge seedlings undergo severe mortality and rarely produce more than one shoot (Selleck et al., 1962) which almost never produces an inflorescence the first year (Messersmith et al., 1985). However, the seedlings do produce crown buds at the six-leaf stage and these buds are capable of regrowth. Nine of 15 six-leaf seedlings with stems cut off above the buds produced new shoots (Selleck et al., 1962).

LATEX

Leafy spurge produces a latex which is exuded from the plant upon injury. The latex is a white ﬂuid containing a serum with a mixture of compounds in solution or suspension (Lynn and Clevette-Radford, 1987). Some of the common components of latex include enzymes, carbohydrates, lipids, free amino acids, vitamins, and terpenes (Lynn and Clevette-Radford, 1987). 176 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

The latex of leafy spurge is produced and stored in non-articulated laticifer cells which are present throughout the plant. Laticifer initials arise in the meristematic mass of the developing embryo. Each initial is capable of unlimited growth (Mahlberg and Sabharwal, 1968). The laticifer system has been suggested to serve as a nutritional reservoir, conductive system, water balance reg- ulator, defense system from herbivores, and as a storage system for cellular byproducts. The presence of terpenes and alkaloids supports the idea that laticifers function as a storage and/or excretory system for secondary plant products (Biesboer and Mahlberg, 1978). Starch grains present in the laticifers are not utilized as an energy reserve but rather in wound closure (Biesboer and Mahlberg, 1981). The latex of the genus Euphorbia may serve as a physical and chemical defensive mechanism. Chewing and piercing-sucking mouthparts of insects become clogged with latex when feeding on leafy spurge (Best et al., 1980). Diterpenes and triterpenes isolated from the latex are potential anti-herbivore compounds. The diterpenes, ingenol, 12-deoxy-phorbol, and their ester derivatives, are found in latex of many plants in the genus Euphorbia (Evans and Kinghorn, 1977). Latex containing ingenol and 12-deoxy-phorbol has been used as an insecticide, ﬁsh poison, arrow poison, and treatment for sca- bies and ringworm (Evans and Kinghorn, 1975). Therefore, the latex may dis- courage herbivores, except for those which have evolved adaptations for feeding on this plant (Best et al., 1980).

ROOT RESERVES

Leafy spurge and other perennial weeds store reserve carbohydrates and nitrogenous compounds in their roots. Carbohydrates are the major energy storage molecules in leafy spurge (Cyr and Bewley, 1990) and other perennial plants (Arny, 1932). In temperate species, excess photoassimilate is stored as nonstructural polysaccharides in underground storage tissue during summer and fall. These polysaccharides are hydrolyzed to soluble sugars while overwintering and used for maintenance respiration and as an energy source to fu- el growth in the spring before leaves are capable of photosynthesis (Cyr and Bewley, 1989). Although carbohydrates are quantitatively the most important form of energy reserve, nitrogenous compounds may be as important qualita- tively (Cyr and Bewley, 1989). Amino acids and soluble protein are cycled similar to sucrose and may be a critical storage currency (Cyr and Bewley, 1990). Total nonstructural carbohydrate (TNC) is a measure of carbohydrate available for metabolism or translocation. Total nonstructural carbohydrate is made up of a soluble fraction, predominantly sucrose, and an insoluble fraction, starch (Lym and Messersmith, 1987). The seasonal pattern of root TNC content is variable. Lym and Messersmith (1987) reported on leafy spurge root TNC averaged over four growing seasons. They found that TNC was lowest in April, increased to a maximum in July, August, and September and then declined through October. This is in disagreement with what Arny (1932) found. Arny described the total readily available carbohydrate content of leafy spurge roots as being fairly high in April, decreasing in May, gradually increasing Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 177 through October, and then declining in November. In studies conducted at SD- SU, TNC in both shallow roots (0 to 15-cm depth) and root crown followed the pattern described by Lym and Messersmith (Scholes, 1996). Lym and Messersmith (1987) found a high correlation between TNC and temperature, but no correlation with precipitation patterns. Leafy spurge utilized soluble sugars when stressed by high temperatures and replaced those sugars during periods of cool temperatures. Mowing and smother treatments, lowered TNC concentrations in both crown and root tissues (Scholes, 1996) with crown tissue more affected by treatment. Variations in root reserves of other perennial weeds such as Canada thistle (Cirsium arvense [L.] Scop.), perennial sowthistle (Sonchus arvensis [L.]), quackgrass (Elytrigia repens [L.] Nevski), and ﬁeld bindweed (Convolvulus arvensis [L.]) have been studied. Root reserve patterns for these species are dis- similar. For example, Canada thistle root carbohydrate reserves decline from May through June or July and then increase through September (Tworkoski, 1992). The decline in root reserves appeared to be related to growth early in the season but in early fall photoassimilate movement in the phloem was basipetal, increasing root reserve storage. Perennial sowthistle root carbohydrate levels follow the same general seasonal pattern as Canada thistle, with reserves lowest during ﬂowering (Arny, 1932). Quackgrass TNC was lower than that of Canada thistle, perennial sowthistle, or leafy spurge, particularly during October and November (Arny, 1932). Quackgrass carbohydrate levels remained fairly constant between late April and early November. Cyr and Bewley (1989) found seasonal variation in free amino acids, soluble protein, and nitrate plus nitrite levels in leafy spurge roots. Free amino acids, soluble protein, and nitrate + nitrite were lowest in August and Septem- ber but increased rapidly by October because of remobilization during senescence. Abundance of individual amino acids changed throughout the year, indicating different roles in the nitrogen budget of leafy spurge. Weekly treatments of defoliation and decapitation of leafy spurge that began in August decreased soluble protein and free amino acids in leafy spurge roots by Novem- ber and nitrate by December (Cyr and Bewley, 1990). These decreases may have been caused by remobilization of nitrogen from the root to meristem important for regeneration, i.e. adventitious root buds.

CULTURAL CONTROL

Cultural control of weeds is, strictly, the use of control practices on cultivated land. Derscheid et al. (1985) expanded this deﬁnition to include all non- chemical methods including grazing livestock. Sheep and goat grazing will be considered as biological control here. Intensive cultivation can be effective in eliminating leafy spurge stands. However, much of the land infested with leafy spurge is pasture or rangeland that cannot be cultivated. The use of competitive crops is another cultural control method but few crops can outcompete leafy spurge without the complementary use of cultivation or herbicides (Der- scheid et al., 1960). Plastic smother and mowing are other potential cultural control methods. 178 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Frazier (1943) investigated the effects of 7-day and 14-day intervals of cultivation on field bindweed root carbohydrate reserves. The longer cultivation interval depleted carbohydrate reserves more than the 7-day interval. The 14- day cultivation interval allowed new field bindweed shoots to emerge about 6 days after cultivation, and these shoots continued to draw on carbohydrate reserves for 8 days after emergence. The shoot emergence decreased root reserves without allowing replacement of photoassimilates. Covering the soil surface with clear or black polyethylene for control of pathogens, nematodes, and weeds has been used as an alternative to chemical control methods. The method has been termed solarization or plastic mulch but will be referred to as plastic smother here. Plastic smother captures radiant heat energy from the sun that alters the microclimate in the soil rhizosphere (Al-Masoom et al., 1993), and causes soil temperatures to increase and nitrate leaching to decrease (Clarkson, 1960). Plastic smother can reduce weed populations by decreasing the number of viable seeds or by killing seedlings and mature plants. Weed seed viability can be reduced by modifying soil temperature, moisture, carbon dioxide, oxygen, and ethylene with plastic smother (Egley, 1990). Moist heated soil is much more detrimental to seeds than dry heated soil. Egley (1990) found that dry soil maintained at 60º C for 7 days only slightly decreased seed survival. How- ever, moist soil maintained at 60º C for 7 days increased seed mortality from 70 to 100%. Moist soil maintained at an elevated temperature can break mechanical (seed coat) seed dormancy. Those seeds not succumbing to the high temperature will be induced to germinate with the seedling likely to perish in the hot soil (Egley, 1990). Plastic smother or solarization has been used most often in areas where the growing season is preceded by a hot, dry period such as in California, Israel (Horowitz et al., 1983), India (Kumar et al., 1993), Greece (Vizantinopoulos and Katranis, 1993), Jordan (Abu-Irmaileh, 1991), and the United Arab Emirates (Al-Masoom et al., 1993). In these areas, plastic smother is typically used as a pre-plant treatment. Duration of the plastic smother treatment varies between 1 week to 2 months. Horowitz et al. (1983) studied duration of plastic smother and its effects on weed emergence after removal of the plastic. A treatment period of 2 weeks controlled most weed species for at least 1 year. A 6 week plastic smother treatment gave 100% control of common purslane 3 months after treatment (MAT), henbit 4 MAT, and field bindweed 2 MAT. In contrast to the use of plastic smother as a pre-plant soil sterilization technique in agriculture, horticulturists use plastic as a mulch to smother weeds during crop growth or as a long-term weed suppression measure in landscaped areas (Martin et al., 1991). Crops such as tomato, basil, parsley, and rosemary can be grown through holes in the plastic that suppresses most weed growth. Additional advantages to this technique include an increase in soil temperature, conservation of moisture, reduction of nutrient leaching, reduction of soil com- paction, and increased CO2 levels (Ricotta and Masiunas, 1991). Because elevated soil temperatures under plastic quickly decrease with depth (Horowitz et al., 1983), roots of crops grown through plastic can grow in warm soil without drying. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 179

Black plastic reduced both leafy spurge stem density and biomass and flowering was prevented (Clay and Scholes, 1997). Two months after application, biomass was reduced by 66%. After 2 continuous years of treatment, virtually no leafy spurge was present (Clay and Scholes, 1997). The problem with this treatment is that all other vegetation under the plastic died also. When treatment was removed after one season, leafy spurge reinfested the areas rapidly, with biomass equal to control areas by early June of the second year and grass biomass less than 200% of the control (Clay and Scholes, 1997). Ar- eas that are treated in this manner must be reseeded with competitive grass species such as "Bozoisky" Russian wild rye (Elymus junceusi Fisch.) or “Ro- dan” western wheatgrass (Agropyron smithii Rydb.) (Christianson and Lym, 1984) in order to supress reinfestation. Mowing is a method used to prevent seed production or lower root reserves. By mowing at 14 day intervals the spurge plant uses stored carbohydrates to regenerate stems and leaves but does not regenerate the root reserves through photosynthesis (Derscheid et al., 1985). Removing the stem will release the root and shoot adventitious buds from apical dormancy. This will stimulate growth of lateral buds, possibly increasing stem density. Peters and Lowance (1978) mowed western ironweed (Vernonia baldwinii) and gray goldenrod (Solidago nemoralis) multiple times as a means of control. Both plants are perennials with the ability to produce shoots from adventitious root buds. Initially the mowing treatment caused an increase in western ironweed density because of its ability to produce shoots from adventitious root buds. However, western ironweed shoot density was reduced 81% after 2 years of mowing 3 times per year. Gray goldenrod stem density also was reduced by repeated mowing (Pe- ters and Lowance 1978). Gray goldenrod stand density was reduced up to 85% after one year of 2 mowing treatments per year and up to 97% after two years of repeated mowing. Mowing at two and four week intervals reduced leafy spurge biomass by > 95% after the first treatment (Clay and Scholes, 1997). After two years of treatment, virtually no leafy spurge was left in the mowed areas and plants that remained did not flower. When treatment was removed after the first year, leafy spurge recovered, however, grass biomass was 50 to 100% more than that of the control.

CHEMICAL CONTROL

The use of herbicides to suppress leafy spurge is currently the most widely used control method (Alley and Messersmith, 1985). Bakke (1937) discussed the application of sodium chlorate, sodium chlorate plus glue and sulfuric acid, sodium chlorate plus calcium chloride (Atlacide), creosote-kerosene, sulfuric acid, ammonium thiocyanate, and potassium chlorate on leafy spurge. The sodium chlorate plus glue and sulfuric acid was very effective (Bakke, 1937) but was expensive and an extreme fire hazard (Alley and Messersmith, 1985). The introduction of 2,4-D [2,4-dichlorophenoxy)acetic acid] in the 1940's and dicamba (3,6-dichloro-2-methoxybenzoic acid) and picloram (4-amino-3,5,6- 180 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) trichloro-2-pyridinecarboxylic acid) in the 1960's began the modern era of herbicide use. Today, picloram, dicamba, and 2,4-D are the most commonly used herbicides for leafy spurge control (Alley and Messersmith, 1985). Picloram at 2.24 kg ha-1 (2 lb acre-1) is the most consistently effective herbicide treatment for leafy spurge (Alley and Messersmith, 1985). However, in a study investigating the economics of leafy spurge control with herbicides, Gylling and Arnold (1985) found that the 2.24 kg ha-1 picloram rate was the most expensive of all treatments in the study, and the increased forage available with this treatment did not pay for the treatment. Lym and Messersmith (1985) concluded that an annual spring or fall application of 0.28 kg ha-1 of picloram in combination with 1.12 kg ha-1 of 2,4-D was the most economically effective treatment. Derscheid et al. (1960) experimented with several combinations of 2,4-D and brome grass to reduce leafy spurge stands. In combination with the brome grass, the herbicide treatments reduced leafy spurge stands up to 52% the first year and 81% the second year. Stand reduction was assessed in May of the year following treatment. Alley and Messersmith (1985) summarized numerous weed control guides, bulletins, etc. that suggested 1.12 to 2.24 kg ha-1 2,4-D will control the aboveground portion of the leafy spurge plant and seedlings, and usually prevent seed production for the year of application. A heavier rate applied at least twice a year for several years was considered necessary to reduce density of established stands. Bybee (1979) reported that 2,4-D at 1.12 kg ha-1 prevented seed production but had no effect on the leafy spurge stand the next year. Splitting the 2,4-D application into 10-0.11 kg ha-1 treatments applied every two weeks gave very good control of leafy spurge, much better than a single 1.12 kg ha-1 spring application (Clay and Scholes, 1997). Effects of the split treatment were not evident until after the fourth application but by the end of the season, only a trace of leafy spurge was present in the plots. Grass biomass was 60% greater in these plots than the check in the first year and was 100% greater in the second season of treatment. Plots recovering from one year of treatment had 100 to 200% more grass than an untreated check.

BIOLOGICAL CONTROL

Biological control of weeds is the use of natural enemies to reduce weed populations to levels below an economic threshold (Harris et al., 1985). Bio- logical control is based upon the assumption that introduced pests attain high population levels by escaping a “balance of nature existing in its native habitat” (Wilson and Huffaker, 1976). The aim of biological control is to establish a balance between the plant and its natural enemies introduced as biocontrol agents in the habitat the pest is infesting. Successful establishment of biological weed control agents may reduce herbicide use and the amount of energy used for chemical, mechanical, or physical weed control. In most cases, plants considered for biological control are perennials inhabiting relatively low value land (Andres et al., 1976) or ecologically sensitive areas such as riparian habitats or shelterbelts. In these areas, chemical or me- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 181 chanical control methods are often unusable, uneconomical, or physically impossible. While cattle avoid leafy spurge, goats and sheep are considered to be biological control agents of it. Johnston and Peake (1960) reported on the use of sheep to control leafy spurge more than 30 years ago. They found that five years of sheep grazing gave 98% control of leafy spurge. Since then, there have been conflicting reports on the attractiveness of leafy spurge to sheep. Landgraf et al. (1984) reported that sheep will consume up to 50% of their diet in leafy spurge. However, Kronberg and Walker (1993) reported that sheep did not consume leafy spurge because fermentation of leafy spurge in the ru- men either did not detoxify the aversive compound of leafy spurge or created a flavor aversion by producing a chemical irritant. In contrast, sheep exposed to goat digesta of leafy spurge did not develop a flavor aversion. Lambs can be trained to consume leafy spurge and introduction to the weed at an early age impacts later consumption (Walker et al., 1992). This early learning by lambs is important to being a selective grazer and, therefore, an effective biocontrol agent. While references to grazing leafy spurge with goats are scarce, it is not an uncommon practice. Goats appear to accept leafy spurge as a food source more willingly than sheep. This may be explained by the finding that aver- sions encountered by cattle (Kronberg et al., 1993) and sheep while or after di- gesting leafy spurge are not encountered by goats (Kronberg and Walker, 1993). Insect introductions have become a major weed biocontrol tactic in the United States. The insects may feed on the root system, foliage, vascular tissue, or flowers, all of which reduce the vigor or reproductive capacity of the plant. Insects that attack the plant where damage is most stressful to the plant are the most successful biocontrol agents. Introduced biocontrol agents must meet certain criteria (Andres et al., 1976). The most important criterion is that the insect be host specific. Poten- tial biocontrol insects are tested rigorously to ensure that they do not attack desirable plants that may be associated with the weed to be controlled (Andres et al., 1976). The bioagent must be adapted or adaptable to the environment that the introduced weed species is inhabiting. The introduced agent also must be able to locate the host plant and reproduce quickly enough to affect the given weed. Finally, the insect must be introduced free from its own parasites and diseases (Andres et al., 1976). Leafy spurge is a good candidate for biological control because it is a perennial plant that often inhabits low value land. Several species of insects associated with the leafy spurge complex in Europe and Asia have been introduced as biological control agents in the United States and Canada. The spurge hawkmoth (Hyles euphorbiae L.), a defoliator, was released in Canada and then in the United States as the first biological control attempt on leafy spurge in 1965 (Harris, 1984). The hawkmoth is very susceptible to predation by several species of insects (Forwood and McCarty, 1983) and, therefore, was difficult to establish. In addition, a virus associated with this species has prevented an increase in the population (Spencer, 1994). Hyles euphorbiae has been estab- 182 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) lished in Ontario and at one site in Montana, but the populations are not pro- lific enough to be of any value in controlling the leafy spurge (Harris, 1984). Two moths with root boring larvae, Chamaesphecia tenthrediniformis D. & S. and C. empiformis Esp., were released in the United States and Canada between 1970 and 1974. These moths never became established (Harris, 1984). Oberea erythrocephala Schrank is a beetle whose larvae mine leafy spurge root and stem tissue. It was released in Canada (Harris, 1984), Montana, Ore- gon, and Wyoming between 1980 and 1984 (Rees et al., 1986). This long- horned beetle is well established in Montana and seems to prefer areas where leafy spurge stem diameter is 2.5 mm or greater (Spencer, 1994). The population growth of this beetle is very slow and took about 10 years to increase to damaging levels. The leafy spurge gall fly [Spurgia esula (Bremi)] has been successfully established in the United States and Canada (Julien, 1987). The gall fly reduces seed production through the formation of galls on the terminal ends of shoots rather than reducing the stem density of the spurge. Gall fly establishment thus far has been limited to sheltered areas, specifically tree belts. Gall flies may complete up to five life cycles during a normal season (Plant Protection and Quarantine, 1989). The gall fly could be important in reducing seed production while other biocontrol agents are being established to reduce the density of leafy spurge. Current efforts are focused on Aphthona flea beetles, which have shown some signs of being successful biological control agents. Six species of flea beetles [A. abdominalis Duftschmid, A. cyparissiae (Koch), A. czwalinae Weise, A. flava Guill., and A. nigriscutis Foudr.] have been released as biological control agents for leafy spurge in the United States. Sommer and Maw (1982) reported that the apex and ventral groove of the aedeagus, two characteristics used to distinguish between Aphthona spp. are not reliable for positive identification. McDaniel et al. (1992) demonstrated that other aedeagus (male reproductive structure) characteristics could be used to reliably separate A. cyparissiae, A. flava, A. nigriscutis, and a native flea beetle Glyptina atriven- tris Horn. Aphthona cyparissiae was first released as a biocontrol agent in North America in 1982. It was originally collected from Austria, Hungary, and Switzerland. Aphthona czwalinae was first released in 1985, with original stock coming from Austria and Hungary. Aphthona flava was collected in Hungary and Italy and released in 1982. In 1983 Aphthona nigriscutis was released after being collected in Hungary (Julien, 1987). The five univoltine flea beetles (A. cyparissiae, A. czwalinae, A. flava, A. lacertosa, and A. nigriscutis) have similar life cycles. In Europe, males and females of these Aphthona beetles emerge simultaneously in June in an approximate 1:1 ratio. Oviposition begins one week after emergence and continues for up to 14 weeks. Under laboratory conditions at 20º C eggs hatched at 16.9 ± 1.8 days (A. cyparissiae) and 18.6 ± 1.4 days (A. flava). The larvae progress through 3 instars. Larvae feed into the fall and form a cell for overwintering. Aphthona cyparissiae consistently survived a temperature of -7.2º C and A. nigriscutis consistently survived down to -5.5º C. Two A. cyparissiae larvae Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 183 survived down to -13.0º C, indicating the ability to supercool. Under laboratory conditions, pharate pupal development took between 28 and 57 days. Pu- pal duration lasted another 20 days (Sommer and Maw, 1982). Univoltine flea beetles attack leafy spurge in two ways. The main impact is from larvae mining the roots, although adults feed on foliage. Root mining by the first instar larvae tends to be on the new filamentous roots, and as the larvae progress through the second and third instars they feed on older, perennial roots (Sommer and Maw, 1982). Root feeding by Aphthona spp. beetles reduces water and nutrient uptake by destroying the vascular tissue and provides an entrance for plant pathogens. The root feeding also may reduce carbohydrate reserves by isolating root fragments (Maw, 1981). The four species of Aphthona released to date are reported to have different habitat preferences. Aphthona nigriscutis prefers dry, coarse-soiled sites that occur on knolls or south-facing slopes. Aphthona cyparissiae prefers open, mesic environments, such as in swales, whereas A. flava prefers mesic, shaded sites. Aphthona czwalinae establishes best in moist sites that are open or shaded (Plant Protection and Quarantine, 1989). Preliminary indications are that A. lacertosa survives best in habitats where grass overtops leafy spurge. The different habitat preferences of these flea beetles allows them to attack leafy spurge in the various environmental inhabits. A multivoltine species of flea beetle, A. abdominalis Duftschmid, was screened and approved for release in the United States in 1993 (Fornasari, 1993). Aphthona abdominalis larvae feed on roots, adventitious buds, and underground shoots. Adults feed on stems and leaves of leafy spurge. Unlike the species of Aphthona flea beetle previously discussed, A. abdominalis over- winters as an adult and could potentially produce four generations per year. The distribution of A. abdominalis between Poland and northern Iran (For- nasari, 1993) suggests that it could overwinter throughout the range of leafy spurge in North America, although it may produce less than 4 generations per year in the north. Because of its potential in producing multiple generations per year and the feeding damage it causes to leafy spurge, this flea beetle has great potential as a biological control agent for leafy spurge.

SUMMARY

Leafy spurge is a perennial, pernicious weed that primarily infests pasture and rangeland habitats. Control is expensive, and because of it's spreading root system with adventitious buds, must be treated annually. Techniques that stress the plant throughout the growing season appear to give better control for a longer period of time than single treatments. Research into control techniques using insects and bacteria may provide the key to successful leafy spurge management in the future. 184 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

LITERATURE CITED

Abu-Irmaileh, B.E. 1991. Weed control in squash and tomato ﬁelds by soil solarization in the Jordan Valley. Weed Res. 31:125-133. Alley, H.P. and C.G. Messersmith. 1985. Chemical control of leafy spurge. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3. Weed Sci. Soc. Amer. pp. 65-78. Al-Masoom, A.A., S. Abdur-Rahman, and S. Itani. 1993. Soil solarized for weed management in U.A.E. Weed Tech. 7:507-510. Andres, L.A., C.J. Davis, P. Harris, and A.J. Wapshere. 1976. Biological control of weeds. In C.B. Huffaker and P.S. Messenger, (eds.) Theory and Practice of Biological Control, Academic Press, NY, pp. 481-499. Arny, A.C. 1932. Variations in the organic reserves in underground parts of ﬁve perennial weeds from late April to November. Minn. Agr. Exp. Sta. Tech. Bull. 84. Bakke, A.L. 1936. Leafy spurge, Euphorbia esula L. Iowa Agr. Exp. Sta. Bull. 198:209-245. Bakke, A.L. 1937. Control of leafy spurge, Euphorbia esula L. Iowa Agr. Exp. Sta. Bull. 222. Bakshi, T.S. and R.T. Coupland. 1959. An anatomical study of the subterranean organs of Euphorbia esula in relation to its control. Can. J. Bot. 37:613-620. Best, K.F, G.G. Bowes, A.G. Thomas, and M.G. Maw. 1980. The biology of Canadian weeds. 39. Euphorbia esula L. Can. J. Plant Sci. 60:651-663. Biesboer, D.D. and P.G. Mahlberg. 1978. Accumulation of non-utilizable starch in laticifers of Euphorbia heterophylla and E. myrsinites. Planta 143:5-10. Biesboer, D.D. and P.G. Mahlberg. 1981. A comparison of alpha-amylases from the latex of three selected species of Euphorbia (Euphorbiaceae). Amer. J. Bot. 68:498-506. Blockstein, D.E., B.D. Maxwell, and P.K. Fay. 1987. Dispersal of leafy spurge seeds (Euphorbia esula) by mourning doves (Zenaida macroura). Weed Science 35:160-162. Bybee, T.A. 1979. Factors affecting leafy spurge control including leafy spurge reestablishment, herbicide application dates, herbicide translocation, and root carbohydrates. Ph.D. Thesis, North Dakota State Univ. 88 pp. Christianson, K. and R. Lym. 1984. Herbicides and grass competition for leafy spurge. Leafy Spurge News. 16(3):12. Clarkson, V.A. 1960. Effect of black polyethylene mulch on soil and microclimate temperature and nitrate level. Agronomy J. 52:307-309. Clay, S.A. and C.M. Scholes. 1997. Alternative methods for leafy spurge (Eu- phorbia esula) control. Weed Sci. (in review) Coupland, R.T. and J.F. Alex. 1955. Distribution of vegetative buds on the underground parts of leafy spurge (Euphorbia esula L.).Can. J. Ag. Sci. 35:76- 82. Croizat, L. 1945. “Euphorbia esula” in North America. Am. Midl. Nat. 33:231- 243. Cyr, D.R. and J.D. Bewley. 1989. Carbon and nitrogen reserves of leafy spurge (Euphorbia esula) roots as related to overwintering strategy. Phys. Plant. 77:67-72. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 185

Cyr, D.R. and J.D. Bewley. 1990. Seasonal variation in nitrogen storage reserves in the roots of leafy spurge (Euphorbia esula) and responses to decapitation and defoliation. Phys. Plant. 78:361-366. Derscheid, L.A., K.E. Wallace, and R.L. Nash. 1960. Leafy spurge control with cultivation, cropping and chemicals. Weeds 8:42-56. Derscheid, L.A., L.J. Wrage, and W.E. Arnold. 1985. Cultural control of leafy spurge. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3. Weed Sci. Soc. Amer. pp. 57-64. Dunn, P.H. 1979. The distribution of leafy spurge (Euphorbia esula) and other weedy Euphorbia spp. in the United States. Weed Sci. 27:509-516. Dunn, P.H. 1985. Origins of leafy spurge in North America. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3. Weed Sci. Soc. Amer. pp. 7- 13. Ebke, D.H. and M.K. McCarty. 1983. A nursery study of leafy spurge (Euphor- bia spp.) complex from North America. Weed Sci. 31:866-873. Egley, G.H. 1990. High-temperature effects on germination and survival of weed seeds in soil. Weed Sci. 38:429-435. Evans, F.J. and A.D. Kinghorn. 1975. The succulent Euphorbias of Nigeria. Part I. Lloydia 38:363-365. Evans, F.J. and A.D. Kinghorn. 1977. A comparative phytochemical study of the diterpenes of some species of the genera Euphorbia and Elaeophorbia (Eu- phorbiaceae). Bot. J. Linn. Soc. 74:23-35. Fornasari, L. 1993. Life history of the flea beetle, Aphthona abdominalis Duftschmid, on Euphorbia esula L. (leafy spurge) in Italy. Biological Con- trol 3:161-175. Forwood, J.R. and M.K. McCarty. 1983. A nursery study of leafy spurge (Eu- phorbia esula) in Nebraska with the spurge hawkmoth (Hyles euphorbiae). Weed Sci. 28:235-240. Frazier, J.C. 1943. Food reserve depletion and synthesis in field bindweed, Con- volvulus arvensis L., as related to 7-day and 14-day intervals of cultivation. Plant Phys. 18:315-323. Galitz, D.S. and D.G. Davis. 1983. Leafy spurge physiology and anatomy. ND Farm Res. Bull. 40:20-26. Gylling, S. R. and W. E. Arnold. 1985. Efficacy and economics of leafy spurge (Euphorbia esula) control in pasture. Weed Sci. 33:381-385. Harris, P. 1984. Euphorbia esula-virgata complex, leafy spurge and E. cyparissiae L., cypress spurge (Euphorbiaceae). Pp. 83-88 In Biological control programmes against insects and weeds in Canada 1959-1968. Tech. Com. CIBC 4. 266 pp. Harris, P., P.H. Dunn, D. Schroeder, and R. Vonmoos. 1985. Biological control of leafy spurge in North America. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3. Weed Sci. Soc. Amer. pp. 79-92. Harvey, S.J., R.M. Nowierski, P.G. Mahlberg, and J.M. Story. 1988. Taxonomic evaluation of leaf and latex variability of leafy spurge (Euphorbia spp.) for Montana and European accessions. Weed Sci. 36:726-733. Holden, A.N.G. and P.G. Mahlberg. 1992. Application of chemotaxonomy of leafy spurges (Euphorbia spp.) in biological control. Can. J. Bot. 70:1529- 1536. 186 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Horowitz, M., Y. Regev, and G. Herzlinger. 1983. Solarization for weed control. Weed Sci. 31:170-179. Johnston, A. and R.W. Peake. 1960. Effect of selective grazing by sheep on the control of leafy spurge (Euphorbia esula L.). J. Range Manage. 13:192-195. Julien, M.H. 1987. Biological control of weeds: A world catalogue of agents and their weeds. 2nd ed. CAB International Institute of Biological Control, Un- win Brothers Ltd., Surrey. 144 pp. Kronberg, S.L., R.B. Muntifering, E.L. Ayers, and C.B. Marlow. 1993. Cattle avoidance of leafy spurge: A case of conditioned aversion. J. Range Man- age. 46:364-366. Kronberg, S.L. and J.W. Walker. 1993. Ruminal metabolism of leafy spurge in sheep and goats: A potential explanation for differential foraging on spurge by sheep, goats, and cattle. J. Chem. Ecol. 19:2007-2017. Kumar, B., N.T. yaduraju, K.N. Ahuja, and D. Prasad. 1993. Effect of soil solarization on weeds and nematodes under tropical Indian conditions. Weed Res. 33:423-429. Landgraf, B.K., P.K. Fay, and K.M. Havstad. 1984. Utilization of leafy spurge (Euphorbia esula) by sheep. Weed Sci. 32:348-352. Lym, R.G. and C.G. Messersmith. 1985. Cost effectiveness of leafy spurge control during a five-year management plan. ND Farm Res. 43:7-14. Lym, R.G. and C.G. Messersmith. 1987. Carbohydrates in leafy spurge roots as influenced by environment. J. Range Manage. 40:139-144. Lynn, K.R. and N.A. Clevette-Radford. 1987. Biochemical properties of latices from the Euphorbiaceae. Phytochemistry 26:939-944. Mahlberg, P.G. and P.S. Sabharwal. 1968. Origin and early development of nonarticulated laticifers in embryos of Euphorbia marginata. Amer. J. Bot. 55:375-381. Manners, G.D. and D.G. Davis. 1987. The characterization of esulone C and chemotaxonomy of jatrophane diterpenes in leafy spurge. Phytochemistry 26:727-730. Martin, C.A., H.G. Ponder, and C.H. Gilliam. 1991. Evaluation of landscape fab- rics in suppressing growth of weed species. J. Environ. Hort. 9:38-40. Mattson, W.J. 1980. Herbivory in relation to plant nitrogen content. Ann. Rev. Ecol. Syst. 11:119-161. Maw, E. 1981. Biology of some Aphthona spp. (Col.:Chrysomelidae) feeding on Euphorbia spp. (Euphorbiaceae), with special reference to leafy spurge (Euphorbia sp. near esula). M.S. Thesis, Univ. of Alberta. McDaniel, B., S.A. Clay, and C. Scholes. 1992. Morphology of three imported Aphthona flea beetles used as biological control agents of leafy spurge. S.D. Agr. Exp. Sta. Bull. TB98. Messersmith, C.G., R.G. Lym, and D.S. Galitz. 1985. Biology of leafy spurge. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3 Weed Sci. Soc. Amer. pp. 42-56. Myers, G.A., C.A. Beasley, and L.A. Derscheid. 1964. Anatomical studies of Eu- phorbia esula L. Weeds 12:291-295. Nissen, S.J. and M.E. Foley. 1987. Correlative inhibition and dormancy in root buds of leafy spurge (Euphorbia esula). Weed Sci. 35:155-159. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 187

Nissen, S.J., R.A. Masters, D.J. Lee, and M.L. Rowe. 1992. Comparison of restriction fragment length polymorphisms in chloroplast DNA of ﬁve leafy spurge (Euphorbia spp.) accessions. Weed Sci. 40:63-67. Noble, D.L., P.H. Dunn, and L.A. Andres. 1979. Proceedings of Leafy Spurge Symposium, pp. 8-14. Pemberton, R.W. 1988. Myrmecochory in the introduced range weed, leafy spurge (Euphorbia esula L.). Am. Midl. Nat. 119:431-435. Peters, E.J. and S.A. Lowance. 1978. Effects of multiple mowing on western ironweed (Vernonia baldwinii) and gray goldenrod (Solidago nemoralis). Weed Sci. 26:190-192. Plant Protection and Quarantine. 1989. Biological control of leafy spurge (Eu- phorbia esula L.). USDA-Animal and Plant Health Inspection Service, unpublished. Radcliffe-Smith, A. 1985. Taxonomy of North American leafy spurge. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3. Weed Sci. Soc. Amer. pp. 57-64. Raju, M.V.S. 1985. Morphology and anatomy of leafy spurge. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3. Weed Sci. Soc. Amer. pp. 26- 41. Raju, M.V.S., T.A. Steeves, and R.T. Coupland. 1963. Developmental studies on Euphorbia esula L.: Morphology of the root system. Can. J. Bot. 41:579-589. Rees, N.E., R.W. Pemberton, A. Rizza, and P. Pecora. 1986. First recovery of Oberea erythrocephala on the leafy spurge complex in the United States. Weed Sci. 34:395-397. Ricotta, J.A. and J.B. Masiunas. 1991. The effects of black plastic mulch and weed control strategies on herb yield. Hortscience 26:539-541. Schimming, W.K. and C.G. Messersmith. 1988. Freezing resistance of overwintering buds of four perennial weeds. Weed Sci. 36:568-573. Scholes, C.M. 1996. Suppression of leafy spurge (Euphorbia esula L.) with alternative chemical, mechanical, and biological methods. PhD thesis. SD- SU. 131 pg. Selleck, G.W., R.T. Coupland, and C. Frankton. 1962. Leafy spurge in Saskatchewan. Ecol. Mono. 32:1-29. Sommer, G. and E. Maw. 1982. Aphthona cyparissiae (Koch) and A. ﬂava (Guill.) (Coleoptera:Chrysomelidae): Two candidates for the biological control of cypress and leafy spurge in North America. CIBC Report. 60 pp. Spencer, N. 1994. Introduced insects for leafy spurge control. Leafy Spurge News 16(2):4-5. Stahevitch, A.E., C.W. Crompton, and W.A. Wojtas. 1988. Cytogenetic studies of leafy spurge, Euphorbia esula, and its allies (Euphorbiaceae). Can. J. Bot. 66:2247-2257. Steeves, T.A. and I.M. Sussex. 1989. Patterns in plant development. Cambridge Univ. Press. 388 pp. Tworkoski, T. 1992. Developmental and environmental effects on assimilate partitioning in Canada thistle (Cirsium arvense). Weed Sci. 40:79-85. Valcarce, R.V., G.G. Smith, J.M. Torell, and J.O. Evans. 1989. Differentiation of leafy spurge (Euphorbia esula L.) accessions based on chemical constitu- 188 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

tion using curie-point pyrolysis-mass spectrometry-pattern recognition. J. Analytical Applied Pyrolysis 17:5-24. Vizantinopoulos, S. and N. Katranis. 1993. Soil solarization in Greece. Weed Res. 33:225-230. Walker, J.W., K.G. Hemenway, P.G. Hatﬁeld, and H.A. Glimp. 1992. Training lambs to be weed eaters: Studies with leafy spurge. J. Range Manage. 45:245-249. Watson, A.K. 1985. Introduction—The leafy spurge problem. In A.K. Watson (ed.), Leafy Spurge. Monograph Series No. 3. Weed Sci. Soc. Amer. pp. 1- 6. Wilson, F. and C.B. Huffaker. 1976. The philosophy, scope, and importance of biological control. In C.B. Huffaker and P.S. Messenger, (eds.) Theory and Practice of Biological Control, Academic Press, NY, pp. 3-15. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 189

STRESS EFFECTS AND EVOLUTION OF DIRECT DEVELOPMENT IN ECHINOID ECHINODERMS

Leland G. Johnson and Teresa L. Fitzgerald-Brown Department of Biology Augustana College Sioux Falls, SD 57197

Keywords

Development and evolution, Stress and evolutionary divergence, Direct development, Echinoid development, Lytechinus pictus, Heliocidaris tuberculata

Many invertebrate animals develop indirectly via feeding larval stages and metamorphosis into adults. However, others produce larger eggs, bypass feeding larval stages, and develop adult structures directly. Indirect development with feeding larvae is a highly conserved ancestral pattern, but direct development with the elimination of feeding larval stages has arisen independently in numerous lineages. The widely distributed origins of direct development and the occurrence of closely related species which differ in the presence or absence of feeding larvae raise evolutionary questions. How has early development been reorganized in direct developers? And, what selective factors might drive the transition from indirect to direct development with its larger per offspring investment? In this paper, we consider several hypotheses regarding selective forces in the evolution of direct development, and we suggest the additional hypothesis that stress effects on early developmental processes could act selectively in the divergence of direct and indirect development. The reorganization of early development that has occurred during the evolution of direct development has been investigated in some detail in echinoid echinoderms. During early ontogeny of indirectly developing sea urchins (Fig. 1), gut morphogenesis and the assembly of the larval skeletal spicules by clustered primary mesenchyme cells are important steps in development of the feeding pluteus larva (Okazaki, 1975; McClay et al. 1991; Ettensohn, 1991; Decker and Lennarz, 1988). But early development is highly modiﬁed in direct developing urchins (Williams and Anderson, 1975; Raff, 1987; Wray and Raff, 1991; Parks et al. 1988; Raff, 1992a; Wray, 1992, 1994; Emlet, 1995). Lar- val guts and skeletons of direct developers are reduced or absent and they have modiﬁed cleavage and gastrulation patterns. There are heterochronies (timing differences) in the origins of various cell lineages and in gene expression patterns. Several aspects of adult development are initiated earlier than in indirectly developing species. As with other invertebrate groups, indirect development is the ancestral pattern among echinoid echinoderms. Direct development has arisen independently in several lineages (Emlet et al. 1987; Emlet, 1990; Wray, 1995), and there is considerable developmental diversity among direct developing echinoids (Wray and Raff, 1991a,b; Wray, 1996). 190 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Despite the progress that has been made in analysis of the evolutionary reorganization of early development of direct developing echinoids, questions remain concerning selection in the evolutionary divergence of direct and indirect development among echinoids, as they do for other marine invertebrate groups (Thorson, 1950; Vance, Figure 1. A composite sketch showing sev- 1973; Christiansen and eral major features of early development in Fenchel, 1979; Strathmann, an indirectly developing sea urchin embryo, 1985; Emlet et al. 1987; and the dimensions measured for gut length Wray and Raff, 1991; to gastrula length ratio calculations. At the Havenhand, 1995: Herrera, end of the cleavage cell divisions, the blastu- et al., 1996). Loss of larval la consists of a hollow sphere of cells sur- features associated with di- rounding a fluid-filled cavity, the blastocoel. rect development has not Primary mesenchyme cells leave the blastula placed evident constraints wall, migrate over its inner surface, and form on adult morphology nor clusters at distinct loci, where they cooperate led to obvious morphologi- to initiate skeleton development by organiz- cal innovation in adults. ing triradiate, crystalline spicules. One side Thus, selective forces ap- of the blastula wall indents (invaginates) to pear to have acted directly initiate gut development. The gut lengthens, on embryonic and larval nearly spans the cavity, and then bends to development. However, one side at its tip to contact the oral ectoderm adaptive significances of A gut length to gastrula length ratio of one in- developmental pattern dif- dicates completion of these processes, not ferences are difficult to re- complete spanning of the blastocoel cavity. solve, especially where closely related species oc- cupy common or very similar habitats, but employ quite different developmental strategies (Williams and Anderson, 1975; Raff, 1987; Johanneson,1988; Byrne and Barker, 1991; Levin, 1984; Levin et al. 1987, 1991; Bridges, 1993). The production of numerous feeding larvae by indirectly developing species promotes dispersal, but larvae are at risk of predation and accidental mortality during their sometimes lengthy period in the plankton. Direct development with parental brooding, demersal development, or with only a brief period spent in the water column reduces risk exposure of the less numerous offspring. Fluctuating larval food supply also may selectively favor direct development. Yet another possibility is that selection for accumulation of materials reserved for support of juvenile development after metamorphosis could favor larger egg sizes (Emlet and Heogh-Guldberg, 1997). Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 191

However, advantages achieved through direct development must be weighed against the greater parental energy investment in each offspring. Though some caution is needed regarding conclusions about energy investment drawn from egg or embryo size alone (Bridges, 1993), utilization of direct development with production of large eggs decreases gamete numbers (e. g., McEdward and Janies, 1993) and is likely to reduce the potential for dispersal. One type of hypothesis concerning forces that might drive the transition to direct development in echinoids is based on observed plasticity in the development of pluteus larvae which is correlated with experimental reduction of the biomass of developing embryos by separation of blastomeres of two-cell and four-cell embryos (Sinervo and McEdward, 1988; McEdward, 1996; Her- rera, et al., 1996 ) or with experimental or natural variations in the exogenous food supply available to feeding larvae (Strathmann et. al. 1992). The hypothesis suggests that genetic changes affecting oogenesis could increase the en- dogenous nutrient supply and could parallel changes in developmental pattern made possible by already available plasticity in larval development. Further development of this hypothesis depends upon several factors including identification of selective forces favoring increased egg size and elucidation of developmental plasticity in early stages of indirect development that is even greater than that observed for dimensions of the pluteus or for development subsequent to the beginning of feeding. Increased probability of fertilization success in situations of sperm limitation is a potential selective force favoring increased egg size, but the significance of this factor depends upon the extent to which sperm limitation affects fertilization kinetics in nature (Levitan, 1993). Furthermore, increased egg size may not have been a prerequisite to evolution of direct development because development of direct developing species can proceed normally after experimental reduction of embryo biomass, even reduction into the size ranges of indirectly developing species (Okazaki and Dan, 1954; Henry and Raff, 1990; Wray and Raff, 1991; Hoegh-Guldberg and Emlet, 1997; Emlet and Heogh- Guldberg, 1997). Thus, direct development might have emerged first with selection for the larger egg size that would better support nonfeeding development acting subsequently. Raff (1992b) has concluded that early developmental mechanisms previously thought to be quite constrained can, in fact, evolve readily and thus permit rapid evolution of early development. Diverse environmental factors affect the embryonic and larval morphogenesis of indirectly developing sea urchins (e. g., Dan and Okazaki, 1956; Bran- driff et al. 1975; Riederer-Henderson, 1988; Johnson et al. 1989; Fujiwara et al. 1990). Thus, we have considered the possibility that disturbance of early development by environmental stresses could be a selective factor in evolutionary divergence of early developmental patterns, and we have examined some developmental disturbances in embryos of two indirectly developing sea urchins, Lytechinus pictus and Heliocidaris tuberculata experimentally subjected to an environmental stress, lowered ambient temperature. Lytechinus pictus embryos were transferred from 16º C to lower temperatures just after hatching, the time when gastrulation and subsequent processes 192 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) critical to the assembly of the feeding pluteus larva begin. This is also the time when hatched embryos in nature begin to swim actively and enter the water column. Embryos transferred to 11º C developed normally, but development was abnormal in many of the embryos transferred to 8.5º C or 6º C. Gut development in temperature-stressed L. pictus embryos (Fig. 2) was highly variable. Although some embryos eventually completed gut elongation at 8.5º C and a few did so at 6º C, most did not. Low temperatures impair sea urchin microtubule assembly (Suprenant and Marsh, 1987), but sea urchin gastrulation appears to be microtubule independent (Anstrom, 1989; Hardin, 1987). Low temperatures also affect actin filaments in sea urchin eggs and embryos (San- tella and Monroy, 1989).

B A

Figure 2. Scatter plots of gut length to gastrula length ratios for: A) Control embryos maintained continuously at 16º C, B) Embryos transferred to 8.5º C as newly- hatched blastulae, and C) Embryos transferred to 6º C as newly- hatched blastulae. Numbers of embryos at points that represent more than one embryo are given C only for embryos with completed gut development. Lytechinus pictus adults were obtained from Pacific BioMarine, Venice, CA. Fertilization and cleavage took place in Woods Hole Formula artificial seawater at 16º C. New- ly hatched blastula-stage embryos were transferred to seawater in controlled temperature chambers at 8.5º C (eggs from three females) and 6º C (eggs from five females) where their subsequent development was observed and compared with control embryos maintained at 16º C (eggs from eight females). Samples of swimming embryos were removed at intervals. Gut length and gastrula length measurements were made with an eyepiece micrometer. Sampling was continued in lower temperature cultures until most embryos stopped swimming, dropped to the bottom of the culture, and ceased developing. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 193

Observation of skeletal spicule development in control embryos by polar- izing microscopy initially reveals a small birefringent dot within each of two clusters of primary mesenchyme cells. These rudimentary spicules grow and become triradiate. Thus, the normal progression is: no visible spicules to two spicule dots to two triradiate spicules (Okazaki, 1975; McClay et al. 1991; Et- tensohn, 1991; Decker and Lennarz, 1988). Figure 3 compares spicule development in control and chilled L. pictus embryos near the mid-point of gut development. Nearly all embryos developing at 6º C produced either abnormal sets of skeletal spicules or no spicules. Only one among all the embryos observed at 6º C had both a completed gut and apparently normal spicules. Some cold-exposed L. pictus embryos not included in the data summarized in Figures 2 and 3, displayed additional abnormalities similar to those resulting from other experimental treatments (Dan and Okazaki, 1956; Brandriff et al. 1975; Riederer-Henderson, 1988; Fujiwara et al. 1990; Katow, 1989; Nocente- McGrath et al. 1990). A few exogastrulated; their guts extended outward rather than inward, just as occurs following some chemical treatments. Many chilled embryos developed no guts, and some developed an abnormally short, side- directed gut. Some chilled embryos produced so many mesenchyme-like cells that the blastocoel (inner cavity) became crowded with cells. We observed birefringent spicule spots in a few of these "packed blastula" embryos. Thus, at least some of the entering cells differentiated as skeleton-producing cells. In Heliocidaris tuberculata, only “packed blastula” embryos developed at 8º C (Table 1), while nearly all hatched embryos produced normal plutei at 12º C No data were obtained for H. tuberculata development at temperatures between 8º C and 12º C. After observing these stress-associated disturbances of early stages of indirect development and noting those observed by others, we hypothesize that selection favoring emergence of direct developmental pathways could result when environmental stresses interfere with cellular processes that are essential for morphogenesis of feeding larvae. These stress-sensitive developmental processes could be bypassed by accelerated (heterochronic) initiation of adult development. Coincidentally or subsequently, there would also be selection favoring increased egg nutrient stores which facilitate non-feeding, yolk-supported (lecithotrophic) development. Thus, environmental stresses could dis- ruptively select for divergence of direct developing strains from more devel- opmentally resilient ones capable of developing feeding larvae under the existing conditions and thereby persisting in use of indirect development. There are broad theoretical questions about how selection acts on developmental processes (Maynard Smith et al. 1985; Wolpert, 1990), and complex questions concerning evolution-environmental stress relationships (Lawrence, 1991; Parsons, 1991, 1994). However, our hypothesis does suggest some spe- ciﬁc and immediate questions for further investigation. Are there spatial or temporal disturbances of gene expression patterns in environmentally-stressed indirectly developing embryos? Do cells in their “packed blastulae” express genes speciﬁc to primary mesenchyme cells? Is there any evidence of heterochronic expression of genes associated with later development (e.g. adult 194 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 3. Spicule development in embryos having gut length to gastrula length ratios between 0.45 and 0.55 (see legend Fig. 2). Spicules, which are crystalline and birefringent, were examined using polar- izing optics. The “Abnormal Set” category includes embryos with only one dot, one “dash” (an elongate, curved spicule), more than two dots or dashes, or other spicule sets never seen in control embryos. The “Nor- mal (?) Set” category includes embryos with one dot and one dash, two dashes, or one triradiate spicule and one other spicule. Similar differences in patterns of spicule development were observed in embryos with greater gut elongation. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 195 skeleton development) in such stressed embryos? Results of these suggested investigations and others could lead to reﬁnement or rejection of this hypothesis.

ACKNOWLEDGEMENTS

Funding was received from the Augustana Research and Artist Fund and the Visiting Investigator Program of the Australian Institute of Marine Science. Ian Hutton provided H. tuberculata collected at Lord Howe Island.

LITERATURE CITED

Anstrom, J. A. 1989. Sea urchin primary mesenchyme cells: Ingression occurs independent of microtubules. Dev. Biol. 131: 269-275. Brandriff, B., R. T. Hinegardner and R. Steinhardt. 1975. Development and life cycle of the parthenogenetically activated sea urchin embryo. J. Exp. Zo- ol. 192: 13-24. Bridges, T. S. 1993. Reproductive investment in four developmental morphs of Streblospio (Polychaeta: Spionidae). Biol. Bull. 184: 144-152. Byrne, M. and M. F. Barker. 1991. Embryogenesis and larval development of the asteroid Patiriella regularis viewed by light and scanning microscopy. Biol. Bull. 180: 332-345. Christiansen, F. B. and T. M. Fenchel. 1979. Evolution of marine invertebrate reproductive patterns. Theor. Popul. Biol. 16: 267-282. Dan, K. and K. Okasaki. 1956. Cyto-embryological studies of sea urchins. III. Role of the secondary mesenchyme cells in the formation of the primitive gut in sea urchin larvae. Biol. Bull. 110: 29-42. Decker, G. L. and W. J. Lennarz. 1988. Skeletogenesis in the sea urchin embryo. Development 103: 231-247. Emlet, R. B. 1990. World patterns of developmental mode in echinoid echinoderms. Pages 329-335, in M. Hoshi and O. Yamashita, eds. Advances in Invertebrate Reproduction 5. Elsevier, Amsterdam. Emlet, R. B. 1995. Larval spicules, cilia, and symmetry as remnants of indirect development in the direct developing sea urchin Heliocidaris erythrogramma. Dev. Biol. 167:405-415. Emlet, R. B., L. R. McEdward, and R. R. Strathmann. 1987. Echinoderm larval ecology viewed from the egg. Pages 55-136, in M. Jangoux and J. M. Lawrence, eds. Echinoderm Studies. A. A. Balkema, Rotterdam. Emlet, R. B. and O. Hoegh-Guldberg. 1997. Effects of egg size on postlarval performance: Experimental evidence from a sea urchin. Evolution 51: 141- 152. Ettensohn, C. A. 1991. Mesenchyme cell interactions in the sea urchin embryo. Pages 175-202 in J. Gearhart, ed. Cell-Cell Interactions in Early De- velopment. Wiley-Liss, New York. Fujiwara, A., K. Kataoka, K. Mikami-Takei, E. Tazawa, and I. Yasumasu. 1990. Vegetalization induced by procaine and tetracaine in sea urchin embryos. Dev. Growth Diff. 5: 489-495. 196 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Hardin, J. 1987. Archenteron elongation in the sea urchin embryo is a microtubule-independent process. Dev. Biol. 121: 253-262. Havenhand, J. N. 1995. Evolutionary ecology of larval types. Pages 79-122 in L. McEdward, ed. Ecology of Marine Invertebrate Life. CRC Press, Boca Ra- ton, FL. Herrera, J. C. , S. K. McWeeney, and L. R. McEdward. 1996. Diversity of en- ergetic strategies among echinoid larvae and the transition from feeding to nonfeeding development. Oceanol. Acta 19:313-321. Hoegh-Guldberg, O. and R. B. Emlet. 1997. Energy use during the development of a lecithotrophic and a planktrotrophic echinoid. Biol. Bull. 192: 27-40. Johannesson, K. 1988. The paradox of Rockall: why is a brooding gastropod (Littorina saxitalis) more widespread than one having a planktonic larval dispersal stage (L. littorea)? Mar. Biol. 99: 507-513. Johnson, L. G., R. M. Monger, and G. K. May. 1989. Evans blue treatment promotes blastomere separation and twinning in Lytechinus pictus embryos. Dev. Biol. 131: 276-279. Katow, H. 1989. Characterization of vegetal plate cells separated from cy- tochalasin B-treated blastulae of the sea urchin, Clypeaster japonicus. Dev. Growth Diff. 31: 439-446. Lawrence, J. 1991. Analysis of characteristics of echinoderms associated with stress and disturbance. Pages 11-26 in T. Yanagisawa, I. Yasumasu, C. Oguro, N. Suzuki, and T. Motokawa, eds. Biology of Echinodermata. A. A. Balkema, Rotterdam. Levin, L. A. 1984. Multiple patterns of development in Streblospio benedicti Webster (Spionidae) from three coasts of North America. Biol. Bull. 166: 494-508. Levin, L. A., H. Caswell, K. D. DePatra, and E. L. Creed. 1987. Demographic consequences of larval development mode: Planktotrophy vs. lecithotrophy in Streblospio benedicti. Ecology 68: 1877-1886. Levin, L. A., J. Zhu, and E. Creed 1991. The genetic basis of life-history characters in a polychaete exhibiting planktotrophy and lecithotrophy. Evo- lution 45: 380-397. Levitan, D. R. 1993. The importance of sperm limitation to the evolution of egg size in marine invertebrates. Am. Nat. 141: 517-536. Maynard Smith, J. R., R. Burian, S. Kaufman, P. Alperch, J. Campbell, B. Good- win, R. Lande, D. Raup, and L. Wolpert. 1985. Developmental constraints and evolution. Quart. Rev. Biol. 60:265-287. McClay, D. R., J. Morrill, and J. Hardin. 1991. Archenteron morphogenesis in the sea urchin. Pages 15-30 in J. Gearhart, ed. Cell-Cell Interactions in Early Development, Wiley-Liss, New York. McEdward, L. R. 1996. Experimental manipulation of parental investment in echinoid echinoderms. Am.. Zool. 36:169-179. McEdward, L. R. and D. A. Janies. 1993. Life cycle evolution in asteroids: What is a larva? Biol. Bull. 184: 255-268. Nocente-McGrath, C., R. McIsaac, and S. G. Ernst. 1991. Altered cell fate in Li- Cl-treated sea urchin embryos. Dev. Biol. 147: 445-450. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 197

Okazaki, K. 1975. Normal development to metamorphosis. Pages 177-232 in G. Czihak, ed. The Sea Urchin Embryo: Biochemistry and Morphogenesis, Springer-Verlag, Berlin. Okazaki, K, and K. Dan. 1954. The metamorphosis of partial larvae of Per- onella japonica Mortensen, a sand dollar. Biol. Bull. 106: 83-99. Parks, A. L., B. A. Parr, J-E. Chin, D. S. Leaf, and R. A. Raff. 1988. Molecu- lar analysis of heterochronic changes in the evolution of direct developing sea urchins. J. Evol. Biol. 1: 27-44. Parsons, P. A. 1991. Evolutionary rates: Stress and species boundaries. Ann. Rev. Ecol. System. 22: 1-18. Parsons, P. A. 1994. Habitats, stress, and evolutionary rates. J. Evol. Biol. 7: 387-397. Raff, R. A. 1987. Constraint, ﬂexibility, and phylogenetic history in the evolution of direct development in sea urchins. Dev. Biol. 119: 6-19. Raff, R. A. 1992a. Direct-developing sea urchins and the evolutionary reorganization of early development. BioEssays 14: 211-218. Raff, R. A. 1992b. Evolution of developmental decisions and morphogenesis: The view from two camps. Development 1992 Supplement: 15-22. Riederer-Henderson, M. A. 1988. Effects of theophylline on expression of the long cilia phenotype in sand dollar blastulae. J. Exp. Zool. 246: 17-22. Santella, L. and A. Monroy. 1989. Cold shock induces actin reorganization and polyspermy in sea urchin eggs. J. Exp. Zool. 252: 183-189. Sinervo, B. and L. R. McEdward. 1988. Developmental consequences of an evolutionary change in egg size: an experimental test. Evolution 45: 885- 899. Strathmann, R. R. 1985. Feeding and nonfeeding larval development and life- history evolution in marine invertebrates. Ann. Rev. Ecol. System. 16: 339- 361. Strathmann, R. R., L. Fenaux, and M. F. Strathmann. 1992. Heterochronic developmental plasticity in larval sea urchins and its implications for evolution of nonfeeding larvae. Evolution 46: 272-286. Suprenant, K. A. and J. C. Marsh. 1987. Temperature and pH govern the self- assembly of microtubules from unfertilized sea-urchin egg extracts. J. Cell Sci. 87: 71-84. Thorson, G. 1950. Reproduction and larval ecology of marine bottom invertebrates. Biol. Rev. 25: 1-45 . Vance, R. R. 1973. More on reproductive strategies in marine benthic invertebrates. Am. Nat. 107:353-361. Williams, D. H. C. and D. T. Anderson. 1975. The reproductive system, embryonic development, larval development and metamorphosis of the sea urchin Heliocidaris erythrogramma (Val.) (Echinoidea:Echinometridae) Aust. J. Zool. 23: 371-403. Wolpert, L. 1990. The evolution of development. Biol. J. Linn. Soc. 39: 109- 124. Wray, G. A. 1992. Rates of evolution in developmental processes. Am. Zo- ol. 32: 123-134. Wray, G. A. 1994. Developmental evolution: New paradigms and paradoxes. Dev. Genetics 15: 1-6. 198 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Wray, G. A. 1995. Punctuated evolution of embryos. Science 267: 1115-1116. Wray, G. A. 1996. Parallel evolution of nonfeeding larvae in echinoids. Syst. Biol. 45: 308-322. Wray, G. A. and R. A. Raff. 1991a. The evolution of developmental strategy in marine invertebrates. Trends Ecol. Evol. 6: 45-50. Wray, G. A. and R. A. Raff. 1991b. Rapid evolution of gastrulation mechanisms in a sea urchin with lecithotrophic larvae. Evolution 45: 1741-1750. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 199

COMPARING ESTIMATES OF SOME BIVARIATE SURVIVAL FUNCTIONS UNDER RANDOM CENSORSHIP

Bonaventure A. Anthonio College of Natural Sciences Dakota State University Madison, SD 57042

ABSTRACT

In many estimation problems, it is inconvenient or impossible to make complete measurements on all members of a random sample. The problem is to estimate the distribution of the life times under random censoring. The onedimensional random censoring model has been treated in great detail in the recent literature beginning with the landmark papers of Kaplan and Meier (1958). Various methods for the nonparametric estimation of a bivariate survival function in the presence of censoring have been proposed by a number of authors. In the present paper, the performance of some of these competing estimators were compared. Two pathdependent estimators which were proposed by Burke (1988) was considered. An extension of the Kaplan and Meier (1958) estimator to the case of bivariate censored data as was proposed by Korwar (1982) was also included in the study. Other competing estimators were the path dependent and closed form estimators due to Campbell and Foldes (1980). An estimator proposed by Tsai, Leurgans and Crowley (1986) was also investigated. A simple scheme for generating the random variables which were used in this study as well as a comparison of the performance of the estimators is presented.

INTRODUCTION

Censorship

In lifetesting, medical followup and other ﬁelds, the observation of the occurrence time of the event of interest (called a death) may be prevented for some of the items of the sample by the previous occurrence of some other event (called a loss). In medical followup studies to determine the distribution of survival times after treatment, contact with some individuals may be lost before their death, and others may die from causes which we may desire to ex- clude from consideration. An incomplete observation is said to be censored and its numerical value can be referred to as a limit of observation. These limits of observations are constants or values of other random variables, which are assumed to be independent of the complete observations. 200 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Bivariate Censored Data

The case of censored bivariate data was considered. To model the censoring, consider a sequence of independent random censoring vectors

from a bivariate distribution.

Let G(s,t) = P(C > s, D > t). Further, let be independent pairs of random variables from a joint distribution function

Fd (s,t) = P(XO ≤ S, YO ≤ t) where is a random sample of pairs of lifetimes from .

The observed quantities are where

Objective

The objective is to estimate the distribution function and its corresponding survival function in the presence of censoring. A number of authors have proposed estimators for the survival function above. We shall introduce the work of some of these authors with a view towards comparing the proposed estimators.

The Kaplan Meier Product-Limit Estimator: Kaplan and Meier (1958) proposed a productlimit estimator for the uni- variate survival function F(s) = P(XO > S). The proposed estimator is given by: Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 201

≤ ≤ ≤ and where X(1) X(2) … X(n) is an arrangement in ascending order of magnitude of the Xi values.

SOME COMPETING ESTIMATORS OF THE BIVARIATE SURVIVAL FUNCTION

Korwar, R.M., and Dahiya, R.C (1982)

Korwar and Dahiya (1982) extended the work of Kaplan and Meier (1958) to the case of bivariate right censored data. Their proposed estimator is given by:

Campbell, G., and Foldes, A. (1982)

Campbell and Foldes (1982) proposed two pathdependent estimators for the bivariate survival function. Each is a product of two onedimensional Ka- planMeier productlimit estimators. They employed a harzard function approach to estimate – loge F(s, t) and hence F(s, t). Their resulting estimators are: 202 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Where

The following estimator is based on the hazard function

Burke, M. D. (1988)

Burke (1988) suitably modiﬁed the approach of Campbell and Foldes (1982) to ensure that the estimators satisfy the important monotonicity requirements of Survival functions. He also employed the harzard gradient approach identical to the one above to arrive at additional estimators. His proposed estimators are: Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 203

Where

The following estimator is based on the hazard gradient approach:

Tsai, W., Leurgans, S., and Crowley, J. (1986)

Tsai et al, (1986) presented a new class of estimators of the bivariate survival function. This new class is kernel and bandwidth dependent but not 204 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) pathdependent. It is based on the decomposition of the survival function in terms of estimable functions. Their proposed estimator is given by:

Where, Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 205

NUMERICAL RESULTS

Random samples of various sizes were simulated from a given bivariate distribution. There were three different samples corresponding to each sample size: (a) 10% censoring (b) 40% censoring (c) 50% censoring

Generation of Random Variables

Let X1, X2, and X3 be independent and have gamma distributions with in- O dex parameters P1, P2 and P3 respectively. Then X = XI + X3 and O O Y = X2 + X3 have a bivariate gamma distribution. The joint p.d.f of X and YO is given by h(xO, yO) =

For P1 = P2 = P3 = 1, this reduces to 206 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Using the above p.d.f. it can be shown that the corresponding survival function is given by:

The censoring variables were generated using

(a)

(b)

Measure of Goodness of Performance

As a measure of goodness of performance of an estimator, many authors use the mean integrated square error, abbreviated as M.I.S.E. If f is a true survival function and is its estimator, then the M.I.S.E is computed by

The function w is referred as the weight function and is often taken to be identically 1. The integrated square error however, is some what difﬁcult to obtain numerically so it was decided to form the average square error,

at the sample points. For each sample size, the average value of the average square errors was computed. The median value of the average square errors Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 207 were also computed. The motivation behind this is based on the simple real- ization that the median is affected less than the mean by an occasional large error. As a measure of dispersion, the standard deviation of the average errors were also found.

CONCLUSION

This research is limited in scope in the sense that: (i) only one distribution was considered in the generation of the random variables. (ii) The comparisons made among the various estimators were based upon computations made at the sample points. (iii) The observations made are therefore only valid within the framework of the study done.

It was apparent from the numerical results that the dispersion of the average square errors tends to zero with increasing sample size. As expected, it tends to zero quite rapidly with reduced censoring. This did not seem to be the case for estimator (K). An explanation for this could be found in the fact that in deriving the selfconsistent estimator, the weight of the censored observations was spread on all the observations beyond the censored point. The simulation results thus indicated that estimator (K) gives average square errors that are quite erratic. The study also showed that estimator (C 1) performed as well as estimator (C2). This was a conﬁrmation of what was expected from theory. The same observations were made about estimators (B 1) and (B2). In short, it did not seem to matter whether one used the hazard gradient approach or the KaplanMeier productlimit estimators. Estimators (B 1) and (B2) have fewer support points than the other estimators. Their support points are the observed uncensored points (Xi, Yi ). Thus the survival function need only be computed at such support points, and their estimates were fairly simple to compute. Their computations required the least CPU time. Estimators (C 1) and (C2) gave the least average square errors. Such a performance may be due to the fact that they do not depend on other estimable functions but rely heavily on the data. The study also indicated that the average square errors of estimators (B 1), (B2), (C 1) and (C2) on one hand are comparable to those of estimator (T). It is worth mentioning that estimator (T) was the most complicated and required the most CPU time.

REFERENCES

Burke, M. D. (1988). Estimation of a bivariate distribution function under random censorship. Biometrika, 75, 2, 379382. Campbell, G., and Foldes, A. (1982). LargeSample properties of nonparametric bivariate estimators with censored data. Colloquia Mathematica Societatis Janos Bolyai 32, pp. 10321. Campbell, G. (1981). Nonparametric bivariate estimation with randomly censored data. Biometrika, 68, 2, 41722. 208 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 1. Average of Average Square Errors for 10% Censored Data.

Figure 2. Average of Average Square Errors for 40% Censored Data.

Figure 3. Average of Average Square Errors for 50% Censored Data.

Figure 4. Median of Average Square Errors for 10% Censored Data.

Figure 5. Median of Average Square Errors for 40% Censored Data.

Figure 6. Median of Average Square Errors for 50% Censored Data. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 209

Cherian, K.C. (1941). A bivariate correlated gammatype distribution function. Sankhya, 5, 14154. Efron, B. (1967). The two sample problem with censored data. Proc. Fifth Berkeley Symp. Math. Statist. Prob., 4, 88153. Gumbel, E. J. (1960). Bivariate exponential distribution. Journal of Amer. Statist. Assoc. 55, 698707. Kaplan, E.L., and Meier, P. (19558). Nonparametric estimation from incomplete observations. American Statistical Assoc. Journal, June 1958, 457481 Kiefer, J. (1961). On large deviations of the empirical d.£ of vector chance variables and a law of the iterated logarithm. Paciﬁc Journal of Math., 11 (1961), 649660. Korwar, R.M., and Dahiya, R.C (1982). Estimation of bivariate distribution function from incomplete observations. Commun. Statist. Theor. Meth., 11(8), 887897. Rosemblatt, M. (1956). "Remarks on some nonparametric estimates of a density function". Ann. Math. Statist., 27, 832837. Tsai, W., Leurgans, S., and Crowley, J. (1986). Nonparametric estimation in the presence of censoring. The Annals of Statistics, Vol. 14, No. 4, 13511365 Wegman, E.J. (1972). Nonparametric density estimation. J. Statist. Comput. Simul., Vol. 1, 225245.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 211

DETERMINATION OF OZONE OPTICAL DEPTHS FROM SHADOWBAND RADIOMETER DATA

Stephen D. Hawks and Stephen Schiller South Dakota State University Brookings, SD 57007

Ozone has been a major concern for the past couple of decades. Benefi- cial ozone levels, necessary for blocking harmful ultraviolet radiation, are decreasing in the stratosphere, while polluting ozone levels in the troposphere are increasing. It is essential that we develop methods of detecting and monitoring these ozone levels. The following procedure is one such method. The purpose of this project was to determine ozone concentrations in the atmosphere from optical depth measurements using a sun photometer. The data used for this project were collected by a sevenchannel rotating shadowband multifilter radiometer (NWR7) manufactured by Yankee Environmen- tal Systems. The seven channels consisted of a broadband channel (350 to 1000nm) and six narrowband channels (10 nm). The six narrowband channels have nominal wavelengths of 415, 500, 615, 673, 870, and 940 nm. The 415 and 870 wavelengths are sensitive to the absorption effects of aerosols. The ozone sensitive wavelengths are 500, 615, and 673 nm. The 940 nm wavelength is sensitive to water vapor. The shadowband unit measured the total downwelling flux on a horizontal surface with the shadowband retracted and the hemispherical diffuse sky flux on a horizontal surface with the shadowband blocking the sun. See Figure 1. Two addition measurements are made to correct for excess sky blocked when the shadowband blocks the sun. The difference gives the solar contribution and provides direct solar flux when corrected for the zenith angle of the sun. The detectors in the shadowband unit view a horizontal diffuser that is il- luminated by the sun and sky. The transmission of light through the diffuser must be Lambertian to correctly measure the solar flux at different zenith an- gles. A Lambertian surface receives light with a perfect cosine response. This means that light received is proportional to the angle of incidence. Since no surface has a perfect cosine response, corrections need to be made to the irradiance data which is the density of radiation incident on the diffuser. The corrections can be made if the angular response of the surface is known compared to an ideal cosine response. A program called DOSBAND is used to perform such angular corrections. After the data are angle corrected, DOSBAND calculates direct solar flux by subtracting the diffuse sky flux from the total irradiance. The program to analyze the observational data was written in the pro- gramming language SPLUS and called Sband.testl. First, Sband.testl performs a Langley analysis on the direct solar flux data. The Langley analysis is an application of Beer's extinction law which is represented by

τ I=IO exp(- m). 212 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 1. Diagram of Shadowband Unit (Yankee Environmental Systems, Inc., 1995) at wavelengths that are not effected by water vapor. In the equation, I is the direct solar irradiance, IO is the irradiance that would be measured at the top of the atmosphere, τ is the total column optical depth, and m is the air mass relative to the unit air mass in the zenith direction (King and Byrne, 1976). One equation to calculate air mass is

air mass=[sin(el)+0.50572(6.07995+el)-1.6364]-1 where “el” is the elevation of the sun in degrees (Kasten and Young, 1989). Us- ing geometry, the cosine and the arccosine of the angle of incidence can be substituted for the sin(el) and the angle el, respectively. The above substitution is made since the shadowband unit also records the time of observation and uses this to calculate the cosine of the solar zenith angle. The following equation is obtained when the natural logarithm is taken of both sides of the extinction equation:

τ ln(I) = ln(IO) - m.

A linear plot relation is created between ln(l) and m. See Figure 2. This is the basis for Langley plot analysis where ln(I) is plotted against m. The slope is the τ magnitude of the optical depth and the y-intercept is the value of ln(IO). The plot holds true only for clear sky conditions. The total column optical depth found through Langley analysis is the sum of various extinction components. These include Rayleigh scattering, ozone absorption, and aerosol extinction (King and Byrne, 1976). This can be represented by Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 213

Figure 2. Langley Plot for the Morning of June 2, 1996.

τ λ τ λ +τ λ +τ λ ( )= rayleigh( ) aerosol( ) ozone( )

At the 415 and 870 nn wavelengths only the Rayleigh scattering and aerosol extinction are detected. Sband.testl separates the extinction components using the following procedure. First, the Rayleigh scattering is calculated. This is done by using

τ λ-4 λ-2 λ-4 rayleigh=0.008569 (1+0.0113 +0.00013 )P/P0 where the wavelength λ is in micrometers, and P is the site pressure relative to sea level pressure, P0 (Hansen and Travis, 1974). Rayleigh scattering is then subtracted from the total optical depth. Now the aerosol optical depths are known for 415 and 870 run wavelengths. Angstrom's turbidity formula is used to ﬁnd a rough estimate of the aerosol optical depth at the ozone sensitive wavelengths. The formula is given by

τ λ βλ−α aerosol( )=

τ λ where aerosol( ) is an estimate of the aerosol optical depth since it includes the effects of scattering and absorption by aerosols, β is called Angstrom's turbidity coefﬁcient, α is the wavelength exponent, and the wavelength λ is in micrometers (Iqbal, 1983). 214 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Taking the natural log of both sides of turbidity formula, one obtains

τ λ β α λ ln( aerosol( ))=ln( )- ln( ).

On a loglog plot, the aerosol data are expected to be nearly a straight line. The wavelength exponent and turbidity coefﬁcient are both constants. At 415 and 870 nn, the aerosol optical depth is not contaminated by ozone. Therefore, these values can be used to calculate the constants α and β. The values of the wavelength exponent and turbidity coefﬁcient are calculated by

α τ τ λ λ =-[ln( 415)-ln( 870)]/[ln( 415)-ln( 870)]

β τ λ−α = aerosol/

Once α and β are known, the aerosol estimates for ozone sensitive wave- τ λ lengths can be found. Subtracting aerosol( )from the remaining optical depth gives an estimate for the ozone optical depth. This value is then used to ﬁnd an estimate for the amount of ozone in cm(STP). The ozone concentration estimate can be determined by solving the following equation for η

τ η λ ozone= a( )

τ η where ozone is ozone optical depth, is the total ozone content in atm-cm, and (λ) is the ozone absorption coefficient per centimeter of pure gas at STP (King and Byrne, 1976). Since the ozone amount found here is just a rough estimate, Sband.testl performs a correction procedure. To do this, the program increases the amount of ozone by 0.01 cm and calculates the ozone optical depth. The program then subtracts the calculated value from the original optical depth to find the new aerosol optical depth. The preceding procedure is repeated for one hundred steps above and below the first estimate for the ozone amount. The program checks to find which set of aerosol optical depths causes the minimum root mean square error in the slope of aerosol loglog plot. See Figure 3. The ozone amount that corresponds to this aerosol data set represents an estimate of the true ozone amount in cm. The above procedure was applied to data collected by four Shadowband units at Lunar Lake, Nevada, on June 2, 1996. Sband.testl calculated ozone amount of 0.230, 0.293, 0.310, and 0.259 cm. A remote sensing group from the University of Arizona found 0.239 cm of ozone independently, from their own observations recorded at the same site and time (Thome, 1996).

BIBLIOGRAPHY

Hansen and Travis. 1974. Space Science Reviews. 16:527610. Iqbal, Muhammad. 1983. An Introduction to Solar Radiation. Academic Press Canada, Ontario. Kasten and Young. 1989. Applied Optics. 28:47354738. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 215

Figure 3. Aerosol Optical Depth with Adjusted Ozone Removed. The curve is the minimum root mean square ﬁt of Angstrom’s turbidity formula to the ln (optical depth) at ﬁve wavelengths (solid squares).

King, Michael D., and Byrne, Dale M. 1976. A Method for Inferring Total Ozone Content from the Spectral Variation of Total Optical Depth Obtained with a Solar Radiometer. Journal of the Atmospheric Sciences. 33:22422251. Thume, Kurtis J. 1996. Vicarious Calibration Comparison Results. Optical Sci- ence Center, University of Arizona, Tucson, Arizona. (preprint). Yankee Environmental Systems, Inc. 1995. MFR7 Rotating Shadowband Ra- diometer Installation and User Guide. Yankee Environmental Systems, Inc., Turners Falls, MA.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 217

STRENGTHENING AND DEFORMATION MECHANISMS IN THE FIRST GENERATION OF AL-LI ALLOYS 2091 AND 8090

Dajun Chen and Glen A. Stone Department of Metallurgical Engineering, South Dakota School of Mines and Technology, Rapid City, South Dakota 57701

INTRODUCTION

Using lightweight materials would give great benefit to the aircraft industry, both in efficiency and performance. Aluminum alloys presently comprise as much as 80% of the weight of a commercial airliner. For this reason, reducing the density of aluminum alloys used in aircraft construction is now generally accepted as the most effective way to reduce the weight of an aircraft. It is wellknown that addition of lithium (the lightest metallic element) to aluminum significantly increases the ambient temperature elastic modulus while simultaneously reducing the density (1-4). Previous research indicates that each weight percent of Li added to an Al alloy reduces the density approximately by 3% and increases the elastic modulus approximately by 6% (5). As a result, Li- containing Al alloys have attracted great interest from the aerospace industry. Al-Li alloys were originally designed to be a direct weight-saving substitute for conventional Al alloys and to provide about 10% density savings. The aerospace industry considers this number as a worth target. Unfortunately, the first generation of Al-Li alloys such as 8090, 2090, and 2091 have failed to gain widespread acceptance in the marketplace because adding lithium gives rise to a variety of problems that makes them less attractive than conventional Al alloys. One major problem of performance is that these alloys have lower elongation and lower toughness than incumbent alloys. The main purpose of this paper is to investigate the strengthening and deformation mechanisms in the first generation Al-Li alloys 2091 and 8090. In addition to this work, a parallel study on a second generation of Al-Li alloy 2097 is nearly completed. Such investigations are significant because the scientific understanding generated could provide a direction for future alloy development.

MATERIALS AND EXPERIMENTAL

The alloys 2091 and 8090 used in this research were supplied by the Alu- minum Company of America (ALCOA) Technical Center in the form of recrystallized and unrecrystallized rolled plates. The chemical compositions of the alloys are listed in Table I. All samples were cut into 0.20 mm thick sheets from the rolled plates by using a low speed saw. Samples were then mechanically ground to about 0.03 mm thickness by carefully using silicon carbide polishing paper. Sample and 218 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) polishing paper were continuously bathed in water during thinning. Three mm disks were punched after this procedure to prepare thin foils for TEM examinations. TEM specimens were prepared by using a Fischione Ion Mill (Model 3000). During this procedure the specimen temperature was controlled below 30º C. All TEM examinations were conducted using a Hitach MODEL H-7000 (FA) electron microscope at 100 kv accelerating voltage.

Table I. Chemical composition of the research alloys. Units are weight percent. (U—unrecrystallized, R—recrystallized)

Li Cu Mg Zn Fe Si Ti Zr Al

2091 U 1.78 1.94 1.44 0.02 0.1 0.04 0.023 0.11 bal 2091 R 1.88 2.11 1.54 0.02 0.06 0.04 0.033 0.09 bal 8090 U 2.23 1.08 0.64 0.02 0.04 0.04 0.014 0.02 bal 8090 R 2.55 1.08 0.68 0.02 0.03 0.04 0.011 0.04 bal

RESULTS AND DISCUSSION

Transmission electron microscopy study shows that the microstructures of δ both alloys consist of uniformly distributed ’ (Al3Li) precipitates, small amount β of plate-like S’ (Al2CuMg) and ’ (Al3Zr) precipitates. Although the main Li-con- δ taining precipitate ’ (Al3Li) cannot be seen in the bright field image, it can be easily identified in the selected area diffraction pattern by recognizing the su- δ perlattice spots (Fig. 1). This is because that ’ is an ordered phase with L12 δ structure (6,7). Figure 1 is a centered dark field image that shows the ’ (Al3Li) precipitates in the recrystallized sample of 2091 alloy. These precipitates are homogeneously coherent with the matrix and ranged in size from 25 to 50 nm. β Some ’ (Al3Zr) precipitates can also be identified in this centered dark field β image. The arrow in Figure 1 points to a typical ’ (Al3Zr) precipitate. Figure 2 is a centered dark field image of 8090 recrystallized sample. Us- ing a superlattice diffraction spot, δ’ precipitates also appear in the image. Compared with Figure 1, the δ’ particles in 8090 have higher density and smaller size (about 15~30 nm). This is consistent with the fact that the Li content is higher in 8090 than in 2091 and indicates that the volume fraction of δ’ precipitates increases with the increasing of Li concentration. In Figure 3, a deformation band can be found along the [110] direction of the matrix. It is clear from this image that δ’ precipitates were sheared in this deformation band. This leads to a fragmentation of the δ’ precipitates in this band. Obviously, this phenomenon is the result of a “cutting mechanism” by slip motion of dislocations. Figures 4 and 5 provide two additional examples that dislocations cut through the δ’ precipitates. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 219

Figure 1. Centered dark ﬁeld image of Figure 2. Centered dark ﬁeld image of alloy 2091 (recrystallized ) on a ( 0 1 alloy 8090 (recrystallized ) on a ( 1 0 1 ) spot (B =[ 0 1 -1] ). The majority 0 ) spot ( B =[ 0 1 1 ] ). Higher den- δ δ of precipitates are ’ (Al3Li). The ar- sity of ’ (Al3Li) precipitates are visi- β row points to a ’ (Al3Zr) precipitate. ble.

Figure 3. Centered dark ﬁeld image of alloy 8090 (recrystallized ) on a ( 1 1 Figure 4. Centered dark ﬁeld image of 0 ) spot ( B =[ -1 1 2 ] ). A deforma- alloy 2091 (recrystallized ) on a ( 1 0 tion band is clearly visible in the [ 1 1 0 ) spot ( B =[ 0 1 1 ] ). Particles 0 ] direction. sheared by dislocation motion are observed in this image. 220 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

It is well-known that when dislocations cut through particles, there are several factors that must be considered which will increase the critical resolved shear stress: 1) the energy required to create an additional particle/matrix interface, 2) straining of the lattice caused by coherency between the matrix and the particles due to difference in lattice spacing, 3) changes in basic properties of dislocation such as the line-tension and self- energy due to the difference in Figure 5. Centered dark field image of elastic shear modulus between the alloy 2091 (recrystallized ) on a ( 1 0 matrix and particles, and 4) the en- 0 ) spot ( B =[ 0 1 1 ] ). Particles ergy required to produce an anti- sheared by dislocation motion are ob- phase boundary (APB) at the slip served in this image plane in the particle if the particle is ordered. Of these four factors, the last two may dominate. As mentioned above, the addition of Li to Al significantly increases the modulus, so δ’ precipitates should have higher modulus than the matrix. This means the dislocation line-tension and self-energy must increase when it cuts through the δ’ precipitates. On the other hand, the energy of anti-phase boundary (APB) formed after dislocation cutting cannot be negligible. This is because the energy of APB is usually about 10 times more than precipitate/matrix interface energy of ordered coherent precipitates (8). These factors will lead to the strengthening of the alloy. From the above results and discussion, it is clear that the δ’ precipitates are the main strengthening phase in both Al-Li alloys researched. The deformation model in these alloys is mainly the dislocation cutting mechanism. The deformation process is the continuous fragmentation of δ’ precipitates. Under the dislocation cutting mechanism, the size of δ’ precipitates becomes smaller and smaller with the increasing of deformation. This is verified by the results obtained from the unrecrystallized samples of both research alloys. Figure 6 is a centered dark field image showing the δ’ precipitates in an unrecrystallized sample of alloy 8090 by imaging a superlattice reflection. The size of δ’ precipitates in this image is much smaller (ranging from 5 to 10 nm) than that observed in the recrystallized sample. This is the result of “cutting” by a number of dislocations and verifies that the deformation model is mainly the dislocation cutting mechanism. A noticeable feature in the unrecrystallized sample is that there exist some very large δ’ precipitates (size about 40-80 nm) (Fig. 6) which were not found in the microstructures of recrystallized samples. These large δ’ precipitates still have the ordered structure, but unlike smaller particles, they have distinct edges with the matrix. This means that these large δ’ precipitates have already Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 221 lost the coherency with the matrix. As a result, they are also visible in the bright field image (Fig. 7). The formation of these large δ’ precipitates may be directly associated with the deformation model. As verified above, the deformation model in the alloys is likely the dislocation cutting mechanism and the size of δ’ precipitates will become smaller and smaller with increasing deformation. It is reason- able to speculate that when the δ’ precipitates are cut to a certain Figure 6. Centered dark field image of size, such as their critical size for alloy 8090 (unrecrystallized ) on a ( 0 1 nucleation, some types of precipi- 1 ) spot ( B =[ 0 1 -1 ] ). Some very large tate reversion will take place in the δ’ particles are visible in this image. deformed alloy. In other words, these small particles will collect again to form the large particles. The driving force of this process will come from the high particle/matrix interface energy caused by particle fragmentation. Alternatively, the high density of δ’ precipitates caused by high concentration of lithium in the alloys may increase the tendency of precipitate reversion. Because these large δ’ precipitates have large size and have lost coherency with the matrix, the dislocation cannot cut through them. The only way that dislocations can pass these large particles is to make loops around them, as shown in Figures 8 and 9. These phenomena indicate that accompanying the

Figure 7. Alloy 2091 (unrecrystallized) (a) bright field image (B =[ 0 0 1] ), (b) centered field image on a (1 0 0 ) spot. Large δ’ particles are visible in both bright and centered dark field images. 222 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Figure 8. Bright ﬁeld image of alloy Figure 9. Weak beam image of alloy 2091 (unrecrystallized ) ( B =[ 0 0 1 ] 8090 (unrecrystallized ) ( g / 3 g and ). The dislocation makes an arched B =[ 0 1 1 ] ). A dislocation tries to shape between two large δ’ particles. make a loop around the large d’ particles. formation of these large δ’ precipitates, the deformation model in both alloys has been changed from the dislocation "cutting" mechanism to a dislocation “looping” mechanism. Under the dislocation “looping” mechanism, each dislocation passing the particle will leave one loop around the particle. It can be anticipated that as many dislocations pass the large δ’ precipitates by the “looping” mechanism, a dislocation pile-up will be formed in the front of the particle. This is shown in Figure 10. The dislocation pile up will cause a serious stress concentration in the front of the particle. In a conventional treatment, this stress concentration can be ex- τ τ τ pressed as = n o , where o is the applied shear stress on the slip plane, n is the number of dislocation in the dislocation pile-up. When this stress concentration increases and reaches to a certain value, the following has been sug- Figure 10. Bright ﬁeld image of alloy gested: 1) Stop the dislocation 2091 (unrecrystallized ) ( B =[ 0 0 1 ] source so that it cannot generate ). The dislocation pile-ups were new dislocations, or 2) Break the formed in the front of the large δ’ par- particle, which will cause micro- ticles. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 223 cracks at the site of the particle. Both of these could be responsible for the low ductility and low fracture toughness in the alloys.

CONCLUSION

In the ﬁrst generation of Al-Li alloys 2091 and 8090, the δ’ phase is the main strengthening phase. The volume fraction of δ’ precipitates is increased with the increasing Li concentration. The deformation model in these alloys is mainly the dislocation “cutting” mechanism, which leads to a continuous fragmentation of δ’ precipitates during deformation. Some very large δ’ particles which had lost coherency with the matrix were found in the unrecrystallized alloys. It is believed that their formation is directly related with the deformation model and high concentration of lithium in the alloys. Accompanying the formation of these large δ’ particles, the deformation model changes from “cutting” to “looping.” The latter will cause a serious stress concentration in the front of these large δ’ particles, and could be responsible for the low ductility and low fracture toughness observed in the ﬁrst generation of Al-Li alloys.

ACKNOWLEDGMENT

The authors wish to thank Dr. R. J. Rioja (Alcoa Technical Center) for supplying the materials and for his guidance and suggestion during this research. Thanks are also addressed to South Dakota NSF/EPSCoR program for the ﬁ- nancial support for this research. The use of the Electron Microscopy Facility within the Engineering and Mining Experiment Station at SDSM&T is gratefully acknowledged.

REFERENCES

I. M. LeBaron, USA Patent No. 2381 219 (1945). E. J. Lavernia & N. J. Grant, J. Mater. Sci., 23, 1521 (1987). A. K. Vasudevan,R. C. Malcolm, W. G. Fricke & R. J. Rioja,Final Report, Naval Air Systems Command, Contrac No. N00019-80-C-0569, June 30 (1985). E. A. Starke, Jr., T. H. Sander, Jr. & I. G. Palmer, J. Met., 33, 24 (1981). K. K. Sankaran & N. J. Grant, Aluminum-Lithium Alloys, p. 205, TMS-AIMS, Warrendale, PA (1981). D. B. Williams & J. W. Eddington, Acta. Metall., 24, 323 (1976). J. D. Kim & J. K. Park, Metall. Trans., 24A, 2613 (1993). J. D. Verhoeven, Fundamentals of Physical Metallurgy, p. 408, by John Wiley & Sons, Inc., New York (1975).

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 225

ILLUMINANCE FROM A POINT SOURCE IS AN OSCILLATING FUNCTION OF DISTANCE IN A HYPERSPHERICAL UNIVERSE

by Richard P. Menzel

ABSTRACT

The universe is to be those ordered quadruples of real numbers (x,y,z,w) (“points”) in a 4-dimensional Euclidean space which satisfy the equation of the hypersphere x2 + y2 + z2 + w2 = 1, where we use the radius of the universe (called “ur”) as the unit of length measurements. It is shown that the illuminance of the surface of a sphere contained in this universe, by a point light source at its center, is proportional to cosecant2 (s radians), where s is the numerical value of the length of the radius of the sphere (expressed in ur units). A conclusion: a luminous body near a point in the universe antipodal to Earth may appear just as bright to an observer on Earth as if it were near Earth.

THE ARGUMENT

(Misner, Thorne & Wheeler, 1973), pages 723-724, discuss a “hyperspherical” universe in which the points are (only) those ordered quadruples of real numbers (x,y,z,w) in a 4-dimensional Euclidean space which satisfy the equation x2 + y2 +z2 + w2 = a2, and we let a = 1. The appendix discusses 4-dimensional Euclidean space, the universe and the surface area of a sphere contained in the universe (which is 4π sine2 (s radians) ur2, where s is the numerical value of the length of the radius of the sphere). The reader may prefer to read the appendix next. We will derive the illuminance law, using this formula for surface area. Assume light paths in an empty region of space are contained in great circles. The amount of energy per second passing through the surface of a sphere of radius s ur, from a point source at the center of the sphere, equals the amount of energy passing through a unit area of the sphere per second (i.e. the illuminance) times the area of the sphere. So energy per second equals illuminance times 4πsine2 (s radians) ur2 and illuminance is proportional to 1 ÷ sine2 (s radians) = cosecant2 (s radians). From a graph of y = csc2 (s radians) we see that a luminous body reced- ing from Earth seems (from Earth) to become progressively dimmer until s (the numerical value of its distance from Earth) is π/2, at which point it is 1/4 of the distance around the universe: then it becomes brighter until it approaches the antipodes (s = π), at which time its illuminance approaches an infinite amount. (It approaches infinity not because the luminous body is emitting an infinite amount of light, but because the surface area of the sphere around the luminous body, 4π sine2 (s radians), is approaching zero ur2. When s = π, all light 226 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) emanating from the luminous body will be concentrated on the point antipodal to it. The identities sine2 (s radians) = sine2 ((π ± s) radians) = sine2 ((2π - s) radians) imply the illuminance on Earth from a star at distance s ur from Earth is the same as that from a star of equal intensity at distance s ur from the antipodes, or from a star at distance 2π - s from Earth. A star would appear equally bright (and might have different red shifts) if viewed from opposite directions (viewed from Earth). (The star might move during the difference in trav- elling time for the two light rays).

SPECULATIONS

Find a cepheid variable near the antipodes. It should appear as bright as if it were equally near Earth and as large (subtending the same angle with great-circle sides) and it might have detectable lateral motion. (One can cross many time zones in a 5 minute walk near the North Pole.) Are some “meteors” near the antipodes? Were the dinosaurs cooked by a radiant body passing through the antipodes? What would gravitational waves from a supernova do to the antipodal region?

APPENDIX

In our analytic 4-dimensional Euclidean geometry we call the set of all quadruples of real numbers (x,y,z,w) “points”, we define the distance between points (x,y,z,w) and (a,b,c,d) to be [(x-a)2 + (y-b)2 + (z-c)2 + (w-d)2]1/2, we define a “vector” with “initial point” (a,b,c,d) and “terminal point” (x,y,z,w) to be (x-a,y-b,z-c,w-d), we define two vectors (A,B,C,D) and (E,F,G,H) to be “perpendicular” if their “scalar product” AE+BF+CG+DH = 0, we define the “sum” of these vectors as (A+E,B+F,C+G,D+H) and we define the product of number k and vector (A,B,C,D) as (kA,kB,kC,kD). Thus a vector (x,y,z,w) may be written as a linear sum of mutually perpendicular vectors (1,0,0,0), (0,1,0,0), (0,0,1,0), (0,0,0,1) with the equation (x,y,z,w) = x(1,0,0,0) + y((0,1,0,0) + z(0,0,1,0) + w(0,0,0,1). For the interested reader, (Sommerville, 1958) discusses 4-dimensional Euclidean geometry analytically and (Manning, 1914) discusses it synthetically, as does a chapter of (Wylie,1964). As we said, the universe is to consist only of the points (x,y,z,w) which satisfy x2 + y2 + z2 + w2 = 12. If paths are restricted to this universe, it has been shown that a shortest path between two points in the universe lies in a “great circle” ie. the circle in which the universe intersects the plane which passes through the origin (0,0,0,0) and the two points. (Reducing the dimension by one, recall that one passes Labrador when flying from New York City to Lon- don: the shortest path on the surface of the Earth lies in a great circle.) We assume that light in a nearly empty region of our universe follows a path which is in a great circle. This universe may be parameterized with the equations x = sine χ sine θ cos φ , z = sine χ cos θ, w = cos χ, y = sine χ sine θ sine φ , because substitution of these expressions into x2 + y2 + z2 + w2 and the use of trigonometric identities gives 1. The universe is swept out with 0 ≤χ≤π, Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 227

0≤θ≤πand 0 ≤φ ≤2π. Taking the intersection of the universe with the plane which contains point P and the positive w axis (see diagram), we may interpret χ to be the angle between the segments ON and OP (from the equation w = cos χ), and the arc of the great circle connecting N and P has a length whose numerical value s equals the numerical value of χ, if χ is in radians. We derive the formula for surface area of a sphere. Just as a circle is the set of all points in a plane which are equidistant from a ﬁxed point (the center), so a sphere in 4-dimensional space is the set of all points in “hyperplane Ax + By + Cz + Dw +E = 0” which are equidistant from a center.

We ﬁrst show the intersection of hyperplane w = w0 and our “hypersphere” universe is a sphere, if the intersection contains more than one point. From the diagram, if (x,y,z,w0) is distance 1 ur from (0,0,0,0) (ie. is in the universe), then 2 1/2 it is constant distance (1 - w0 ) ur from “center” (0,0,0,w0) (ie. is on a sphere 2 1/2 of this radius). Conversely if (x,y,z,w0) is (1 - w0 ) ur from (0,0,0,w0) (ie. if it is on the sphere), then it is 1 ur from (0,0,0,0) (ie. it is in the universe). See the diagram. The points in the universe whose arc-distances from the

North Pole (0,0,0,1) have numerical value s0 lie in the hyperplane w = w0 = cos(s0 radians). From preceding argument, the intersection of this hyperplane 2 1/2 ⏐ and the universe is the sphere with radius (1 - cos (s0 radians)) ur = sine(s0 ⏐ π 2 radians) ur and center (0,0,0,w0). With this radius, the surface area is 4 sine 2 2 2 2 2 (s0 radians) ur . The equations of the sphere are x + y + z = sine (s0 radians) and w = cos (s0 radians). One may also compute the surface area of this sphere in the universe with χ center (0,0,0,1) and radius s0 ur with calculus. Since s0 = 0 substitute s0 into the parametric equations of the universe to get the surface in terms of two param- θ φ θ φ θ eters: ie. x = sine s0 sine cos , y = sine s0 sine sine , z = sine s0 cos , (with 0 ≤θ≤πand 0 ≤φ≤2π). Integrate.

ACKNOWLEDGEMENT

The author is grateful to Professor Donald Abraham of USD for encouragement, for suggestions and for reading the paper. 228 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

BIBLIOGRAPHY

Manning, H.P., 1914. Geometry of Four Dimensions. Dover Publications, New York City. Misner, C.W., Thorne, K.S. & Wheeler, J.A., 1973. Gravitation. W.H. Freeman & Co., San Francisco. Sommerville, D.M.Y., 1958. An Introduction to the Geometry of N Dimensions. Dover Publications, New York City. Wylie, C.R.Jr., 1964. Foundations of Geometry. McGraw-Hill, New York City, San Francisco, Toronto, London. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 229

SOURCES OF NATURAL RESOURCE INFORMATION, NON-CONSUMPTIVE USE OF WILDLIFE, AND CHANGES IN SOUTH DAKOTA RESIDENTS' ATTITUDES TOWARD HUNTING BETWEEN 1973 AND 1989

Nancy J. Dietz Wildlife and Fisheries Sciences Department

Kenneth F. Higgins National Biological Service, South Dakota Cooperative Fish and Wildlife Research Unit

Robert D. Mendelsohn Rural Sociology Department

South Dakota State University Brookings, SD 57007

ABSTRACT

Trends in South Dakotan’s attitudes toward hunting and their participation in hunting were evaluated from responses to mail questionnaire surveys conducted in 1973 and 1989. Completed questionnaires were returned by 66% of 2,841 drivers license holders, ages 16 years and older. Seventy-eight percent of the respondents had hunted at sometime during their life. During the 16- year interval between surveys, changes in responses to three stimuli statements suggest an increase in South Dakotans’ positive attitudes about hunting in their state. Ninety-nine percent of respondents enjoyed seeing or watching wildlife, but only 53% of respondents were willing to pay a fee of $2 or more to view wildlife on public lands. Fifty-six percent of respondents would be willing to pay a special tax on non-hunting equipment to help fund non-consumptive wildlife programs. About 54% of respondents did not receive any conservation, hunting, or wildlife magazines nor watch wildlife-oriented programs on television. Determination of trends in residents’ attitudes toward, and their participation in, hunting and other wildlife related recreation is necessary for a wildlife management agency to be responsive to all segments of the public. Knowl- edge of residents’ attitudes toward hunting, and the proper use of this knowledge enhances an agency’s quality of public service and aids in development and administration of programs to promote and regulate hunting recreation. In 1973, Rosonke et al. (1975a and b) conducted a survey of South Dakota residents to determine their attitudes towards hunting, hunters, and wildlife management ofﬁcials. Their results provided baseline data that can be used to as- 230 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) sess future trends in South Dakotans’ attitudes toward hunting. Such information was wanted by the South Dakota Game, Fish and Parks Department to help them make informed management decisions. Using data from Rosonke et al. (1975a and b) for comparison, we conducted this study to determine (1) if any changes had occurred in South Dakotans’ attitudes toward hunting during a 16-year period, from 1973 to 1989, (2) where citizens obtain natural resource information, and (3) kinds of non-consumptive uses of wildlife.

METHODS

A random sample of South Dakota residents was surveyed with a self-administered, mail-back questionnaire in September 1989 (Dietz, 1990). The random sample of 3,228 residents, ages 16 or older, was obtained from the South Dakota Drivers License Bureau’s 1988 population of 523,000 drivers license holders. A total of 387 people were removed from the initial sample due to incorrect addresses, death, poor health, or because they had moved from South Dakota, resulting in a deliverable sample of 2,841 residents. To assess changes in South Dakotans’ general attitudes toward hunting, we used the same stimulus statements and response formats used by Rosonke et al. (1975a). The stimulus statements were: “Hunting helps to preserve the balance of nature,” “There should be more restrictions on hunting,” and “All hunting should be banned in South Dakota” (Rosonke et al., 1975a). Responses were grouped by disagree (Likert scale 1 to 3), undecided (Likert scale 4), and agree (Likert scale 5 to 7). Drivers license holders were surveyed using standard mail survey procedures (Dillman, 1978), except for one modification. Dillman (1978) recommended an initial and three follow-up mailings. We substituted an introductory letter for the third follow-up mailing (Dietz, 1990). Linsky (1975) reported that the use of an introductory letter increased response rates in all 12 studies he examined. The self-administered, mail-back questionnaire asked residents about attitudes towards hunting and hunters, their use of the wildlife resource, hunter behavior witnessed, and wildlife information resources. A 10-percent sample of computer-coded responses was verified; no further verification was necessary because of a < 1.0% coding error. Because of the high response rate and because nonresponse bias is not considered a problem in studies of attitude (Manfredo et al., 1989), we did not estimate nonresponse bias. Chi square analyses (SAS Institute Inc., 1985) were used to examine the significance of associations between the respondents’ attitudes toward wildlife and hunters. A relation was considered statistically significant if P was ≤ 0.05.

RESULTS

Completed questionnaires were returned by 65.6% (n=1,865) of 2,841 drivers license holders, whose ages were ≥ 16 years. Nearly 73% of the respondents (n=1,849) resided east of the Missouri River. According to the 1990 census, 67% of all South Dakotans lived east of the Missouri River, indicating that our sample was representative of the population. The distribution of South Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 231

Dakota residence respondents (n=1,748) was 25.1% from Sioux Falls or Rapid City, 29.1% from cities with populations of 3,000 to 30,000, 24.0% from cities with populations of 2,999 or fewer, and 21.6% from farms or ranches. Re- spondents (n=1,779) were 42.4% female and 57.6% male. Forty percent of the respondents (n=1,728) had resided on a farm or ranch when they were between the ages of seven and 16, whereas only 5.7% had resided in cities of 100,000 or more at the same age. Caucasians comprised 95.9% (n=1,793), and 78.4% (n=1,784) were natives of South Dakota. Almost 28% (n=1,796) of the respondents were 18 to 34 years old, 35.7% were 35 to 54 years old, and 14.6% were 55 to 64 years old. Seventy-eight percent (n=1,852) of the respondents had hunted at sometime during their life. Forty percent (n=1,849) of the respondents who had hunted during their lifetime currently hunt and had hunted every year from 1987 to 1989. Approximately 95% (n=1,405) of respondents had friends who hunt and 75.8% had immediate family members who hunt (n=1,855). About 54% (n=749) of the female respondents and 95.8% (n=1,018) of the male respondents had hunted during their lifetime.

South Dakotans’ Attitudes Toward Hunting

In 1973 and in 1989, most respondents agreed with the statement “Hunt- ing helps to preserve the balance of nature in South Dakota” (Table 1). Ex- amination of the polar positions, i.e., those agreeing versus disagreeing with the stimulus statement, suggested an increase in a positive resident perspective on hunting and a decrease in negative resident perspectives on this stimulus. A positive perspective means respondent attitudes were pro hunting even though they might not hunt. A significant increase in positive evaluations and a decrease in negative evaluations occurred between the 1973 and 1989 time periods was statistically significant (X2=3.79, d.f.=1, P<0.05). Significant differences in the extent of agreement or disagreement with the above statement existed by gender (X2=29.6, d.f.=6, P<0.05), age (X2=30.8, d.f.=18, P<0.05), where the respondent resided between the ages of 7 and 16 (X2=51.8, d.f.=18, P<0.05), and lifetime hunting participation (X2=69.9, d.f.=6, P<0.01). In 1973 (Table 1), a greater percentage of citizens agreed (49.2%) than disagreed (38.4%) with the statement “There should be more restriction on hunting in South Dakota.” By 1989, the proportion of agreements (32.9%) and dis- agreements (49.3%) with the same attitude stimulus statement were virtually re- versed with the proportion of “undecided” respondents also increasing; 12.4% and 17.8%, respectively (Table 1). An analysis of the polar attitude positions indicates there was a significant difference (X2=34.47, d.f.=1, P<0.001) in the opposition to increased restrictions on hunting and the reductions of support for more restrictions. Significant differences in response to the above statement were associated with gender (Table 2), age (X2=44.6, d.f.=18, P<0.05), where the respondent resided between the ages of 7 and 16 (X2=39.1, d.f.=18, P<0.05), and lifetime hunting participation (Table 3). In 1973, most respondents (93.2%) disagreed that “All hunting should be banned in South Dakota,” and only a minority agreed (3.6%) (Table 1). Com- 232 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 1. Frequency and percent responses of South Dakota's residents to three attitude stimulus statements, 1973a and 1989.

Responseb Stimulus Number of Statement Responses Disagree Undecided Agree 1973 1989 1973 1989 1973 1989 1973 1989

“Hunting helps 474 1,850 (14.8%) (11.5%) (5.5%) (5.8%) (79.7%) (82.7%) to preserve the balance of nature in South Dakota”

“There should be 474 1,848 (38.4%) (49.3%) (12.4%) (17.8%) (49.2%) (32.9%) more restrictions restrictions on hunting in South Dakota”

“All hunting should 474 1,852 (93.2%) (93.9%) (3.2%) (2.4%) (3.6%) (3.7%) should be banned in South Dakota” a Data for 1973 were obtained from Rosonke et al. (1975a). b Response is grouped by categories: disagree (Likert 1 and 2), undecided (Likert 3, 4, and 5) and agree (Likert 6 and 7).

Table 2. Percent responses by gender of South Dakota’s residents to the statement: “There should be more restrictions on hunting in South Dakota,” 1973a and 1989.

1973b 1989c Male Female Male Female (n=310) (n=164) (n=1,020) (n=745)

Strongly agree 4.8 7.3 5.4 6.9 Agree 24.2 23.2 10.7 14.9 Somewhat agree 17.2 23.2 14.3 14.1 Undecided 9.7 17.7 14.5 22.3 Somewhat disagree 11.0 9.8 13.2 11.1 Disagree 27.4 17.1 27.8 23.1 Strongly disagree 5.2 1.8 14.1 7.6 a Data for 1973 were obtained from Rosonke et al. (1975b). b X2 = 16.2732 d.f. = 6 P = 0.0124 c X2 = 43.240 d.f. = 6 P = 0.0001 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 233

Table 3. Percent responses by hunting participation of South Dakota’s residents to the statement: “There should be more restrictions on hunting in South Dakota,” 1989.

Hunted during Never lifetime hunted (n = 1,437) (n = 399)

Strongly agree 6.3 5.5 Agree 1.0 17.0 Somewhat agree 13.8 16.3 Undecided 15.1 27.6 Somewhat disagree 12.8 9.8 Disagree 27.4 19.8 Strongly disagree 13.6 4.0

X2 = 72.29 d.f. = 6 P = 0.0001 parable figures were obtained in 1989, when 93.9% of respondents (n=1,852) disagreed and 3.7% agreed that “All hunting should be banned in South Dako- ta” (Table 1). The changes in the percentages between 1973 and 1989 were not statistically significant (X2=0.004, d.f.=1, P>0.05). Significant differences by gender were found in response to the above statement referring that all hunting should be banned in South Dakota (Table 4) and lifetime hunting participation (X2=130.7, d.f.=6, P<0.05) (Table 5).

Non-consumptive Use of Wildlife in South Dakota

Ninety-nine percent of all respondents (n=1,861) enjoyed seeing and watching wildlife. About 70% of respondents (n=1,847) fed songbirds occasionally to frequently whereas about 30% never fed birds. Forty-three percent of respondents or their families took drives in the country at least once a month to observe wildlife (n=1,825). Nearly 32% took drives to observe wildlife two to 10 times a year and 14.7% took such drives only once or less a year. Dur- ing 1988 to 1989, 37.4% of respondents (n=1,798) indicated they had attempt- ed to photograph, paint, or draw wildlife. Close to half (47%) of the respondents (n=1,803) were unwilling to pay an annual fee for observing wildlife on public lands. Of the 53% willing to pay a fee, 13.8%, 18.9%, 12.5%, 2.2%, 3.2%, and 2.4% were willing to pay annually $2, $5, $10, $15, $20, and $25 or greater, respectively to observe wildlife. Near- ly 44% (n=1,814) of respondents did not favor adding a special tax to the sale of non-hunting equipment (e.g., bird feeders, binoculars, cameras, or ﬁlm) to provide funds for a nongame wildlife program in South Dakota. Of the 56% of citizens willing to pay a fee, 25.2%, 23.3%, 6.0%, 0.8%, and 0.6% (n=1,814) would favor a special sales tax of 0.5%, 1-2%, 3-5%, 6-10%, and 11% or greater, respectively on non-hunting equipment. Almost 16% of respondents (n=1,852) 234 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 4. Percent responses by gender of South Dakota’s residents to the statement: “All hunting should be banned in South Dakota,” 1973a and 1989.

1973b 1989c Male Female Male Female (n=310) (n=164) (n=1,019) (n=748)

Strongly agree 0.6 1.2 1.7 2.1 Agree 1.0 0.6 0.6 0.7 Somewhat agree 1.3 3.0 0.9 1.6 Undecided 1.6 6.1 2.1 2.8 Somewhat disagree 3.2 7.9 2.4 5.9 Disagree 52.3 55.5 19.1 33.7 Strongly disagree 40.0 25.6 73.2 53.2 a Data for 1973 were obtained from Rosonke et al. (1975b). b X2 = 20.5822 d.f. = 6 P = 0.0022 c X2 = 79.211 d.f. = 6 P = 0.0001

Table 5. Percent responses by lifetime hunting participation of South Dakota’s residents to the statement: “All hunting should be banned in South Dakota,” 1989. (n=1,839).

Hunted during Never Lifetime Hunted Response (n=1,435) (n=404)

Strongly Agree 1.6 2.7 Agree 0.5 0.7 Somewhat Agree 0.9 2.5 Undecided 1.7 5.2 Somewhat Disagree 2.9 8.7 Disagree 21.0 38.4 Strongly Disagree 71.4 41.8

X2 = 130.684 d.f. = 6 P = 0.0001 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 235 would be willing to donate money to help the South Dakota Department of Game, Fish and Parks purchase areas that would be used for wildlife observation and interpretation, whereas 47.1% were uncertain about such a donation.

Sources of Public Information about Wildlife in South Dakota

Approximately half (54%) of the respondents (n=1,814) did not receive any conservation, hunting, or wildlife magazines, whereas 35.4% received one or two such magazines. Among those respondents receiving conservation magazines (n=780), 50.1% read the South Dakota Department of Game, Fish and Parks’ South Dakota Conservation Digest, 71.2% read popular hunting magazines (e.g., Outdoor Life or Field and Stream), 33.8% read magazines associated with hunting or game conservation organizations (e.g., American Hunter or Pheasants Forever), and 38.3% read technical wildlife or conservation organization publications (e.g., Ducks Unlimited, Audubon). Outdoor magazines identified most important to respondents receiving these types of magazines (n=770) were: Outdoor Life (18.6%), National Geographic (18.4%), Field and Stream (14.0%), South Dakota Conservation Digest (7.3%), and Dakota Game and Fish (5.7%). Nearly 46% of respondents (n=1,728) watched televised conservation or wildlife programs fairly often or frequently. Only 4.7% of respondents (n=1,833) never watched conservation or wildlife programs on television; 49.2% occasionally watched these programs. The most watched programs were: Wild Kingdom (30.5%), National Geographic Specials (25.8%), Nature (12.1%), NOVA specials on animals (11.3%), and South Dakota Outdoor Guide (10.0%). If respondents (n=1,823) needed some facts about wildlife, they obtained information first from: the South Dakota Department of Game, Fish and Parks (SDGFP) (32.3%), magazines or books (27.9%), South Dakota conservation officers (11.8%), or United States Fish and Wildlife Service (USFWS) personnel (7.9%). If they (n=1,786) could not get the wildlife information at their first choice they would next go to: SDGFP (30.4%), USFWS (19.0%), magazines or books (13.4%), or a South Dakota conservation officer (13.3%).

DISCUSSION AND RECOMMENDATIONS

Our findings revealed that resident responses to three statements reflecting attitudes about hunting were more positive in 1989 than in 1973. In 1989, only 6.1% of the citizen respondents were in favor of, or undecided about, a complete ban of hunting in South Dakota. By comparison, 11% of the surveyed citizens in Iowa were opposed to hunting (Dahlgren et al., 1977). South Dako- ta respondents who were in favor of or undecided about a ban on hunting were primarily females or those who had never participated in hunting. Oth- er studies also consistently support the generlization that a higher proportion of individuals with anti-hunting attitudes were female or those who had never participated in hunting (Dahlgren et al., 1977; Kellert, 1978; 1976; Applegate, 1979; 1975; 1973; Kellert and Berry, 1987). The gender influence on attitudes 236 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) among South Dakotans was stronger in 1973 (Rosonke et al., 1975b) than in 1989 (Table 2). In fact, in 1989, twice as many women strongly disagreed with a ban on hunting compared with women in 1973 (Table 4). This increased commitment may be due to an increasing number of women receiving higher educations or working outside the home (Hoffman, 1977). Hoffman (1977) suggests that adult gender roles are converging and, therefore, attitudinal differences between sexes in future generations would decline. Attitudes toward hunting are also influenced by a person’s previous connection to rural and urban culture and their age. Shaw (1977) surveyed members of anti-hunter organizations and found that members came from predominately urban backgrounds. Slightly less than half of the landowners in New York who were reared in metropolitan areas were opposed to or had reservations about hunting (Brown, 1974). Shaw and Gilbert (1974) found that anti- hunting sentiment of college students was strongest among those from large population centers. Other studies (Kellert 1976, 1978) revealed that urban background had similar influences as the above studies. Our results showed that urban childhood residence affected responses to two of three attitude stimulus statements. Apparently, the influence of where a child is reared has similar effects on their attitudes in a less populous state (e.g., South Dakota) as would have been expected in a more populous state (e.g., New York). Peterle and Scott (1977) predicted that anti-hunting sentiments probably would become stronger as the percentage of hunters continues to decline in an urbanized society. In New Jersey, Applegate (1979) found that individuals who opposed deer hunting were mostly under 30 years of age. Age did not influence the general attitude of our sample respondents towards hunting. However, age was a factor in responses about hunting regulations and concepts, although the importance of age was not consistent. The majority of South Dakota respondents did not support more restrictions on hunting. These results indicate that in general current hunting regulations are perceived to be restrictive enough. Most South Dakotans support hunting as recreation and for management of game populations. Apparently, the current generation of South Dakotans has not been strongly influenced by the anti- hunting movement. In South Dakota, an individual’s decision to hunt does not seem to reflect national or regional social pressures because nearly 80% of the surveyed citizens had hunted during their lifetime and 86% were non-consumptive users of wildlife. Most South Dakotans enjoy wildlife regularly by either feeding birds or by viewing, photographing, or drawing wildlife. Residents also show their interest in wildlife by purchasing conservation, hunting, or wildlife magazines and by viewing wildlife on television programs. Given this high interest in wildlife among South Dakota residents, wildlife managers might want to consider these nonconsumptive uses of wildlife when developing and implement- ing management plans, so that all citizens can enjoy the bountiful wildlife resource of the state. Dietz (1990) indicated that young people in South Dakota are exhibiting a rural-urban difference in attitudes toward hunting, with urban youth having less favorable attitudes toward hunting. Most South Dakota adults come from Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 237 a rural background, even though a larger percentage of the state’s population now resides in urban areas. The recent population trend toward greater urbanization is projected to continue until the year 2000. In light of this trend, we recommend conducting another survey of South Dakotans’ attitudes toward hunting in the late 1990’s. Its objective would be to determine if the current higher urban residence of today’s youth will be an influence on their future attitudes.

ACKNOWLEDGMENTS

Financial support was provided by the National Riﬂe Association of Amer- ica Grants-In-Aid Program (GIA 90-11), Washington, D.C., and by the Federal Aid to Wildlife Restoration P-R Project W-75-R, Study No. 7540 to the South Dakota Cooperative Fish and Wildlife Research Unit, in cooperation with the South Dakota Department of Game, Fish and Parks, U.S. Fish and Wildlife Ser- vice, South Dakota State University, and the Wildlife Management Institute. W. Plucker, S. Clifford, W. Higgins, J. Ray, and K. Luttschwager assisted with survey preparation for mailing or data entry. G. Morris and L. Kronizal entered and veriﬁed data and assisted with all aspects of the project. B. Dirks, K. Jenkins, R. Johnson, S. Kohn, R. Stover, and R. Warhurst provided manuscript reviews.

LITERATURE CITED

Applegate, J. E. 1973. Some factors associated with attitude toward deer hunting in New Jersey residents. Trans. N. Am. Wildl. & Nat. Resour. Conf. 38:267-273. Applegate, J. E. 1975. Attitudes toward deer hunting in New Jersey: a second look. Wildl. Soc. Bull. 3:3-6. Applegate, J. E. 1979. Attitudes toward deer hunting in New Jersey: a decline in opposition. Wildl. Soc. Bull. 7:127-129. Brown, T. L. 1974. New York landowners’ attitudes toward recreation activities. Trans. N. Am. Wildl. & Nat. Resour. Conf. 39:173-179. Dahlgren, R. B., A. Wywialowski, T. A. Bubolz, and V. L. Wright. 1977. Inﬂu- ence of knowledge of wildlife management principles on behavior and attitudes toward resource issues. Trans. N. Am. Wildl. & Nat. Resour. Conf. 42:146-155. Dietz, N. J. 1990. Surveys of citizens’ attitudes towards hunting, hunters and wildlife in South Dakota. M.S. Thesis, South Dakota State University, Brookings. 238pp. Dillman, D. A. 1978. Mail and telephone surveys: A total design method. John Wiley & Sons, New York. 325pp. Hoffman, L. W. 1977. Changes in family roles, socialization, and sex differences. Am. Psychologist. 32:644-657. Kellert, S. R. 1976. Perceptions of animals in American society. Tran. N. Am. Wildl. & Nat. Resour. Conf. 41:533-546. Kellert, S. R. 1978. Attitudes and characteristics of hunters and antihunters. Trans. N. Am. Wildl. & Nat. Resour. Conf. 43:412-423. 238 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Kellert, S. R. and J. K. Berry. 1987. Attitudes, knowledge, and behavior toward wildlife as affected by gender. Wildl. Soc. Bull. 15:363-371. Linsky, A. S. 1975. Stimulating responses to mailed questionnaires: a review. Public Opinion Quarterly. 39:82-101. Manfredo, M. J., J. J. Sneegas, B. Driver, and A. Bright. 1989. Hunters with dis- abilities: a survey of wildlife agencies and a case study of Illinois deer hunters. Wildl. Soc. Bull. 17:487-493. Peterle, T. J., and J. E. Scott. 1977. Characteristics of some Ohio hunters and nonhunters. J. Wildl. Manage. 41:386-399. Rosonke, J. R., R. T. Wagner, R. M. Dimit, and R. L. Linder. 1975a. Attitudes of South Dakota residents toward hunting, hunters and game ofﬁcials. Re- search in Sociology and Hunting in South Dakota, Hunting and Wildlife Re- ports No. 2. 12pp. Rosonke, J. R., R. T. Wagner, and R. M. Dimit. 1975b. Factors associated with varying attitudes among South Dakotans toward hunting, hunters and game ofﬁcials. Research in Sociology and Hunting in South Dakota, Hunt- ing and Wildlife Reports No. 5. 40pp. SAS Institute Inc. 1985. SAS user’s guide: statistics. SAS Institute Inc., Cary, NC. Shaw, W. W. 1977. A survey of hunting opponents. Wildl. Soc. Bull. 5:19-24. Shaw, W. W. and D. L. Gilbert. 1974. Attitudes of college students toward hunting. Trans. N. Am. Wildl. & Nat. Resour. Conf. 39:157-162. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 239

STUDENT OPINIONS ABOUT HUNTING IN SOUTH DAKOTA

Robert D. Mendelsohn Rural Sociology Department

Nancy J. Dietz Wildlife and Fisheries Sciences Department

Kenneth F. Higgins Cooperative Fish and Wildlife Research Unit

South Dakota State University Brookings, SD 57007

ABSTRACT

Recent trends have indicated declines in hunter recruitment in South Dako- ta. The present study focuses on two potential reasons for declining participation: one—structural changes in the family unit affecting children’s acculturation toward hunting; and, two—urbanization affecting the recruitment of young hunters. Data was gathered from a stratified random sample of public school, junior high schools in South Dakota. With support from the South Dakota Office of Public Schools, strong cooperation was elicited from the participating schools, enabling the development of a data base carefully profiling attitudes of students across the geographic, socioeconomic, standard demographic variables characteristic of the state as a whole. Student global attitudes about hunting, e.g., whether “it was okay to hunt wild animals” as well as specific self attitudes about hunting reflected not only family composition, but also degrees of urbanization. In general, students from single parent-headed households and from larger urban school districts were the most negative about hunting. Related factors in the broad category of availability of alternative recreational activities were also found to be associated with negative attitudes about hunting. Other studies reveal similar findings. Conclusions pointed to increasing educational and marketing programs directed at young hunters recruitment across the state.

INTRODUCTION

The past decade has witnessed a steady decrease in hunter participation in South Dakota (Dietz, 1990). Although many factors may be responsible for this decline, eight have been offered as most important. They are: (1) individuals lack personal leisure time; (2) there have been shifts from hunting to other forms of recreation, particularly those that are more family oriented, such as camping, jogging, biking, hiking, and ﬁshing; (3) hunting license fees and 240 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) equipment may be becoming too costly; (4) game populations are too low; (5) fewer hunting lands are accessible to hunters as private landowners close their property to hunting or charge access fees; (6) the population of hunters is aging; (7) people are changing their attitudes about hunting, (8) there is insuf- ﬁcient recruitment of youthful hunters for the future. The present study investigations asks:

How do structural conditions of urbanization and parental composition affect the attitude formation and recruitment of young hunters in South Dakota?

RELEVANT LITERATURE

As with the decline of adult hunter participation in general, multiple factors may contribute to the decrease in hunter youth recruitment in South Dako- ta. Research has shown that rural children who are introduced to hunting at an early age continue to exhibit consistent participation (Brown, et al. 1987). O’Leary, et al. (1987) found an association between days spent hunting per year and age of initial experience i.e., the younger a hunter was at initiation the more involved they were in the sport. By contrast, when hunting was not introduced to an individual at a young age (16 years or younger), desertion rates from hunting were greater; particularly among those individuals who began hunting after the age of 20 (Applegate, 1977). In Pennsylvania, Sofranko and Nolan (1972) found that increased hunting frequency during the formative years predisposed citizens into more frequent participation as adults. In a similar study, Purdy et al. (1985) found that people who were initiated to hunting during post-adolescence were about twice as likely to desert hunting activities within a few years of initiation as those who were initiated in early adolescence. Hunting is a traditional rural form of recreation. Twenty-five years ago, Hendee (1969) predicted that trends toward urbanization would be expected to affect, negatively, hunting participation. A collection of studies (Klessig and Hale, 1972; Langenau and Mellon-Coyle, 1977) indicate individuals from urban areas tend to begin hunting at a later age. This urban-rural connection suggests that as rural areas come to more closely mirror their urban counter- parts, the proportion of hunters will continue to decline. Attitudes towards hunting are also influenced by citizens’ previous connection to rural and urban culture as well as their age. Shaw (1977) surveyed members of anti-hunter organizations and found that members came from predominately urban backgrounds. In New York state, slightly less than half of landowners who were reared in metropolitan areas were opposed to or had reservations about hunting (Brown 1974). Shaw and Gilbert (1974) discovered anti-hunting sentiment of college students was strongest among those from large population centers. In other studies (Kellert 1976, 1978), urban childhood residence affected responses to two of three stimulus statements about the “desirability of hunting.” Peterle and Scott (1977) reported that anti-hunting sentiments will probably become stronger in an urbanized society with a Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 241 declining percentage of hunters. In New Jersey, Applegate (1979) found individuals who opposed deer hunting were generally less than 30 years old. Clearly the rural versus urban environment (within which early socialization of young persons occurs) influences development of favorable or unfavorable attitude dispositions toward hunting as a form of recreational activity. Beyond the broad cultural shifts associated with rural-to-urban changes, the once common, traditional rural family units, facilitators in the transmission of hunting values from one generation to the next, are rapidly disappearing (Brown et al., 1987). Nationally the number of single-parent households has doubled since 1970, primarily due to divorces (USDC, 1986). In 1985, 15 million children under the age of 19 were living with only one parent. Of these, 90% resided with their mothers. These single-parent homes are dominated by women, most of whom do not hunt, and, therefore, do not follow the patterns of transmitting traditional hunting values, from the father to son or daughter. (Brown et al., 1987). Finally, single parent households may not be able to afford the increasing costs of hunting for “even when female-headed families are not poor, they are much less likely than are other families to be affluent, for the major source of high family income is two earners” (Stark, 1989:392).

STUDY OBJECTIVES

Based on suggestions from the literature concerning hunter recruitment factors, the following research objectives were developed:

(1) determine the extent to which urbanization has affected the recruitment and attitude development of young hunters in South Dakota, and (2) determine whether structural changes in family composition inﬂuence young persons’ attitudes about hunting in South Dakota.

METHODS

Sampling

Based upon prior research describing the optimal ages for hunter recruitment (Appleton, 1977), a decision was made to sample young persons at the junior high school level. South Dakota schools were grouped into eleven size classifications: less than 100; 100-199; 200-299; 300-399; 400-499; 500-599; 600- 699; 700-799; 1000-1,499; 1,500-4,999, and 5000 or more students. Using the eleven classifications, a stratified random sample of 6th, 7th and 8th grade students was obtained from all public school districts within South Dakota.

Survey Instruments

In August 1989, a letter of support for the student questionnaire was so- licited and received from the South Dakota Department of Education and Cul- tural Affairs, Division of Education, State Superintendent’s Office of Public Schools. A six page, teacher-supervised, self-administered student question- 242 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) naire was developed. It contained 75 questions which could be answered in 5 to 10 minutes. Student survey answers were recorded on machine readable answer sheets. Forty-eight of the 75 questions on the student questionnaire were adopted from a questionnaire presented to a sample of Michigan youth (Pomerantz 1977); the other 27 questions were develpped by the authors (Di- etz, 1990). Before final distribution, the questionnaire was pretested with a sample of students in a local, Brookings County junior high school. Each teacher administering the survey was first contacted via telephone and informed of the procedures for the administering the survey, after which written instructions along with the correct number of questionnaires and all related relevant materials, e.g., pencils, for the class of students were sent by mail to those teachers. SDGFP posters and calendars were provided for each participating school and student respondents were given SDSU pencils to keep after completing the surveys. Completed questionnaires were returned by all school districts for a total 1,408 student respondents.

RESULTS

South Dakota student respondents (N=1,400) were 52.7% male and 47.3% female. Less than 1% were 5th graders; 36.1%, 31.7%, and 31.4% were 6th, 7th, and 8th graders, respectively. Student ages ranged from 10 to 16 years. Twen- ty-two percent, 32.9%, 31.6%, and 10.4% were 14, 13, 12, and 11 years of age, respectively. Almost 67% of student respondents resided in town and 32.8% in the country. Of those students who resided in the country (N=438), 61.6% of their families operate a farm or ranch. Approximately 60% of South Dakota student respondents had gone hunting, 17.3% had never gone hunting, but would like to, and 22.5% had never gone hunting and had no interest in hunting. Almost 97% of student respondents had gone fishing. Nearly 41% of the students who had gone hunting in the past (N=832) possessed a hunting license in 1989. Almost 81% of the student respondents had shot a BB gun or air rifle, and 61.9% had shot a firearm, for example, a rifle, pistol, or shotgun.

Parental Composition and Acculturation to Hunting in South Dakota

Approximately 83% of South Dakota student respondents lived with both parents, which may include a step parent, and 13.8% lived with one parent. No significant differences existed by SDGFP region and with whom a student lived most of the time (Table 1). Of those students who lived with one parent (N=183),most of the time, 74.3% lived with their mother while 21.9% lived with their father. Significant differences existed by parental situation and participation of an immediate family member in hunting. Approximately 75% of two-parent families and 63.2% of one parent families had at least one member of their immediate family who hunted (Table 2). In 1990, 16.1% of the sample of student respondents agreed with the statement “I think all hunting of wild animals should be against the law”, and 67.0% disagreed with the statement (Table 3). No significant differences were found Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 243

Table 1. Percent of South Dakota sixth, seventh, and eighth grade student respondents’ parental situation according to their residence in South Dakota Game, Fish and Parks (SDGFP) region, 1990. (N= 1,353)

Percent by SDGFP Region Parental Situation Total I II III IV

One Parent 13.8 14.4 10.7 16.1 12.8 Two Parentsa 82.6 81.0 86.4 80.0 84.8 Otherb 3.6 4.6 2.9 3.9 2.4

X2 = 7.823 d.f. = 6 P = 0.251 a May include a step parent. b Includes grandparent/s, aunt and/or uncle, or other situation.

Table 2. Percent of South Dakota sixth, seventh, and eighth grade student respondents’ parental situation and hunting participation of an immediate family member, 1990.

Percent by Parental Situation Member of Immediate One Parenta Two Parents Otherb Family Hunts (N=185) (N=1,117) (N=47)

Yes 63.2 75.3 63.8 No 36.8 24.7 36.2

X2 = 13.954 d.f. = 2 P = 0.001 a May include a step parent. b Includes grandparent/s, aunt and/or uncle, or other situation.

to exist by parental-family composition and extent of agreement or disagreement with this statement. A minority (26.8%) of student respondents agreed with the statement “I think hunting wild animals for enjoyment is O.K.,” and 57.5% disagreed with the statement (Table 3). Signiﬁcant differences in the extent of agreement or disagreement with the above statement existed by parental composition (Table 4.0). Students who lived in one parent families disagreed (64.0%) more often than students who lived in two parent families (56.1%). Eighty-eight percent of respondents agreed with the statement “I think hunting wild animals for food is O.K.” (Table 3). No signiﬁcant differences existed by parental composition and agreement or disagreement with the above statement. More than half (58.0%) of the student respondents agreed, 23.8% didn’t know, and 18.2% disagreed with the statement “I think it is O.K. for oth- 244 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 3 Frequency and percent responses of South Dakota's sixth, seventh, and eighth grade students to ﬁve stimulus statements, 1990.

Responsea Stimulus Statement Disagree Don’t Know Agree

“I think all hunting of wild animals should 941 237 227 be against the law.” (67.0%) (16.9%) (16.1%) (N = 1,405)

“I think hunting wild 808 220 377 animals for enjoyment (57.5%) (15.7%) (26.8%) is O.K.” (N = 1,405)

“I think hunting wild 56 113 1,235 animals for food is (4.0) (8.0%) (88.0%) O.K.” (N = 1,404)

“I think it is O.K. 256 334 815 for other people to (18.2%) (23.8%) (58.0%) hunt wild animals.” (N = 1,405)

“I think it is O.K. 360 294 751 for me to hunt wild (25.6%) (20.9%) (53.5%) animals.” (N = 1,405) a Response is grouped by categories: agree (likert 1 and 2), don't know (likert 3), and disagree (likert 4 and 5).

er people to hunt wild animals” (Table 3). No signiﬁcant differences existed by parental situation and agreement or disagreement with this statement. Nearly 54% agreed and 25.6% of student respondents disagreed with the statement “I think it is O.K. for me to hunt wild animals” (Table 3). Signiﬁcant differences in the extent of agreement or disagreement with the above statement existed by parental composition (Table 4). Those students who lived with two parents agreed most often (54.3%) with the above statement compared to those students who lived with one parent (Table 5).

Urbanization and the Recruitment of Young People into Hunting in South Dakota

In 1989, statistically signiﬁcant differences (p<.05) were found to exist by school size classiﬁcation and agreement-disagreement responses to the follow- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 245 ing statements: “I think all hunting of wild animals should be against the law”, “I think hunting wild animals for enjoyment is O.K.”, “I think it is O.K. for other people to hunt wild animals” and “I think it is O.K. for me to hunt wild animals” (Tables 6, 7, 8).

Students attending larger schools tended to be more negative towards hunting. Signiﬁcant differences (p<.05) existed between student residence in

Table 4. Percent responses by parental situation of South Dakota's sixth, seventh, and eighth grade students to the statement “I think hunting wild animals for enjoyment is O.K.”, 1990. (N = 1,363)

Percent by Parental Situation One Parenta Two Parents Otherb Response (N=189) (N=1,125) (N=49)

Strongly Agree 6.9 6.2 0.0 Agree 18.0 20.7 32.6 Don't Know 11.1 17.0 4.1 Disagree 25.4 23.9 20.4 Strongly Disagree 38.6 32.2 42.9

X2 = 18.939 d.f. = 8 P = 0.015 a May include a step parent. b Includes grandparent/s, aunt and/or uncle, or other situation.

Table 5 Percent responses by parental situation of South Dakota's sixth, seventh, and eighth grade students to the statement “I think it is O.K. for me to hunt wild animals,” 1990. (N = 1,363)

Percent by Parental Situation One Parenta Two Parents Otherb Response (N=189) (N=1,125) (N=49)

Strongly Agree 22.2 23.9 36.7 Agree 24.9 30.4 32.7 Don’t Know 20.1 21.5 12.2 Disagree 13.2 11.7 0.0 Strongly Disagree 19.6 12.4 18.4

X2 = 20.379 d.f. = 8 P = 0.009 a May include a step parent. b Includes grandparent/s, aunt and/or uncle, or other situation. 246 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 6. Percent responses by school size classiﬁcation of South Dakota's sixth, seventh, and eighth grade students to the statement “I think all hunting of wild animals should be against the law,” 1990.

Percent by School Size Classiﬁcationa 123456 Responseb (N=27) (N=47) (N=44) (N=65) (N=90) (N=121)

Disagree 66.7 74.4 88.6 70.8 78.9 71.9 Don’t Know 22.2 12.8 6.8 10.8 12.2 16.5 Agree 11.1 12.8 4.6 18.4 8.9 11.6

Percent by School Size Classiﬁcation 7 8 9 10 11 Responseb (N=150) (N=150) (N=151) (N=234) (N=313)

Disagree 67.3 62.7 60.2 66.2 63.0 Don’t Know 20.0 20.0 16.6 18.4 16.9 Agree 12.7 17.3 23.2 15.4 20.1

X2 = 34.650 d.f. = 20 P = 0.022 a 1= under 100; 2= 100-199; 3= 200-299; 4= 300-399; 5= 400-499; 6= 500-599; 7= 600-699; 8= 700-999; 9= 1,000-1,499; 10= 1,500-4,999; 11= 5,000 or above. b Response is grouped by categories: agree (likert 1 and 2), don't know (likert 3), and disagree (likert 4 and 5).

Table 7. Percent responses by school size classiﬁcation of South Dakota's sixth, seventh, and eighth grade students to the statement “I think hunting wild animals for enjoyment is O.K.,” 1990.

Percent by School Size Classiﬁcationa 123456 Responseb (N=27) (N=47) (N=44) (N=65) (N=90) (N=120)

Disagree 48.2 55.3 59.1 50.8 38.9 52.5 Don’t Know 11.1 19.2 18.2 10.8 24.4 12.5 Agree 40.7 25.5 22.7 38.5 36.7 35.0

Percent by School Size Classiﬁcation 7 8 9 10 11 Responseb (N=150) (N=150) (N=151) (N=234) (N=314)

Disagree 61.4 60.0 62.9 52.1 65.9 Don’t Know 13.3 16.7 14.6 18.0 14.0 Agree 25.3 23.3 22.5 29.9 20.1

X2 = 42.274 d.f. = 20 P = 0.003 a 1= under 100; 2= 100-199; 3= 200-299; 4= 300-399; 5= 400-499; 6= 500-599; 7= 600-699; 8= 700-999; 9= 1,000-1,499; 10= 1,500-4,999; 11= 5,000 or above. b Response is grouped by categories: agree (likert 1 and 2), don't know (likert 3), and disagree (likert 4 and 5) Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 247

Table 8. Percent responses by school size classiﬁcation of South Dakota's sixth, seventh, and eighth grade students to the statement “I think it is O.K. for other people to hunt wild animals,” 1990.

Percent by School Size Classiﬁcationa 123456 Responseb (N=27) (N=47) (N=44) (N=65) (N=90) (N=121)

Disagree 11.1 21.3 2.3 20.0 8.9 14.1 Don’t Know 14.8 17.0 25.0 15.4 23.3 25.6 Agree 74.1 61.7 72.7 64.6 67.8 60.3

Percent by School Size Classiﬁcation 7 8 9 10 11 Responseb (N=149) (N=150) (N=151) (N=234) (N=314)

Disagree 19.5 21.3 21.2 16.2 22.0 Don’t Know 26.8 24.7 21.2 23.1 26.4 Agree 53.7 54.0 57.6 60.7 51.6

X2 = 31.816 d.f. = 20 P = 0.045 a 1= under 100; 2= 100-199; 3= 200-299; 4= 300-399; 5= 400-499; 6= 500-599; 7= 600-699; 8= 700-999; 9= 1,000-1,499; 10= 1,500-4,999; 11= 5,000 or above. b Response is grouped by categories: agree (likert 1 and 2), don't know (likert 3), and disagree (likert 4 and 5).

Table 9. Percent responses by school size classiﬁcation of South Dakota’s sixth, seventh, and eighth grade students to the statement “I think it is O.K. for me to hunt wild animals,” 1990.

Percent by School Size Classiﬁcationa 123456 Responseb (N=27) (N=47) (N=44) (N=65) (N=90) (N=121)

Disagree 22.2 25.5 11.4 26.2 15.5 19.0 Don’t Know 14.8 14.9 18.2 16.9 18.9 26.4 Agree 63.0 59.6 70.4 56.9 65.6 54.6

Percent by School Size Classiﬁcation 7 8 9 10 11 Responseb (N=149) (N=150) (N=151) (N=234) (N=314)

Disagree 25.5 30.7 27.8 24.8 30.2 Don’t Know 20.8 20.7 15.2 20.5 25.8 Agree 53.7 48.7 57.0 54.7 44.0

X2 = 35.676 d.f. = 20 P = 0.017 a 1= under 100; 2= 100-199; 3= 200-299; 4= 300-399; 5= 400-499; 6= 500-599; 7= 600-699; 8= 700-999; 9= 1,000-1,499; 10= 1,500-4,999; 11= 5,000 or above. b Response is grouped by categories: agree (likert 1 and 2), don't know (likert 3), and disagree (likert 4 and 5). 248 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Table 10. Percent responses by residence in South Dakota Game, Fish and Parks (SDGFP) region of South Dakota’s sixth, seventh, and eighth grade students to the statement “I think hunting wild animals for enjoyment is O.K.,” 1990.

Percent by SDGFP Region I II III IV Responsea (N=287) (N=287) (N=464) (N=354)

Disagree 58.2 52.6 64.4 52.3 Don't Know 13.9 19.5 13.8 16.1 Agree 27.9 27.9 21.8 31.6

X2 = 18.857 d.f. = 6 P = 0.004 a Response is grouped by categories: agree (likert 1 and 2), don't know (likert 3), and disagree (likert 4 and 5).

Table 11. Percent responses by residence in South Dakota Game, Fish and Parks (SDGFP) region of South Dakota's sixth, seventh, and eighth grade students to the statement “I think it is O.K. for me to hunt wild animals,” 1990.

Percent by SDGFP Region I II III IV Responsea (N=287) (N=287) (N=463) (N=355)

Disagree 24.4 25.1 28.7 22.8 Don’t Know 16.0 19.2 24.2 22.5 Agree 59.6 55.7 47.1 54.7

X2 = 15.417 d.f. = 6 P = 0.017 a Response is grouped by categories: agree (likert 1 and 2), don't know (likert 3), and disagree (likert 4 and 5). Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 249

SDGFP region and agreement or disagreement with two stimulus statements, “I think hunting wild animals for enjoyment is O.K.” and “I think it is O.K. for me to hunt wild animals” (Tables 10 and 11). Students who resided in regions 1 and 3 disagreed most often with “hunting wild animals for enjoyment is O.K.”, while students who resided in regions 2 and 3 disagreed most often that “it is O.K. for me to hunt wild animals”.

DISCUSSION AND RECOMMENDATIONS

Ways to Increase Hunter Participation

Clearly, although there are a variety of reasons behind the decline of citizens’ hunting in South Dakota; there also are many ways to encourage participation. A majority (75%) of South Dakotans with children indicated they would encourage their child to take a hunter safety education course (Dietz, 1990). Most adults would also allow their child to participate in a SDGFP sponsored activity of either shooting clay or paper targets or an actual hunting experience. It is disconcerting that a quarter of current hunters in South Dakota who have children between the ages of 10 and 15 have never taken those children hunting (Dietz, 1990). On a local level, many children are interested in hunting and have completed a hunter safety-education course but have never had a hunting experience because they have no family role models who hunt. Lo- cal hunting clubs need to be aware of this problem and should be encouraged to help. For example, an adopt-a-young-hunter program could be started as one solution to non-parent participation. Local hunting club members could become active as surrogate hunting role models in such a program. Also special adult-youth hunts could be established to encourage greater participation in hunting activities. The phenomenon of declining interest in hunting by younger generations has appeared in several states (Gilbert 1977, Applegate 1982, 1989); therefore, in this regard, South Dakota is not unique. Nevertheless, one wonders about the direction of South Dakota’s future with hunter participation. In 1989, six of ten South Dakota student respondents had been hunting; nearly 70% disagreed that hunting should be against the law. South Dakota’s 6th, 7th and 8th grade students had lower percentages of anti-hunting sentiment (16.1%) than 8th and 12th grade students in Missouri (20%)(Stout et al., 1988). Moreover, less than half of 5th grade students in Missouri believed that deer hunting was okay (Glover et al., 1986). In general, South Dakota’s 6th, 7th, and 8th grade student respondents were also more favorable about hunting than were 7th through 12th grade Mighigan student respondents in 1976 (Pomerantz 1977). Almost 90% of South Dakota students agreed it was okay to hunt wild animals for food whereas nationally 46% of 5th and 6th grade students agreed that it was okay to hunt wild animals for food, but almost as many opposed (42%) the activity as supported it (Westervelt and Llywellyn, 1985). Primary socialization of children occurs in the family (Applebaum and Chambliss, 1995). Family structure and identification in early life may lead to 250 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) a persisting set of attitudes (Hollander, 1971). Brown et al. (1987) suggested that changes in the traditional family structure would erode transmission of hunting values among future generations. Nearly 14% of South Dakota student respondents lived in single parent households 74.3% of these were headed by women. This is less than the 1985 national percentage where 20% of households were headed by single parents, and of these 90% were headed by women (USDC, 1986). Parental composition did not influence student responses to either of the following statements: “I think all hunting of wild animals should be against the law” and “I think it is O.K. to hunt wild animals for food”. Parental situation did influence student responses to the following statements: “I think hunting wild animals for enjoyment is O.K.,” “I think it is O.K. for me to hunt wild animals,” and “I think it is O.K. for other people to hunt wild animals.” Those students living in two parent families and other parental situations agreed most often with the above three statements. This may be a result of more two parent families having at least one immediate family member who hunted than one parent families. No other studies have examined this aspect. South Dakota students who opposed hunting tended to be from more pop- ulated areas. Westervelt and Llewellyn (1985), in a nationwide Weekly Reader Survey, found that urban and suburban 5th and 6th grade students were less likely to have fished, hunted, or have a father who goes hunting, and were more often opposed to hunting than were rural students. Lahart (1981) and Pomerantz (1977) also found similar urban-rural differences. In contrast to the present findings and that of others (Pomerantz 1977, Lahart, 1981, Westervelt and Llewellyn, 1985), Kellert and Westervelt (1983) found no urban/rural influence on anti-hunting or animal rights attitudes (moralistic attitude) among Connecticut youth. To a considerable degree, the future of hunting in South Dakota is dependent on the attitudes of the younger generations. In 1989, junior high students were not as enthusiastic about hunting as were adults (Dietz, 1990). In addition, students lacked a complete understanding of conservation and wildlife management concepts and the role of hunting in South Dakota. Several studies (Lahart. 1981; Miller, 1975; Westervelt and Llewellyn, 1985) have suggested students have already developed their adult attitudes by the eighth grade. Accordingly, wildlife programs should be presented by trained guest edu- cators in middle and elementary schools, particularly in schools with large numbers of students. Adults should also be encouraged to support the addition of an environmental conservation and ecology course in the middle school curriculums. Brown, et al. (1987) suggested that efforts are needed to infuse wildlife ecology concepts into public school curricula to ensure that young persons in their formative attitude developmental years may acquire an understanding of the basics for wildlife management. Project WILD was recently activated in South Dakota. Undoubtedly, Project WILD has the potential to as- sist young persons’ appreciation of wildlife; however, it may fail to teach students about the principles of life and death processes. It is recommended that additional student surveys, replicating or similar to the present study, should be repeated at five to ten year intervals to determine if project WILD or other Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 251 resource education projects in South Dakota have resulted in any changes in citizen attitudes towards land ethics and natural resources in general, and towards wildlife and hunting in particular. Education should include contacts with wildlife, at least visual contact, since these experiences are internalized most. Students should also be encouraged to engage in shooting sports, including non-living targets. Additional studies should examine hunter-safety education participants and their desertion patterns and re-examine the effect of age of initiation to hunting on hunter recruitment and persistence in hunting sports. Studies also should be initiated to assess the effectiveness of new education and public relations programs regarding attitudes, effectiveness and efficiency of the wildlife division operations of SDGFP. In summary, while other research has suggested hunting in South Dakota is accepted by a strong majority of all South Dakota citizens (Dietz, 1990), emerging patterns suggest it is less sought out as a recreational activity by today’s South Dakota youth. Directed activities toward reversing these trends would seem logical, not only in light of the eonomic impact of hunting on the state’s economic sectors, but also as an effort at preserving a once common- place, rural recreational tradition.

LITERATURE CITED

Applebaum, R. P., W. J. Chambliss, Sociology, 1995. Harper Collins Publilsh- ers, New York New York. 115. Applegate, J. E. 1973. Some factors associated with attitude toward deer hunting in New Jersey residents. Trans. N. Am. Wildl. & Nat. Resour. Conf. 38:267-273. Applegate, J. E. 1977. Dynamics of the New Jersey hunter population. Trans. N. Am. Wildl. & Nat. Resour. Conf. 42:103-116. Applegate, J. E. 1979. Attitudes toward deer hunting in New Jersey: a decline in opposition. Wildl. Soc. Bull. 7:127-129 Applegate, J. E. 1989. Patterns of early desertion among New Jersey hunters. Wildl. Soc. Bull. 17:476-481. Brown, T. L. 1974. New York landowners’ attitudes toward recreation activities. Trans. N. Am. Wildl. & Nat. Resour. Conf. 39:173-179. Brown, T. L. K. G. Purdy, and G. F. Mattfeld. 1987. The future of hunting in New York. Trans. N. Am. Wildl. & Nat. Resour. Conf. 52:553-566. Dietz, N. 1990. Surveys of Citizens’ Attitudes Toward Hunting, Gilbert, A. H. 1977. Influence of hunter attitudes and characteristics on wildlife management. Trans. N. Am. Wildl. & Nat. Resour. Conf. 42:226-236. Glover, R. L., P. S. Haverland, and D. K. Heard. 1986. Knowledge of Missouri fifth graders about deer. Missouri Dept. of Conservation, Jefferson City, MO. 55pp. Hendee, J. C. 1969. Rural-urban differences reflected in outdoor recreation participation. J. Leis. Res. 1:333-341. Hollander, E. P. 1971. Principles and Methods of Social Psychology. Second ed. Oxford University Press, Toronto, Canada. 711pp. 252 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Kellert, S. R. 1980. Americans’ attitudes and knowledge of animals. Trans. N. Am. Wildl. & Nat. Resour. Conf. 45:111-124. Kellert, S. R. and M. O. Westervelt. 1983. Children’s attitudes, knowledge and behaviors toward animals. U.S. Govt Printing Office, Washington, D.C. 202pp. Klessig, L. L., and J. B. Hale. 1972. A profile of Wisconsin hunters. Wisc. Dept. of Nat. Resour., Tech. Bull. No. 60. 24pp. Lahart, D. E. 1981. The influence of knowledge on young people’s perceptions about wildlife. Proc. Ann. Conf. S.E. Assoc. Fish & Wildl. Agencies. 35:661-668. Langenau, E. E., Jr., and P. M. Mellon-Coyle. 1977. Michigan’s young hunter. Mich. Dept. of Natural Resources, Wildl. Div. Report No. 2800. 54pp. +ap- pendices. Miller, J. D. 1975. The development of pre-adult attitudes toward environmental conservation and pollution. School Sci. and Math. 75:729-737. O’Leary, J. T., J. Behrens-Tepper, F. A. McGuire, and F. D. Dottavio. 1987. Age of first hunting experience: results from a nationwide recreation survey. Leis. Sci. 9:225-233. Peterle, T. J., and J. E. Scott. 1977. Characteristics of some Ohio hunters and nonhunters. J. Wildl. Manage. 41:386-399. Pomerantz, G. A. 1977. Young people’s attitudes toward wildlife. Mich. Dept. of Nat. Resour., Wildl. Div. Report No. 2781. 79pp. Purdy, K. G., D. J. Decker, and T. J. Brown. 1985. New York’s 1978 hunter training course participants: the importance of social-psychological influences on participation in hunting from 1978-1984. Human Dimensions Res. Unit. Ser. No. 85-7, Dept. of Nat. Resour., Cornell Univ., Ithaca, New York. 127pp. Shaw, D. L., and D. L. Gilbert. 1974. Attitudes of college students toward hunting. Trans. N. Am. Wildl. & Nat. Resour. Conf. 39:157-162 Shaw, W. W. 1977. A survey of hunting opponents. Wildl. Soc. Bull. 5:19-24. Sofranko, A. J., and M. F. Nolan. 1970. Selected characteristics, participation patterns, and attitudes of hunters and fishermen in Pennsylvania. Penn. State Univ. Agri. Exper. Stout, R. J., D. K. Heard, and P. S. Haverland. 1988. Knowledge and attitudes of Missouri eighth and twelfth grade students on deer biology and management. Missouri Dept. of Conservation, Columbia, MO. 63pp. U.S.D.C. Bureau of the Census. 1986. Marital status and living arrangements. Current Popn. Reports, Ser. P-20, No. 410. U.S. Govt. Printing Office, Washington, D.C. 91pp. Westervelt, M. O., and L. G. Llewellyn. 1985. Youth and wildlife: the beliefs and behaviors of fifth and sixth grade students regarding non-domestic animals. U.S. Govt Printing Office, Washington, D.C. 78pp. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 253

TRACE METALS IN WATER AND SEDIMENTS OF WETLANDS IN THE RAINWATER BASIN AREA OF NEBRASKA

Christine C. Gordon and Lester D. Flake Department of Wildlife and Fisheries Sciences

Kenneth F. Higgins National Biological Service, South Dakota Cooperative Fish and Wildlife Research Unit

South Dakota State University Brookings, SD, 57007

ABSTRACT

Water column and sediment samples, collected from 55 sites in the rainwater basin area of Nebraska during Spring 1988, were analyzed for total concentrations of 20 metal constituents. The survey was designed to quantify the concentrations of metal constituents in wetlands in an area of intensive agriculture. These wetlands are particularly important to migratory waterfowl and other wetland birds. Trace metals in water column samples were compared with U. S. Environmental Protection Agency (EPA) national criteria for freshwater aquatic organisms. Iron (Fe) and mercury commonly exceeded EPA criteria in study wetlands. Copper (Cu), lead (Pb), and zinc (Zn) EPA criteria depend on water hardness; these elements exceeded EPA criteria in some wetlands for which hardness data were available. EPA criteria for Aluminum (Al), molybdenum (Mo), and strontium (Sr) in water were not available; EPA criteria for trace metals in sediments were not available. Metal concentrations in sediment samples were compared with published crustal abundances from untilled prairie soils in Missouri. Beryllium (Be), mercury (Hg), and zinc (Zn) were in higher concentrations in sediments than in Missouri soils. None of the elements in the sediments were markedly elevated in comparison with background levels reported in the literature.

INTRODUCTION

Intensive agricultural development in Nebraska has prompted concern for possible environmental contamination of wetland areas with concentrations of trace elements resulting from the extensive and intensive use of pesticides, fertilizers, and other agricultural chemicals. Metals may have toxic, sublethal, and latent effects on wildlife if found in sufﬁcient concentrations (Eisler, 1985). In 1983, elevated levels of selenium on the Kesterson National Wildlife Refuge in Central California caused selenium teratogenesis in natural populations of 254 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) aquatic birds (Hoffman et al., 1988). Blus et al. (1987) reported that declines in certain mammal populations in highly contaminated areas near smelters in northern Idaho were attributed to direct and secondary effects of toxic metals, such as lead and cadmium. Since 1970, sport and commercial ﬁshing have been banned in mercury contaminated waters in 26 of the contiguous 48 states (Eisler, 1987). The rainwater basin area of Nebraska (Fig. 1) serves as a major staging area for migratory waterfowl. Each spring since 1975, thousands of migrating waterfowl have succumbed to avian cholera as a result of the bacterium Pas- teurella multocida. Researchers have suggested that the micro-environment of a wetland may be a critical factor involved in the transmission of the disease, the lifespan of the disease organism, or the increased susceptibility of waterfowl to a disease (Friend, 1981; Franson, 1986). The objectives of this study were (1) to determine the concentrations of 20 trace metals in the water columns and sediments of wetland areas in south-central Nebraska; (2) to compare water column concentrations with U. S. Envi- ronmental Protection Agency (EPA) national criteria for freshwater aquatic organisms; (3) to compare metal concentrations in sediments with crustal concentrations in untilled soils characteristic of nonglaciated prairie in Missouri (standards are not available for sediments); and (4) to compare water column and sediment concentrations with values from published studies as available.

STUDY AREA

The rainwater basin area, which is located in south-central Nebraska, occupies approximately 10.8 thousand square km of land area. The basin for- mations characteristic of the area are thought to have been caused by irregular loess deposits, as controlled by topography and modiﬁed by wind (Evans and Wolfe, 1967). Although the wetland basins occur as a continuum, the rain-

Figure 1. Location of the eastern and western regions of the study area in the rainwater basin of Nebraska. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 255 water basin area is administratively divided into an eastern and western region. The main differences between the eastern and western basin areas are that the topography is more rolling and rainfall is greater in the east, allowing for deeper and more permanent wetland basins due to more runoff. Water levels are generally dependent upon direct precipitation, runoff from snowmelt, and irrigation systems. In addition, deep percolation of water from the Tri-County Canal System and from irrigation runoff from row crops has raised the water table 3 m or more in some areas (Spalding, 1981). The up- lands surrounding most wetland basins experience intensive agricultural practices and deep well irrigation, primarily by center-pivot systems. Agriculture was the primary land use within all watersheds. The extent of idle lands sep- arating study wetlands from adjacent agricultural activities varied from approximately 100% native prairie to 100% cultivated farm land (Gordon, 1989).

METHODS

Samples were collected in both the eastern and western regions during March and May of 1988. Forty-nine water column samples and 53 sediment samples were collected from 55 collection sites that included 36 wetland basins, 2 irrigation canals, 11 re-use pits, 2 intensive cattle-use areas, 3 rivers (Platte, Big Blue, Little Blue), and 1 area that was directly affected by an irrigation pivot. One sample collected from a cattle-use area was taken from a dry basin just below a feedlot, and another was collected from a livestock wa- tering area. About 1/3 of the sample sites were in the western region and 2/3 in the eastern region of the rainwater basin (Fig. 1). Water samples were collected in acid-washed, 125-ml, polyethylene bottles which were submerged 5-10 cm below the water surface, and later preserved with nitric acid to a pH of 2. Sediment samples were collected using an epoxy- coated core sampler. A minimum of 30 g of sediment were placed in acid- washed, Qorpak, 118-ml, glass jars with Teﬂon-lined lids. Each sample was a composite of different layers of a single core sample. Sediment samples were kept frozen until analysis. All water column and sediment samples were analyzed for 20 trace metals at the Environmental Trace Substances Research Center, at the University of Missouri, Columbia. At the center, sediment samples were dried, weighed, and further homogenized using a blender. Approximately 0.5 g of sediment or 50 ml of water were digested by one of 2 methods: mercury was analyzed on nitric reﬂux digestates, while the remaining elements were analyzed on nitric- perchloric acid digestates. Concentrations were determined using 3 methods: mercury concentrations were analyzed by cold vapor atomic absorption; arsenic and selenium concentrations were determined using a hydride method; and the other 18 trace metals (Ag, Al, B, Ba, Be, Cd, Cr, Cu, Fe, Mn, Mo, Ni, Pb, Sr, Tl, V, Zn) were analyzed using an inductively coupled plasma (ICP) scan. Instrumentation included a Perkin-Elmer 403AA, Varian VGA-76 hydride generator, Perkin-Elmer Model 603AA or 3030AA, and a Jarrell-Ash Model 1100 Mark III. 256 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Trace metal concentrations in water column samples were compared with EPA national criteria where available for freshwater aquatic organisms (EPA, 1986; EPA, 1987). To date, EPA has not developed a set of sediment quality criteria; therefore, concentrations were compared with the crustal abundances reported by Erdman et al. (1976A) and Erdman et al. (1976B).

RESULTS

Water column elemental concentrations were lower than EPA national criteria for freshwater aquatic organisms with the exception of Cu, Fe, Hg, Pb, and Zn (EPA, 1986; EPA, 1987) (Table 1). Criteria for Cu, Pb, and Zn require knowledge of water hardness. Although these data were not collected as part of this study, information pertaining to water hardness was obtained for 7 of the 56 study areas (Gordon, 1989) and it was determined that Cu, Pb, and Zn exceeded the EPA criteria. Water hardness ranged from 33 mg/L at County Line Marsh to 439 mg/L at Funk Lagoon. Generally, water hardness increases from the eastern region to the western region of the rainwater basin area (Spalding, 1981; Windingstad et al., 1984; Gordon, 1989). Copper was found to be elevated in 2 of the 7 wetlands for which water hardness data were available. Ex- cessive amounts of lead were found in 2 of the 7 wetlands. Zinc concentrations were high in 3 of the 7 wetlands for which water hardness data were available. Water column concentrations for Ag, Be, Cd, Cr, Hg, Mo, Ni, and Tl were found near or below detection limits. National criteria were not available for aluminum, strontium, and vanadium. Most of the elements in sediment samples were in lower concentrations than those reported for unglaciated soils in Missouri. Trace metal concentrations were not unusual for loess and soils on silt deposits (Kabata-Pendias and Pendias, 1984).

DISCUSSION

Copper is an essential micronutrient for living organisms but it can produce toxic effects at elevated levels. Principal sources of copper include the following: metalliferous mining, smelting, industrial emissions and efﬂuents, traf- ﬁc, urban development and dumped waste materials, contaminated dusts and rainfall, sewage sludge, pig slurry, composted refuse, fertilizers, ameliorants, and pesticides (Nriagu, 1979). Copper is widely used as an algicide and herbicide for nuisance aquatic plants, but this use is uncommon in the rainwater basin. Potential copper sources in the study area would primarily be pesticides and fertilizers. Water column concentrations of copper in Nebraska wetlands ranged from <0.004-0.079 mg/L. Spalding (1981) reported concentrations in the groundwater in the Tri Basin area of Nebraska as ranging from <0.002-0.011 mg/L. Copper concentrations in sediments were almost identical to those reported by Erdman et al. (1976A) for unglaciated prairie soils. Mercury from fungicides is a potential source for contamination of rainwater basin wetlands. Although its use as a fungicide has been discontinued in many places, Hg residues may still remain in soils and wetland areas. Mer- Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 257

Table 1. Concentrations (x and range) (mg/l) of trace metals in water column and sediment samples collected from 55 wetlands or other water body sites in the Rainwater Basin of Nebraska. EPA criteria and comparative soil concentrations from unglaciated prairie in Missouri are also presented.

cury concentrations in study wetlands (0.0003 mg/L) exceeded maximum amounts recommended by the EPA for aquatic systems (0.000012 mg/L). Mar- tin and Hartman (1984) found mercury concentrations in sediments to range from 0.02-0.06 mg/kg in the North Central United States. Lake Winnipeg and other Canadian lakes had sediment levels of 0.03-0.28 kg/mg (Allan and Brun- skill, 1977). Levels ranging from 0.03-0.62 mg/kg were reported for sediments in Lake Oahe in South Dakota (Walter et al., 1973). Concentrations in soils collected in eastern North Dakota had a mean concentration of 0.2 mg/kg (Houghton and Briel, 1987). Sediment mercury levels in this study (0.02-0.80 mg/kg) were comparable to other studies but may be of concern in some wetlands. Mercury concentrations in sediments appeared to be clearly elevated in relation to prairie soils in Missouri (Erdman et al., 1976A). Iron is the fourth most abundant element in the make-up of the earth's crust and it is a major component of clay soils (EPA, 1986). Therefore, it is not surprising to find elevated levels of Fe in Nebraska wetlands, where clay soils are dominant (Evans and Wolfe, 1967). Although Fe levels are elevated in comparison to EPA criteria, concentrations found in rainwater basins are considered to be normal. Sediment concentrations were below those observed by Erdman (1976A) for prairie soil in Missouri. Lead enters the environment from several sources. Vehicle exhaust has been a major contributor but restriction on lead use in gasoline has reduced 258 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) this source. Pesticides and lead shot are also sources of Pb (May and McKin- ney, 1981). Oates (1989) examined the incidence of lead shot in 8 rainwater basin wetlands and found an average of 47,217 pellets/ha; steel shot has been required for waterfowl hunting on these areas since 1980. Thus, lead shot may provide a source of contamination for many years. Levels of Pb in the water in rainwater basin wetlands ranged between <0.04-0.17 mg/L. EPA criteria for lead in the water are influenced by water hardness and, in some study wetlands of known hardness, exceeded recommended criteria for aquatic ecosystems; further monitoring of lead concentrations in rainwater basin wetlands is recommended. Sediment concentrations of lead do not appear excessive. Zinc ranks fourth among metals of the world in annual consumption (Nriagu, 1980). Fertilizers and fungicides (EPA, 1987) represent the probable sources for contamination of natural wetlands in Nebraska (EPA, 1987). Zinc occurs naturally in freshwater environments, and exposure to elevated concentrations is one to which species of many groups have adapted themselves (Nriagu 1980). Zinc is an essential micronutrient for all living organisms. Zinc levels in groundwater from irrigation wells in the Tri-basin area of Nebraska were 0.004-0.078 mg/L. The mean concentration for water column samples collected in rainwater basin study wetlands was 0.124 mg/L of zinc. Erdman et al. (1976A) reported levels of zinc in unglaciated prairie soils to be 51 mg/kg (geometric mean). Sediments in the rainwater basin study sites averaged 87 mg/kg of zinc, markedly higher than our sediment concentrations. Selenium is widely distributed in nature and is particularly abundant with sulfide minerals of various metals such as Fe, Pb, and Cu (Eisler, 1985). Sele- nium should be discussed because of the problems this metal can cause in aquatic ecosystems. The recent discovery of high concentrations at Kesterson Reservoir in California has prompted concerns for excessive Se in the environment. Levels of Se reported in water entering Kesterson Reservoir contained 300 ppb; however, water entering the nearby Volta Wildlife Area contained a more normal level of 1 ppb (Ohlendorf et al., 1986). Spalding (1981) reported Se levels of < 0.2-1.9 ppb in groundwater samples collected in the Tri-basin Natural Resources District in Nebraska. Selenium levels in the water reported for study wetlands in the rainwater basin were 0.3-8.0 ppb. Our findings are consistent with background levels reported in the literature. Sediment concentrations of selenium were also at low levels compared to unglaciated prairie in Missouri (Erdman et al., 1976A). Allan and Brunskill (1977) examined the presence of element concentrations in bottom sediments of Lake Winnipeg and other Canadian lakes. Results for the analysis of Be were consistent with the findings of this study. Berylli- um in Canada was present in concentrations which ranged from 1.0-3.1 mg/kg in sediment samples. Rainwater basin samples contained 0.2-2.7 mg/kg of Be. Cadmium should be briefly discussed because it has caused environmental problems elsewhere. Runoff from agricultural areas where phosphate fertilizers have been applied may result in a substantial Cd loading to the aquatic environment (May and McKinney, 1981). Background concentrations of Cd reported in the literature demonstrated considerable variation. Martin and Hartman (1984) reported sediment concentrations ranging from 0.13-1.1 mg/kg Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 259 in pothole wetlands of the North Central United States, including 2 wetlands which were also sampled in this study. Allan and Brunskill (1977) found sediment concentrations in Lake Winnipeg to range from 1.0-3.0 mg/kg. Cadmi- um concentrations ranging up to 13.7 mg/kg were considered to be background levels for Clay and Ball Lakes in Ontario (Jackson, 1978). Based on the concentrations reported in the literature, the findings of this study (<0.2-1.0 mg/kg) did not suggest problems with Cd contamination of water or sediments. Toxic effects are related to the physical and chemical forms of each element, the toxicity of each form, the degree of interaction among the various forms, and the interrelationship between different trace elements. Tolerance ranges of trace metals, as expected, vary among individuals, species, and larger phylogenetic groups. This study was designed to provide baseline data on trace metal concentrations in a variety of environments within the rainwater basin area of Nebraska. The chemical interactions of an aquatic ecosystem are very complex and dynamic and require a thorough examination. Periodic monitoring of trace metals that were near or above EPA criteria or comparable published data is recommended for the rainwater basin wetlands. Additional- ly, laboratory studies could examine the effects of trace metals specific to aquatic species which are found in the rainwater basin area. In general, most trace metals in rainwater basin wetlands are not at concentrations that would cause immediate damage to aquatic ecosystems.

ACKNOWLEDGEMENTS

We thank C. Collins and T. Fairbanks for assistance in sample collection and laboratory preparations. We acknowledge R. M. Wilson, U.S. Fish and Wildlife Service, for assistance with sample collection, preparation of sample logbooks, and for acting as a liaison between South Dakota State University and the analysis laboratory. Dick Gersib, Nebraska Game and Parks Commis- sion, is thanked for his help with sampling site selection. We also express our gratitude to G. Engel, U.S. Geological Survey, for the loan of a core sampler. G.E. Gordon (deceased) is thanked for his encouragement, suggestions, and review of the manuscript.

REFERENCES

Allan, R. J. and G. J. Brunskill. 1977. Relative atomic variation (RAV) of elements in lake sediments: Lake Winnepeg and other Canadian lakes. In: Golterman, H. L. (ed). Interactions between sediments and freshwater. The Hague, The Netherlands. p. 108-120. Blus, L. J., C. J. Henny, and B. M. Mulhern. 1987. Concentrations of metals in mink and other mammals from Washington and Idaho. Environ. Pollut. 44:307-318. Eisler, R. 1985. Selenium hazards to ﬁsh, wildlife, and invertebrates: a synop- tic review. U. S. Fish Wildl. Serv. Biol. Rep. 85(1.5). 57pp. Eisler, R. 1987. Mercury hazards to ﬁsh, wildlife, and invertebrates: a synop- tic review. U. S. Fish Wildl. Serv. Biol. Rep. 85(1.10). 90pp. 260 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Erdman, J., H. T. Shacklette, and J. R. Keith. 1976A. Elemental composition of selected native plants and associated soils from major vegetation-type areas in Missouri. U. S. Geol. Survey Prof. Paper 954-C, C1-C-87. Erdman, J., H. T. Shacklette, and J. R. Keith. 1976B. Elemental composition of corn grains, soybean seeds, pasture grasses, and associated soils from selected areas in Missouri. U. S. Geol. Survey Prof. Paper 954-D, D1-D23. Evans, R. D., and C. W. Wolfe. 1967. Waterfowl production in the rainwater basin area of Nebraska. J. Wildl. Manage. 31:788-794. Franson, J. C. 1986. Immunosuppressive effects of lead. In: Feierabend, J. S., and B. A. Russel (eds). Lead poisoning in wild waterfowl-a workshop. Nat. Wildl. Fed., Washington, D.C. p. 106-109. Friend, M. 1981. Waterfowl management and waterfowl disease: independent or cause and effect relationships? Trans. N. Am. Wildl. Conf. 46:94-103. Gordon, C. C. 1989. The relationship of wetland characteristics to avian cholera (Pasteurella multocida) outbreaks in the rainwater basin area of Nebraska. M. S. Thesis. South Dakota State University, Brookings, SD. Hoffman, D. J., H. M. Ohlendorf, and T. W. Aldrich. 1988. Selenium teratogenesis in natural populations of aquatic birds in central California. Arch. Environ. Contam. Toxicol. 17:519-525. Houghton, R. L., and L. I. Briel. 1987. Arsenic and selenium in irrigable soils of eastern North Dakota. North Dakota Acad. Sci. 85:64. Jackson, T. A. 1978. Biogeochemistry of heavy-metals in polluted lakes and streams at Flin Flon, Canada, and a proposed method for limiting heavy- metal pollution of natural-waters. Environ. Geol. 2:173-189. Kabata-Pendias, A., and H. Pendias. 1984. Trace elements in soils and plants. CRC Press, Boca Raton. Martin, D. B., and W. A. Hartman. 1984. Arsenic, cadmium, lead, mercury, and selenium in sediments of riverine and pothole wetlands of the North Central United States. J. Assoc. Ofﬁc. Anal. Chem. 67:1141-1146. May, T. W., and G. L. McKinney. 1981. Cadmium, lead, mercury, arsenic, and selenium concentrations in freshwater ﬁsh, 1976-77-National Pesticide Monitoring Program. Pestic. Monit. J. 15:14-38. Nriagu, J. O. 1979. Copper in the environment-Part I: Ecological cycling. John Wiley and Sons Inc., New York. 522pp. Nriagu, J. O. 1980. Zinc in the environment-Part I: Ecological cycling. John Wiley and Sons Inc., New York. 453pp. Oates, D. W. 1989. Incidence of lead shot in the rainwater basins of South Central Nebraska. Prairie Nat. 21:137-146. Ohlendorf, H. M., R. L. Hothem, C. M. Bunck, T. W. Aldrich, and J. F. Moore. 1986. Relationship between selenium concentrations and avian reproduction. Trans. N. Am. Wildl. Nat. Res. Conf. 51:330-342. Spalding, M. E. 1981. Areal groundwater quality in the Tribasin Natural Re- sources District. Institute of Agriculture and Natural Resources, Univ. Ne- braska, Lincoln, NE. U. S. Environmental Protection Agency. 1986. Quality criteria for water. U. S. Gov. Printing Off. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 261

U. S. Environmental Protection Agency. 1987. Ambient water criteria for Zinc. U. S. Gov. Printing Off. 207pp. Walter, C. M., F. C. June, and H. G. Brown. 1973. Mercury in ﬁsh, sediments, and water in Lake Oahe, South Dakota. J. Water Pollut. Control Fed. 45:2203-2210. Windingstad, R. M., J. J. Hurt, A. K. Trout, and J. Cary. 1984. Avian cholera in Nebraska's Rainwater Basin. Trans. N. Am. Wildl. Conf. 49:576-583.

Abstracts of Senior Research Papers

presented at The 82nd Annual Meeting

of the South Dakota Academy of Science

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 265

DETERMINING THE PREVALENCE OF WHIRLING DISEASE IN IDAHO RIVERS

Bryan Ledgerwood and Dale Droge College of Natural Sciences Dakota State University Madison, SD 57042-1799

Steve Elle Idaho Department of Fish and Game Nampa, Idaho

ABSTRACT

Whirling disease is a problem that is spreading rapidly through prized trout populations of the Rocky Mountain west. The disease is carried by a protozoan, Myxobolus cerebralis. If the spores are ingested by tubifex worms, the protozoan transforms into a new and deadly form. After leaving the worm, the grappling-hook-shaped spore invades the fish through the skin or gills, passes into the bloodstream, and travels to the head region. The parasite feeds off cartilaginous tissue of the skull and creates five major external characteristics: 1) black tail from the adipose fin to the tail, 2) shortened opercula, 3) scolio- sis of the spine, 4) concavity in the head region, 5) whirling behavior when startled. We conducted a study of the East Fork of the Big Lost River in the Challis National Forest located in central Idaho. We investigated the occurrence of whirling disease by conducting an experiment using lab-reared fish. We put healthy fry in sentinel box exclosures and placed the boxes in the stream. Fry were fed and observed daily. After 10 days the fry were transferred to the fish hatchery and were held in isolation with specific pathogen-free water. We exposed fish during the months of May, June, and July to see if there was any correlation between time of year and number of fry infected. Regardless of time of year, nearly all fish were heavily infected with the parasite. These results indicate the parasite is widely spread throughout the region, even reaching isolated areas such as our study site. This implies that nearly all natural trout populations in the Rocky Mountain area are in danger of contracting whirling disease.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 267

A CHARACTERIZATION OF WATER CHEMISTRY AND PLANKTON FROM FOUR PRAIRIE LAKES

Theodore B. McMillan and Lois Haertel South Dakota State University Brookings SD 57007

ABSTRACT

Plankton and water chemistry were sampled in four glacial prairie lakes on six dates in 1995 and 1996. Phytoplankton and zooplankton were identiﬁed and enumerated. Chemical parameters included nitrate, total kjeldahl nitrogen, total phosphorus, silica, iron, manganese, sodium, potassium, magnesium, calcium, chloride, sulfate, bicarbonate, carbonate and chlorophyll a. Station depth, secchi depth, turbidity, and conductivity were also measured. Two of the lakes were located in moraines and two were located in outwash. One lake was less than 13,000 years old as determined by position, and three were older. Two of the lakes were hypertrophic and two were eutrophic. Two of the lakes were slightly saline in 1995-1996, and two were fresh. Values obtained were compared with previous data bases maintained by the junior author and state agencies from in the 1970's and 1989-1994. Roy Lake (Marshall County) is located in Mankato moraine (less than 13,000 years old). It was slightly saline with the following mean values measured in 1995-6: 1170 uS conductivity, 376 ppm sulfate, 251 ppm bicarbonate, 122 ppm calcium, and 91 ppm magnesium. Roy Lake was eutrophic (7.8 ppb chlorophyll a, 0.03 ppm total phosphorus, 1.31 ppm total nitrogen, 2.32 m secchi depth, and 4.2 ntu turbidity). Mean values for other phytoplankton nutrients were: 0.07 ppm iron, 0.07 ppm manganese, and 8.5 ppm silica. Pickerel Lake (Day County) is older and located in outwash. It was fresh with the following mean values measured in 1995-6: 580 uS Conductivity, 73 ppm sulfate, 189 ppm bicarbonate, 36 ppm calcium, and 34 ppm magnesium. Pickerel Lake was eutrophic (17.7 ppb chlorophyll a, 0.03 ppm total phosphorus, 0.75 ppm total nitrogen, 1.25 m secchi depth, and 7.4 ntu turbidity). Mean values for other phytoplankton nutrients were: 0.17 ppm iron, 0.07 ppm manganese, and 4.5 ppm silica. Bitter Lake (Day County) is older than Pickerel and located in outwash, downstream from Pickerel Lake. Bitter is a saline lake in most years. Howev- er, during 1995-6, its water level rose and it was slightly saline with the following mean values measured in 1995-6: 3758 uS conductivity, 3248 ppm sulfate, 386 ppm bicarbonate, 142 ppm chloride, 509 ppm magnesium, 401 ppm sodium, and 93 ppm potassium. Bitter Lake was hypertrophic (33.7 ppb chlorophyll a, 0.64 ppm total phosphorus, 9.56 ppm total nitrogen, 0.78 m secchi depth, and 38.9 ntu turbidity). Mean values for other phytoplankton nutrients were: 0.20 ppm iron, 0.14 ppm manganese, and 5.0 ppm silica. 268 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

Oak Lake (Brookings County) is located in 14,000 year old moraine. It was fresh with the following mean values measured in 1995-6: 610 uS conductivity, 123 ppm sulfate, 178 ppm bicarbonate, 36 ppm calcium, and 32 ppm magnesium. Oak Lake was hypertrophic (35.6 ppb chlorophyll a, 0.13 ppm total phosphorus, 1.41 ppm total nitrogen, 0.42 m secchi depth, and 37.3 ntu turbidity). Mean values for other phytoplankton nutrients were: 0.28 ppm iron, 0.12 ppm manganese, and 10.3 ppm silica. Comparisons with previous data bases did not indicate major changes between years for Roy, Pickerel and Oak but indicated drastic change between years for Bitter. Major ion concentrations other than calcium were almost an order of magnitude lower in 1995-6 than sampled by the junior author in 1975. In addition, water depth and clarity were much greater in 1995-6 than in all previous years. In previous years Bitter was dominated by non-nitrogen-fixing coccoid bluegreen algae including Anacystis (Microcystis) incerta and cyanea. Anacystis spp. populations were 2 orders of magnitude lower in 1995, however Bitter experienced a heavy bloom of the nitrogen-fixing Aphanizomenon holsaticum (flos-aquae) in midsummer 1995, a species not previously recorded from Bitter. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 269

ROLE OF WETLANDS IN AGRICULTURAL SYSTEMS

D.H. Rickerl Plant Science Department and L.L. Janssen Economics Department South Dakota State University Brookings, SD, 57007. USA

ABSTRACT

Wetlands that formed as a result of glacial activity in the Prairie Pothole Re- gion of North America frequently interact with shallow groundwater systems. The interactions affect water quantity and quality in the agricultural landscape and thus economic returns for farmers. Our project has used a multidisciplinary approach to study environmental and economic impacts of wetlands in agricultural settings. Water budgets were developed to partition water at upland and wetland sites. Results indicated that approximately 60% of the wetland water input was runoff from adjacent fields, and thus a potential for wetland pollution from agricultural management existed. Of the total wetland water budget, wetland storage and recharge to groundwater was nearly 40%. This percentage represents the wetlands’ contribution to surface water detention. The stored water can subsequently be used to replenish soil moisture and or recharge groundwater supplies. Nitrate and ortho-phosphate concentrations, as indicators of water quality, were monitored in wetland surface water, groundwater sampled from wells near the wetland, and groundwater sampled from upland wells. Nitrate concentrations were less in wetland surface water than the discharging groundwater, and were influenced by wetland class. Seasonal wetlands with fluctuating wet and dry cycles were more efficient at removing nitrates than semiper- manent wetlands. Groundwater samples from wells near the wetland margin contained higher concentrations of ortho-phosphate than groundwater sampled at upland sites. The data suggested that the P-sorption capacity of seasonal wetland sediments had been exceeded and soluble phosphorus was being transported to the groundwater. Wetlands are frequently cited as nutrient filters for surface water and groundwater. Although this function helps reduce contamination of surface water and groundwater, it represents a loss of nutrients from the agricultural system. The natural vegetation of wetland margins is very productive. Biomass production of Prairie Potholes rivals that of the tropical rain forest. However, when temporary and seasonal wetlands are farmed through, the natural biomass productivity is not translated to grain yield. Yield measurements collected as a base for economic comparisons established zones of increasing grain yield pro- gressing outward from the wetland. The environmental and economic results indicate that buffer zones may be a management practice which will protect environmental quality of wetlands while increasing economic return.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 271

PALEONTOLOGICAL RESOURCE SURVEY OF STATE-ADMINISTERED LANDS IN SOUTH DAKOTA

Bruce A. Schumacher and James E. Martin Museum of Geology South Dakota School of Mines and Technology Rapid City, SD 57701

Curt Johnson Ofﬁce of School and Public Lands Pierre, SD 57501

ABSTRACT

For the past two years, the Office of School and Public Lands, State of South Dakota has sponsored a state-wide paleontological resource survey. The purpose of the survey is three-fold: to document all state holdings of paleontological significance to provide information for educated decisions concerning fossil resources; to salvage fossil specimens in danger of destruction due to the elements or poaching; and to educate the general public about fossil resources in South Dakota. To date survey crews have examined a significant portion of state holdings in Butte, Custer, Fall River, Harding, Lyman, Meade, Pennington, Perkins, and Stanley counties. Diverse fossil remains have been discovered and salvaged on isolated portions of state land. Salvaged materials include portions of Cretaceous dinosaurs, Cretaceous marine reptiles and invertebrates, and Oligocene mammals. The partial skeleton of a gigantic marine turtle, Archelon, was discovered in Stanley County in 1995 and was recovered in 1996. The skeletal remains of this great turtle are currently being prepared for research and display purposes for the people of South Dakota. The humerus of this specimen is nearly one- half meter in length, indicating a shell length of greater than three meters. Al- though several Archelon skeletons have been collected in South Dakota, this is the first which will remain in the state. We urge the citizens of South Dakota to take an interest in this state’s rich fossil heritage and to support legislation which protects and preserves fossils for future generations.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 273

SEDIMENTOLOGY OF THE COARSE-GRAINED DEPOSITS FROM THE QUATERNARY FOSSIL LAKE AREA IN SOUTH-CENTRAL OREGON

Janet L. Bertog and James E. Martin Museum of Geology South Dakota School of Mines & Technology Rapid City, SD 57701

Allen J. Kihm Department of Earth Sciences Minot State University Minot, ND 58701

ABSTRACT

The Fossil Lake area in Lake County, Oregon, is famous for assemblages of Late Pleistocene vertebrates. However, less is known of the mode of deposition of these fossil specimens. For some time, the area has been known to have been part of a large pluvial lake, Fort Rock Lake. However, the occurrence of large land mammals in layers of sand and gravel-sized deposits interbedded with lake clays and silts, suggests a more complicated depositional history. Over several field seasons, the second and third authors have documented a stratigraphic succession, indicating episodic sedimentation of several disconformity-bounded, fining-upward depositional packages normally containing volcanic ash. A complete fining-upward package consists of a basal gravel grading successively into sand and silt and/or clay, resulting from decreasing energy upward as flooding wanes and waters become deeper within the lake basin. The first author undertook textural analyses of samples from these depositional packages to arrive at scenarios for depositonal environments. One of the documented episodes is a major vertebrate fossil producer. This episode includes a gray-brown sandstone overlain by a claystone layer; the textural curve of the gray-brown sand most closely resembles that of the sandy bedform ar- chitecture, typically deposited by flash floods of ephemeral streams in the upper flow regime. The thick, bedded clay indicates deposition in the deeper, quieter water of a lake. Strata deposited during the succeeding episode also contain abundant fossil vertebrates. This depositional package overlies an apparent disconformity and consists of gravel within sand and pumice ash suggesting sheet flood deposits, capped by an argillaceous siltstone. This package is overlain by a succession of interbedded sand and silt layers whose predom- inant lithofacies are horizontally bedded sandstones grading into fine silts. These unconformity bounded packages repeat until deposition of a coarser, iron-stained, poorly consolidated sandstone with high-angle, trough cross-bed- 274 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) ding. This sand unit represents a back-beach depositional environment and is overlain by a claystone, representing a return to deeper lacustrine conditions. Overall, the Fossil Lake area represents episodic sedimentation, resulting from ephemeral lake development. Complete episodic packages are disconformity bounded and consist of a gravel or sand grading up into a silty claystone; a succession representing the transformation of high-energy sheet flood deposits into low-energy lacustrine deposits. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 275

FOSSIL VERTEBRATES OF THE NIOBRARA FORMATION IN SOUTH DAKOTA

James E. Martin and Bruce A. Schumacher Museum of Geology SD School of Mines & Technology Rapid City, SD 57701

David C. Parris and Barbara Smith Grandstaff Bureau of Natural History New Jersey State Museum Trenton, NJ 08625

ABSTRACT

The Niobrara Formation through the mid-continent has been classically known for its assemblage of fossil fish, birds, and reptiles. Within this area, the Niobrara chalk of South Dakota has received relatively little attention. Only a few fish, a mosasaur, and a plesiosaur have been described from this marine formation heretofore. With additional investigations by the authors, the known vertebrate assemblage from South Dakota now includes: Squalicorax falcatus, Squalicorax kaupi, Cretoxyrhina mantelli, Echinorhynus, Ptychodus mortoni, Pachyrhizodus caninus, Pachyrhizodus minimus, Protosphyraena gladius, Protosphyraena nitida, cf. Apateodus, Ichthyodectes ctenodon, Xiphactinus audux, Bananogmius evolutus, Cimolichthys nepaholica, Stratodus apicalis, Saurodon leanus, Saurocephalus lanciformis, Enchodus shumardi, Enchodus petrosus, Enchodus gladiolus, cf. Lophochelys, Toxochelyidae, Polycotylus latip- innus, Pteranodon, Clidastes propython, Platecarpus tympaniticus, Tylosaurus proriger, cf. Ichthyomis, and Hesperomis regalis. Most of these taxa represent new records from the Niobrara Formation of South Dakota and attribute an early Coniacian to early Campanian age to the unit. The type specimens of Enchodus shumardi Leidy 1856 and Cladocyclus occidentalis Leidy 1856 (a taxon based upon fish scales which are most likely those of Ichthyodectes ctenodon) are from the Niobrara Formation of South Dakota and may have been collected from deposits near the Cheyenne River.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 277

ADDITIONAL QUATERNARY VERTEBRATES FROM THE CHAMBER OF LOST SOULS, WIND CAVE NATIONAL PARK, SOUTHWESTERN SOUTH DAKOTA

James E. Martin and Ruth Anderson Museum of Geology SD School of Mines & Technology Rapid City, South Dakota 57701

ABSTRACT

From deep within Wind Cave, vertebrate remains have been collected by Museum of Geology personnel. These specimens were derived at least 150 m from the nearest present opening, and no remains were encountered anywhere between current openings and concentrations in the Chamber of Lost Souls. The site is only 18.3 m from the surface, suggesting that a former opening in- to the Chamber of Lost Souls could explain the vertebrate accumulation. A rather unusual assemblage including a gastropod (snail), ?urodele (sala- mander), Pseudacris (chorus frog), Bufo (toad), passerine bird, hawk, ?owl, Sorex (shrew), Myotis and Eptesicus (bats), rabbit, Urocyon (gray fox), ?chip- munk, Marmota (woodchuck), Spermophilus (ground squirrel), ?grasshopper mouse, Peromyscus (white-footed mouse), Neotoma (woodrat), Cleithrionomys (redback vole), Microtus (meadow vole), Odocoileus (deer), and Bison. Nearly all the larger bones exhibit evidence of rodent-gnawing, and most are the result of Neotoma, the woodrat. At least some of the elements appear to have been derived from a woodrat (packrat) nest. The occurrence of raptors within the assemblage suggests that some small bones may be derived from pellets. Other bones appear to be derived from creatures which inhabited and expired within the cave. Reworking of most skeletal elements by water perco- lating from above is common.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 279

PRELIMINARY STUDIES ON DISTRIBUTION OF PHYLLACHORA ON POACEAE

A.C. Gabel Department of Biology Black Hills State University Spearﬁsh, SD 57799

L.H. Tiffany Department of Botany Iowa State University Ames, Iowa 50011

M.L. Gabel Department of Biology Black Hills State University Spearﬁsh, South Dakota 57799

ABSTRACT

In this study we collected and identiﬁed species of Phyllachora on grasses from mid-continent United States grasslands and from the pampa of Ar- gentina to determine distribution of the fungi and hosts. Eight species of Phyl- lachora were collected on thirty-six different grass species from mid-continent United States grasslands, with P. luteomaculata on Andropogon gerardii the most frequently collected species. Eight species of Phyllachora have been iden- tiﬁed on eleven different species of grasses from the pampa of Argentina. P. minutissima on species of Paspalum was the most frequently collected species from the pampa. All Phyllachora species and hosts are new records for Ar- gentina. P. cynodontis, P. eragrostidis, P. luteo-maculata and P. vulgata were collected from both countries. Muhlenbergia and Paspalum were hosts in both North America and South America, but all North American host species were different from South American host species.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 281

STUDIES ON A SOUTH DAKOTA POPULATION OF SOYBEAN CYST NEMATODE, HETERODERA GLYCINES

James L. Jones and James D. Smolik Department of Plant Science South Dakota State University Brookings, South Dakota 57007

ABSTRACT

The soybean cyst nematode (SCN), Heterodera glycines Ichinohe, is the most serious disease threat to soybeans throughout much of the soybean growing area of the United States. Yield losses in the North Central Region of the United States are estimated at 50 million bushels per year. The first report of SCN in South Dakota was in Union County in 1995. A 1996 survey identified SCN in several additional locations in Turner County. Crop production practices in SCN-infested areas often include corn-soybean or soybean-soybean rotations which are conducive to rapid increases in SCN populations. A small-plot test and field-length test strips using resistant and susceptible soybean varieties were established in an SCN-infested field in Union County. The objectives of the study were to determine: population dynamics of the soybean cyst nematode over the growing season on a susceptible and resistant soybean variety; SCN population changes in SCN-infested fields planted to corn following soybean; and the effects of SCN on soybean yields and yield components, including number of pods, number of seeds, seed weight per plant, and plant height. Plots planted to the resistant variety Bell and susceptible variety Hardin were sampled bi-weekly to determine population dynamics over the growing season. Five corn fields infested with SCN were also sampled bi-weekly to determine SCN populations on a non-host crop. A significant reduction in SCN populations was measured in plots planted to the resistant soybean variety while populations in plots planted to the susceptible variety increased over the growing season. Populations in SCN-infested corn fields declined over the growing season, although the decline was not as great as that recorded in plots planted to the resistant soybean variety. Yields of resistant soybean varieties were significantly greater (44 to 100 %) than those of susceptible varieties. Yield components of resistant varieties were also generally greater than those of susceptible varieties. In conclusion, the identification of SCN in Union and Turner Counties represents a serious problem for soybean producers in southeastern South Dako- ta. This research indicated that planting susceptible soybean varieties in the presence of SCN will result in increased SCN populations and significant yield loss. Rotation to corn for one year did not greatly reduce SCN populations.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 283

PRELIMINARY EXAMINATION OF THE EFFECT OF DIRECTION OF SLOPE OF FIELD ON GROWTH AND YIELDS OF CORN PLANTS IN NORTHWESTERN IOWA

Troy McKenney and Donna Hazelwood Dakota State University Madison, SD 57042

ABSTRACT

This study compared two fields cropped to corn (Zea Mays L.) during summer 1996. The purpose of this study was to determine if direction of slope of field would have an effect on growth or yield of corn plants. Both fields have been in a corn-soybean rotation and have been managed as no-till for the past four years. Pre-planting fertilizer was applied, and Extrazine was applied post planting, but insecticide was not used. The field that slopes from south to north was planted May 7; the field that slopes from east to west was planted May 9. Both fields were harvested October 14. In each field, fifteen plants were selected at random and were measured periodically. Measurements taken included plant height, date of anthesis, and development of ear. In addition, air temperature (daily highs and lows), wind direction and wind speed were recorded. Insect samples were collected and stored for analysis. Incidence of Ustilago mayids, common corn smut, was also recorded. From both fields, soil samples collected July 23 were sent to the chemistry lab at the University of Nebraska at Lincoln. Nitrogen, potassium, and phosphorus levels were reported to be similar in the two fields, and there were sufficient levels of nutrients in both fields for adequate growth. Preliminary findings include the following similarities for the two fields: 1) erosion was not evident in these no-till fields, 2) although these fields were not sprayed for Pyrausta nubilahs, the common corn borer was not a damaging factor, 3) common corn smut was detected, but at a low level of incidence, and 4) the common grasshopper was detected at a high level of incidence on plants in the fields. Preliminary results indicate differences in rate of growth and yield. Plants in the field that sloped from east to west gave a higher yield, but in the field that slopes from south to north corn grew at a faster rate. This study will be continued in summer l997

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 285

BLACK DOT FLEA BEETLE GENETIC DIVERSITY AS MEASURED AT TWO LOCI—MDH AND PGM

Mark A. Brinkman and Sharon A. Clay Plant Science Department and Nels H. Granholm Department of Biology/Microbiology South Dakota State University Brookings, South Dakota 57007

ABSTRACT

Leafy spurge, Euphorbia esula L., is a noxious perennial weed that occurs as isolated infestations covering nearly 1 million ha in North America. Black dot flea beetles, Aphthona nigriscutis Foudras, have been introduced from Hun- gary for biological control of leafy spurge. Since natural dispersal is limited, flea beetles are collected from established insectaries and distributed to leafy spurge infested areas. In South Dakota, the number of adults released to each site has ranged from 100 to 1000. The purpose of this study was to determine the effects of bottleneck and isolation on black dot flea beetle genetic variability. Beetles were obtained from established sites near Pollock, SD, Valley City, ND, and Theodore Roosevelt National Park (ND) in 1996. We analyzed 48 individuals from each site for genetic variability at 2 loci - malate dehydroge- nase (MDH) and phospoglucomutase (PGM). Chi-square analysis was used to determine allele and genotype frequency deviations from Hardy-Weinberg predicted values. One locus was scored at both MDH and PGM. Four alleles were detected at MDH; three were found at PGM. Rare alleles (number of individuals < 4) in a specific population were pooled with the least common other allele prior to statistical analysis. There were few heterozygous individuals (less than 1%) detected at MDH. Therefore, it was difficult to compare populations based on genetic variability at MDH. However, more than 45% of individuals from the Valley City, ND site were heterozygous at PGM. Over 37% of individuals from Theodore Roosevelt National Park were heterozygous at PGM. More than 93% of individuals from the Pollock site were homozygous at PGM. With “A” and “B” allele frequencies equal, 50% of the individuals in a population in Hardy-Weinberg equilibrium would be expected to be heterozygous. Valley City and Theodore Roosevelt National Park population genotype fre- ≤ ≥ quencies did not deviate significantly (X2 2.48, df = 2, P 0.2892) from Hardy-Weinberg predictions. The number of heterozygous individuals at Pol- lock was significantly (X2 = 22.9, df = 2, P < 0.0001) lower than Hardy-Wein- berg predictions. Population growth at the Pollock site may have been too slow to offset the effects of inbreeding. Heterozygosity in certain insect species has been associated with higher fecundity, male mating capacity, and mobility. Since heterozygosity is generally associated with fitness or vigor, there may be 286 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) a positive correlation between flea beetle heterozygosity and leafy spurge control. Accordingly, the Pollock population may benefit from introduction of individuals from the Valley City site. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 287

ALLELOPATHY IN ECHINACEA ANGUSTIFOLIA D.C. ROOTS

Peter A. Jauert and R. Neil Reese Department of Biology/Microbiology South Dakota State University Brookings, SD 57007

ABSTRACT

The purpose of this study was to examine the allelopathic effects of compounds of different molecular weights taken from Echinacea angustifolia D.C., the common purple coneflower. Roots were collected from ten plants from five geographic locations and from five fresh plants that were grown in the green- house. Water extracts were made from ground dried roots and crushed fresh roots. These extracts were then separated into a high and a low molecular weight fractions using a 10,000 dalton cut-off ultrafilter (Phenomenex, Tor- rence, CA ). Lettuce seeds (Lactuca sativa) were germinated with the high molecular weight, low molecular weight, and a crude extract from each plant. After four days the seed germination was counted and the root lengths were measured. Analyses of variance of the data were made using the general linear model procedure (SAS Institute, Cary, NC). The crude and the low molecular weight extracts exhibited the most severe allelopatic effects. High molecular weight fractions did not cause a significant inhibition of lettuce seed germination. Partitioning of the active fractions with chloroform and methylene chloride are in progress to isolate the biologically active components.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 289

ALLELOPATHIC POTENTIAL OF ECHINACEA ANGUSTIFOLIA D.C.'S ROOT EXTRACTS

Kimberly Piechowski and R. Neil Reese Department of Biology/Microbiology South Dakota State University Brookings, SD 57007

ABSTRACT

The narrow-leaf purple coneflower, Echinacea angustifolia D.C., is a perennial native to North America. E. angustifolia is found on the dry prairies of Texas to Saskatchewan, west to the Rocky Mountains and east into Min- nesota (Foster, 1991). E. angustifolia has been shown to produce many biologically active compounds (Viles and Reese, 1995). The water soluble compounds of E. angustifolia have shown an allelopathic effect on lettuce (Lactu- ca sativa), switchgrass (Panicum virgatum), and prairie dropseed (Sporobolus heterolepis) (Viles and Reese, 1995). The allelopathic potential of ethanol soluble compounds from E. angustifolia roots were examined across five different populations from South Dakota, North Dakota, Wyoming, Kansas, and Ne- braska. A common allelopathy bioassay was performed using lettuce seeds germinated in the presence of ethanol extracts that were brought to dryness and resuspended in water. Analysis of the seed assay was completed by measuring percent germination and shoot elongation. Concentrated root ethanol extracts were analyzed by reverse phase HPLC using a C-18 column. HPLC revealed ten significant peaks in all the populations. Using multiple regression analyses (SAS Institute, Inc., Cary, NC), a correlation was made between germination, shoot elongation, and peak area. These results showed that peak numbers three and six contribute to the allelopathic effects of E. angustifolia‘s roots. Further research is being conducted to collect, purify, and identify the allelopathic compounds of E. angustifolia.

REFERENCES

Foster, S. 1991. Echinacea: Nature’s immune enhancer. Healing Arts Press, Rochester, VT. Viles, A. and R. Neil Reese. 1996. Allelopathic Potential Of Echinacea angustifolia D.C.. Environmental and Experimental Botany 36:39-43.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 291

THYROXINE EFFECTS ON INVERTEBRATE DEVELOPMENT

Leland G. Johnson, Jessica E. Kullbom, Stephanie L. Priesz, Shana N. Strand, and Katherine M. Faber Augustana College Sioux Falls, SD 57197

ABSTRACT

Thyroxine and other thyroid hormones are unique among vertebrate regulatory molecules because they are amino acid derivatives that contain iodine, an element that is very scarce in the environments of many vertebrates. This dependence on a scarce resource raises questions about the evolutionary origins of thyroid hormones. Metabolic effects of thyroid hormones in birds and mammals are very familiar, but the oldest and most basic thyroid hormone functions in vertebrates may be in promoting differentiation and maturation during development. Thus, a search for thyroid hormones’ evolutionary roots might begin with comparative examination of possible roles these molecules play in invertebrate development. For some time the only evidence of developmental effects of thyroxine in invertebrates came from Spangeberg's reports (1974; 1984) that thyroxine affects strobilation and statolith development in a cnidarian, the jellyfish Aurelia. More recently, there have been reports that thyroxine promotes adult rudiment development in late larval stages of sea urchins (Chino, et al., 1994) and ac- celerates late larval development of the crown-of-thorns starfish (Johnson and Cartwright, 1996). We have investigated thyroxine effects in several freshwater invertebrate species and have tested the effect of thyroxine on earlier stages of sea urchin development. Planarian worms Dugesia dorotocephala were decapitated by cutting across the body just posterior to the auricles, and we followed regeneration of worms in spring water (controls) and in several concentrations of thyroxine in spring water (experimentals). We observed no significant difference in regeneration of control and experimental planarian worms. We did several preliminary experiments with thyroxine effects on regeneration in Hydras severed below the hypostome using iridectomy scissors. After 22 hours, we found regeneration was significantly more advanced in thyroxine-treated green Hydras than in controls, and thyroxine associated accelera- tion was still evident up to 30 hours. Later, regeneration in controls caught up. We did not find such a thyroxine effect in brown Hydras. Subsequent results with green Hydras have been somewhat variable, but we think more research is warranted. We investigated thyroxine effects on early embryonic development in the sea urchin Strongylocentiotus purpuratus, and found that thyroxine did not af- 292 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) fect developmental rate from fertilization to hatching or from the mesenchyme blastula stage to assembly of the larval skeleton. In the sea urchin Evechinus chloroticus (Johnson, unpublished), thyroxine accelerated late larval development, as it had in other echinoderms, but actually slowed development from fertilization to the 4armed pluteus stage, and from the 4-armed stage to the 6- armed stage. It is too early to draw firm conclusions about thyroxine effects on invertebrate development, but several premises might aid interpretation of our results and suggest directions for future investigations. 1) The divergence of Cnidari- ans from other animals is regarded to have been very ancient. Thus, thyroxine effects on Cnidarians may have arisen independently of effects observed in more complex invertebrates (echinoderms) and in chordates. 2) Echinoderms and chordates share membership in the deuterostomes, one of the two super groups of modern complex animals. The other group, the protostomes, includes annelids, molluscs, and arthropods. Discovery of thyroxine effects on development of protostome animals would imply that thyroxine's evolutionary roots are very deep, and further investigation might then suggest how ancient the origins of its regulatory functions might have been. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 293

AGOUTI GENE REGULATION OF THIOL (CYS AND GSH) CONCENTRATIONS IN LIVERS OF MICE

J. T. Brunz, R. N. Reese, and N. H. Granholm Department of Biology/Microbiology South Dakota State University Brookings, SD 57007

ABSTRACT

Thiols like cysteine (cys) and glutathione (GSH) are thought to be important regulators of melanogenesis; they promote yellow pigment synthesis (phaeomelanogenesis) by inhibiting tyrosinase activity and by complexing with dopaquinone to form phaeomelanin intermediates (e.g., 5-S-cysteinyldopa). The agouti gene may influence thiol metabolism within hair follicles as well as extrafollicular tissues (J Invest Dermatol 106:559,1996). One breakdown enzyme of GSH, gamma-glutamyltranspeptidase, exhibited no genotype differences within regenerating hair bulbs of agouti (AwJ/AwJ), black (a/a), and yellow (Ay/a) mice (Pigment Cell Res., Suppl. 5:57,1996). The liver is a major source of thiols; liver thiol depots may supply thiols to regenerating hair follicles. The purpose of this study was to determine if the agouti gene influences thiol metabolism. If so, we predict genotype-related patterns of liver thiol metabolism. We assessed the concentration of liver thiols (cys and GSH) via HPLC in three different genotypes—two mutants, Ay/a and a/a and one control, AwJ/AwJ within three distinct age/treatment groups—60 day unplucked (60-U), 60 day plucked/9 day hair bulb regenerated (60-P), and 120 day unplucked (120-U). GSH exhibited a genotype-related concentration hierarchy (AwJ/AwJ > Ay/a > a/a) within 120-U mice. In general, however, there were no striking genotype-specific changes in either cys or GSH levels. In contrast, there were significant age- and plucking-related elevations in GSH but not cys concentrations. Although these data do not support a major role for the agouti gene in thiol metabolism of liver, they do suggest causal relationships between aging, plucking, and liver thiol metabolism. Funded by SDSU-AES089, FOE Ehrmann Cancer Fund, and NIH-AR42757.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 295

DES-ACETYL-ALPHA MELANOCYTE STIMULATING HORMONE AND MELANOGENESIS IN B16 MELANOMA CELLS

P. Ramasastry and N.H. Granholm Department of Biology & Microbiology, South Dakota State University Brookings,S.D. 57007

ABSTRACT

Des-acetyl-alpha MSH (dMSH) is a cleavage product of propiome- lanocortin, a polypeptide hormone precursor synthesized in the pituitary gland. Post-translational acetylation of dMSH results in the production of mono-acetyl- alpha MSH (mMSH), the melanogenically active form of MSH, i.e., the form that stimulates production of eumelanin (black melanin). Ay-induced aberrations in the acetylation of dMSH may be partially responsible for yellow hair of Ay/a mice, because excess dMSH may induce yellow pigment synthesis in pigment cells. The purpose of this study was to test the melanogenic potentials of both dMSH and mMSH to determine if excess dMSH does inhibit eumelanogenesis and promote phaeomelanin synthesis. B16 melanoma cells were treated in culture with mMSH alone (0.2 uM), mMSH (0.2 uM) together with increasing amounts of dMSH (2.0 to 10.0 uM), and dMSH (0.2 uM) alone; we measured the effects of these treatments on four parameters of melanogenesis: cAMP concentration, two tyrosinase activities (TH and DO), and total melanin. In- creasing levels of dMSH in combination with mMSH (0.2 uM) does not cause a reduction in melanogenic parameters and therefore does not promote phaeomelanin synthesis. Furthermore, dMSH alone enhances cAMP, TH, DO, and melanin synthesis at levels comparable to equivalent concentrations of mMSH. These data show that dMSH and mMSH are essentially equivalent in their melanogenic potential. Therefore, it is unlikely that dMSH plays any major role in the regulation of phaeomelanin synthesis in B16 cells in culture. Supported by SDSU-ASE089, FOE Ehrmann Cancer Fund, and NIH-AR42757.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 297

SERUM ESTRADIOL LEVELS IN THE LETHAL YELLOW MOUSE

Maureen Diggins, Dustin Dierks and Julia Spiry Department of Biology Augustana College Sioux Falls, SD 57197

Nels Granholm Department of Biology and Microbiology South Dakota State University Brookings, SD 57007

ABSTRACT

The lethal yellow mutation occurs at the agouti locus on chromosome 2 in the mouse. Mice homozygous for the lethal yellow gene (Ay/Ay) die during development. Heterozygous mice (Ay/a) live and exhibit a collection of characteristics known as the lethal yellow syndrome (LYS). The LYS includes a yellow coat, increasing obesity with age, and progressive infertility. In previous studies, we have reported that the lethal yellow mice (Ay/a), when compared with lean black controls (a/a), exhibit lengthened estrous cycles, depressed ovulation rates, and depressed levels of follicle stimulating hormone (FSH). Our hypothesis for this study is that obese yellow females have increased levels of serum estrogens, both estradiol from ovarian follicles and estrone from adipose cells. Elevated levels of serum estrogens act upon the hypotha- lamus and anterior pituitary, suppressing secretion of FSH. We have adapted a double antibody procedure for radioimmunoassay (RIA) of human serum estradiol for use in mice (Larson, et al., Proc. DD Acad. Sci., 1994). One hundred thirty mice (half lethal yellows and half black controls) were studied. The mice were divided into seven different age groups: 60, 90, 120, 180, 210, 240, and 270 days of age. Females were placed in divided cages with males to induce cycling. Stage of estrous cycle was determined by examination of cell types in vaginal smears. Serum was collected for analysis of estradiol levels in the afternoon of estrous. Estradiol was extracted from the serum with diethyl ether. The RIA for estradiol was conducted using the mod- iﬁed kit mentioned above. The kit was supplied by Diagnostics Products Cor- poration. Serum estradiol levels were indeed higher in yellow mice than in black controls. This was particularly true for mice ages 60-240 days. In lethal yellow mice older than 240 days, serum estradiol levels dropped below those of black controls. When mice from all age groups were compared, the estradiol levels of the obese yellow mice were signiﬁcantly higher than estradiol levels of the lean black controls. Thus elevated levels of serum estrogens in obese mice may be inhibiting secretion of FSH from the anterior pituitary. This may be the 298 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) cause of the depressed ovulation rates, FSH levels, and reproductive success of obese lethal yellow mice. This study was funded by USDA-NRICGP Grant No. 95-37208-2245. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 299

MOLECULAR CHARACTERIZATION OF THE BOVINE HOMOLOG OF THE AGOUTI GENE

M.D. Johansen, R.R.R. Rowland, and N.H. Granholm Department of Biology and Microbiology South Dakota State University Brookings, SD 57007

ABSTRACT

Overexpression of the lethal yellow (Ay) allele in the mouse causes a collection of systemic defects known as the lethal yellow syndrome. These disor- ders include obesity, diabetes, infertility, compromised immunity, increased susceptibility to cancer, and other defects of great interest to both agricultural productivity and human health. Our purpose was to detect, isolate, and characterize the bovine agouti homolog. Bovine genomic DNA was isolated, digested, and hybridized with 32P-labeled murine agouti cDNA. Polymerase chain reaction (PCR) was used to amplify regions of the agouti coding region. PCR primers were designed using human agouti gene sequences. Southern hy- bridizations indicate the existence of a bovine agouti homolog. We constructed a partial restriction enzyme map of the bovine agouti region. PCR of bovine genomic DNA yielded products approximately the same sizes as those of the human and mouse. These products are being cloned and prepared for sequence analysis. Characterization of the bovine agouti homolog may ultimately allow cattle producers to genetically select for rapid growth, high protein/low fat carcasses, enhanced reproduction, and greater vigor. Funded by SDSU-AES (HD089), Eagles’ Ehrmann Cancer Fund, and NIH(AR42757).

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 301

BOVINE VIRAL DIARRHEA VIRUS INFECTS CAMELID CELL LINES

A.A. Ali1,3, A.A. Salama1, T.A. Aboellail2,3 and C.C. Chase3

1Department of Microbiology Facility of Veterinary Medicine Zagazig University, Zagazig, Egypt

2Department of Pathology Facility of Veterinary Medicine Cairo University, Cairo, Egypt

3Department of Veterinary Science, ADRDL South Dakota State University Brookings, SD, 57007

ABSTRACT

The recent isolation of bovine viral diarrhea virus (BVDV) from Camelidae (dromedary camel and llama species) opens the door for investigating the biology of the virus in cell lines derived from those species. Growth characteristics and kinetics of BVDV infectivity were measured in camelid cell lines (camel turbinate and llama kidney) in comparison to bovine cell lines (bovine turbinate and Madin Darby bovine kidney). Four BVDV strains, two cytopathic (CP: Type I, NADL and Type II, 125) and two non cytopathic (NCP: Type I, New York-1 and Type II, 890 ) were used to infect the cell lines at multiplici- ty of infection (MOI) of 5. The affinity of the virus to different cell lines was followed up and screened by direct immunofluorescence methods using BVDV-fluorescein isothiocynate conjugate before the development of the cytopathic effect. After the first passage, the cell lysates were harvested and used to infect susceptible cells in serial passage with subsequent virus titration of each passage. The inoculum was decreased by one MOI per passage and then maintained at a MOI of 1 from fifth to seventh passage. After the seventh passage,the infectivity of NADL and 125 strains adapted on camel turbinate (CT), llama kidney (LMK-I) or bovine turbinate (BT) cells was performed on MDBK cells. One step growth curves of cytopathic strains was done on both camelid and bovine cell lines at MOI of 1. Non cytopathic strains (New York-1 and 890) were passaged in both CT and LMK-I for three blind passages and detected by immunofluorescence. Viral replication of BVDV biotypes (CP and NCP) in camelid cells was low at the first passage. With NADL strain, the cytopathic ef- 6 fect first developed at the third passage in CT with a titer of 9x10 TCID50/ml, while in LMK-I, the cytopathic effect developed at the fifth passage with a titer 6 of 5x10 TCID50/ml. With strain 125, the cytopathic effect developed at the 7 fourth passage in CT with a titer of 1.6x10 TCID50/ml and at the sixth passage 302 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997)

6 in LMK-I with a titer of 1.6x 10 TCID50/ml. The cytopathic effect reached its maximum with complete destruction of the cell at the seventh passage with both camelid cell lines. The growth kinetic curve of BVDV CP biotypes was measured for 96 hours. After adsorption, the titer of NADL strain in CT was 3 2 1.6xl0 TCID50/ml and in LMK-I was 9x10 TCID50/ml, and of strain 125, the titer 3 2 in CT was 2.8x10 TCID50/ml and in LMK-I was 5x10 TCID50/ml. The maximum 6 6 titer of NADL strain in CT was 9x10 TCID50/ml and in LMK-I was 5x10 6 TCID50/ml , and that of 125 strain in CT was 9x10 TCID50/ml and in LMK-I was 6 1.6x10 TCID50/ml after inoculation for 72 hours. Adaptation and propagation of NCP strains was clear after three blind passages. In conclusion, the BVDV strains grew to higher titers in camelid cell lines than bovine cells and grew to a higher titer with more progressive cytopathic effect in turbinate cells than kidney cells. This work indicates that BVDV grows well in camelid cells. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 303

USE OF DIGITONIN TO DIFFERENTIALLY STRIP THE TEGUMENTAL MEMBRANE SURROUNDING THE STROBILA AND BLADDER REGIONS OF TAENIA TAENIAEFORMIS STROBILOCERCI

E.J. Olson, D.J. Robison, S.R. Duimstra and M.B. Hildreth Departments of Biology/Microbiology and Veterinary Science South Dakota State University Brookings, SD 57007

ABSTRACT

The strobilocercus or larval stage of the cat tapeworm, Taenia taeniaeformis, is surrounded by an epithelial, syncytial tegument composed of a single, distal sheet of cytoplasm that completely encompasses each worm. The function of the cestode tegument is analogous to intestinal enterocytes, both being responsible for absorbing nutrients while also resisting various intestinal proteases and lipases. The tegument covering the anterior portion of the strobilocercus (scolex-strobila) is resistant to intestinal digestion; however, the tegument surrounding the posterior end of the strobilocercus (bladder), is susceptible to digestion. Differences in the susceptibility of these two regions may result from differences in the membrane composition from each region. Mills et al. (1984) described a method for isolating and characterizing the tegumental membrane from entire T. taeniaeformis strobilocerci by incubating the worms in a 0.1 % (w/v) digitonin solution for 10 minutes, and in 0.85 % saline for 10 minutes, followed by vortexing for 10 seconds. The purpose of this study was to evaluate the use of this stripping technique for isolating the teguments from the strobila and the bladder regions of T. taeniaeformis strobilocerci. Tae- nia taeniaeformis strobilocerci were cut into halves (strobila and bladder), and each of these components were incubated in the digitonin solution for either 10 or 20 minutes. Scanning electron micrographs (SEM) of the stripped carcasses suggest that only some of the tegument was stripped off the strobila after 10 minutes in digitonin, but most of the tegument was removed from the sttrobila after 20 minutes. Increasing the incubation time to 20 minutes for the strobila allowed for much more of the tegument to be removed. Following an incubation period of 20 minutes in digitonin, virtually all of the bladder tegument had been removed; however, most of the bladders had collapsed and some damage to the underlying basal lamina occurred. This damage may enable proteins and lipids from the underlying tissues to contaminate the tegumental membrane fraction. An incubation time of 10 minutes caused most of the bladder tegument to be removed (even without vortexing), but better preserved the basal lamina and eliminated bladder collapse. Our results suggest that doubling the stripping time-period described by Mills et al. (1984) improved removal of the strobilar tegument, and demonstrates an additional dif- 304 Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) ference between the teguments from the two regions. [Supported in part by NSF-EPSCoR Project # OSR-9452894.]

REFERENCE

Mills, G.L., S.C. Cooley and J.F. Williams. 1984. Lipid and protein composition of the surface tegument from larvae of Taenia taeniaeformis. J. Parasit. 70(2):197-207. Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 305

MATHEMATICAL ANALYSIS OF A MODEL FOR THE GROWTH OF A SINGLE SPECIES CELL: RIBOSOME DEPENDENCE MODEL

A. S. Elkhader Northern State University Aberdeen, SD

ABSTRACT

A chemostat is a laboratory device used in ecological studies of microor- ganisms. In 1996 H. Yun, J. Hong, and H. Lim [YHL] developed a mathematical model that describes the growth of species in an ideal chemostat. An important feature of this model is the dependence of the cell growth rate function on the ribosome content of the cell. The variables in the model are: the nutrient, the cell mass, and the total ribosome concentration in the cell. The parameters in the model are: the washout rate, the input concentration rate, the maximum growth rate, and the half saturation rate for the cells and the ribo- somes. The model consists of three ordinary differential equations of nonlinear type. Numerical simulation by YHL showed an increase in the ribosome is followed by an increase in the cell mass. However, no mathematical treatment of the model was given. This work gives a mathematical analysis of the YHL model. The analysis supports the conclusions of YHL. The analysis starts by presenting basic properties of the model. This includes establishing boundedness and nonnegativity of solutions. Then the conservation principle was used to reduce the three-dimensional system into a planar system. The resulting two-dimensional system has three possible equilibria, only one of which has a nonzero cell mass. The local stability of these equilibria is established. A Poincare-Bendixon Theorem using the Dulac criterion eliminates periodic orbit solutions and cycles of steady states in the planar region. Thus the global stability of the nontrivial equilibrium point is proved. Numerical examples support the conclusions of this analysis.

REFERENCE

Yun, H. S, Hong, J., Lim, H. C. 1996. Regulation of Ribosome Synthesis in Es- cherichia Coli: Effects of Temperature and Dilution Rate Changes. Biotech- nology and Bioengineering, vol. 52, Pp.615-624.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 307

MATHEMATICAL MODELS FOR THE VISUALIZATION OF JUGGLING

Richard D. Simmons and Douglas J. Peters College of Natural Sciences Dakota State University Madison, SD 57042

ABSTRACT

A mathematician and a physicist joined forces in the experimental and theoretical investigation of the motion of objects being juggled. From the analysis of the motion, practical and realistic mathematical models for juggling were developed. The mathematical models were ultimately used in an interactive computer program for the visualization of juggling. The user can choose or custom- design the juggling objects and the juggling pattern, and the program produces a three-dimensional animation of the user's request. The video analysis software package World-in-Motion was used to explore the kinematics of actual juggled objects by tracing the coordinates of points on the objects. The motion was further analyzed using Newtonian mechanics, and mathematical models were developed. The software package Maple V, Release 4 for Windows, was used to create the visualizations. Further rendering of the visualizations beyond the scope of the graphics capabilities of Maple V would produce more aesthetically pleasing results but was not a part of this project. In modeling the juggling process for an interactive animation, there are a number of desirable characteristics; for example, the ability to change the type of pattern, the dimensions of the pattern, the speed of the juggling, the number of juggled objects, and the type of objects. All of these features were employed within the Maple V program. Kinematics was used to understand the nature of the juggling motions and to add realism to the animations. Factors other than realism were considered in the design of the models; for example, smoothness, symmetry, and mathematical simplicity. A simple model is easier to program and does not take as much computer time to run. The juggling experience of the mathematician/author was also used in the development of the models. This project is part of broader collaborative efforts in the College of Natu- ral Sciences at Dakota State University. A goal of these efforts is to use technology in the integrated study of mathematics and science. Another goal is the development of original activities and curriculum materials for active learning. It was our goal in this project to create a product that is both educational and fun to use. The physics, geometry, and computer graphics come together in a virtual world where anyone can experience a part of the joy of juggling; a world where the only physical limitations are those of the hardware and software.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 309

MODIFICATION OF CARBON ELECTRODE SURFACES

Royce C. Engstrom, Brian D. Lamp and Betsy B. Ratcliff Department of Chemistry University of South Dakota, Vermillion, SD 57069

ABSTRACT

Glassy carbon electrodes have been subjected to anodization, followed by modification on a microscopic scale to create domains of electrochemical activity in a matrix of rich functionalities available for derivatization. The microactivated sites have been characterized with respect to their physical morphology, electrochemical kinetic behavior and chemical composition. Physical characterization was done with atomic force microscopy, which indicated that the mechanism of formation of the microactivated sites involves hydroxide etching of the anodization layer, but that the mechanism is more complex than simple hydroxide-induced dissolution. Furthermore, the anodization layer has a thickness of approximately 80 nm with sharply defined boundaries. Kinetic behavior was investigated with electrogenerated chemiluminescence imaging, in which the potential dependence of light generation showed faster kinetics at the microsites than at freshly polished glassy carbon. Electrochemical microscopy using the Fe(II)/Fe(III) system showed that when first formed, the microsites are likely devoid of carbonyl functionalities, but that slow oxidation occurs to form carbonyls.

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 311

CREATING AND IMAGING MICROSTRUCTURES ON SURFACES

Miles Koppang, Robert Carson, Paul Verbanac and Wenchun Chao Department of Chemistry, University of South Dakota Vermillion, SD 57069

ABSTRACT

We have investigated self-assembled monolayers (SAMs) of alkyl thiols on gold for their use in modification of an electrode surface. By derivatization of a pendent functional group of the SAM, a variety of functionalities or chemicals, such as enzymes, could be attached to the surface. Instead of derivatizing all pendent groups, our goal is to restrict derivatization to pendent groups located within microscopically-sized domains on the surface. We have used two approaches to modify the SAM surface, 1) removal of SAMs and 2) derivatization of pendent amines of SAMs within micrometer-sized regions. In the first, 2+ an electron transfer mediator, tris(2,2’-bipyridyl)ruthenium(II) (Ru(bpy)3 ), was used to create micrometer-sized “holes” or “defects” in an SAM on a polycrys- talline gold electrode by a technique we refer to as “localized-reductive des- 2+ orption.” In this technique, Ru(bpy)3 was reduced at a carbon microelectrode µ 2+ positioned within 2 = m of the SAM surface. The reduced Ru(bpy)3 diffused to the SAM, transferred its electrons to the adsorbed thiols causing the thiols to desorb within a microscopic domain. Electrogenerated chemiluminescence imaging (ECL) with luminol in alkaline peroxide was used with a microscope- based imaging system to observe the deliberately-produced “defect” sites or “holes” in the alkyl thiol coating. The effectiveness of the reductive desorption decreased significantly when oxygen was present. Reduction of oxygen at -1.0 V vs. Ag/AgCI, prior to mediator generation at -2.0 V, yielded more effective thiol desorption. The precathodization step reduced dissolved oxygen pre- 2+ venting the oxygen from scavenging electrons from the reduced Ru(bpy)3 . In the second approach, pendent amine groups of SAMs were derivatized with a fluorescent tagging reagent, dansyl chloride. Spatially-resolved imaging of the subsequent fluorescent monolyer was performed with chemiluminescence imaging using a peroxylate ester, bis(2,4,6-trichlorophenyl) oxalate (TCPO) and hydrogen peroxide. The fluorophoric tags were excited through chemical energy transfer from the unstable, dioxetanedione intermediate, and the fluores- cence was observed through the microscope-based imaging system.

South Dakota Academy of Science

1997 Junior Academy Best Presentation

Proceedings of the South Dakota Academy of Science,Vol. 76 (1997) 315

STUCK WITH DUCK YUCK?

Justin L. Herreman Stevens High School, Rapid City, SD

Canyon Lake, a twenty-six acre lake located in Rapid City, South Dakota, has a large population of ducks and geese that reside year-around within its confines. Last year the city of Rapid City spent over two million dollars on improvements in an effort to improve flow and quality of the water. The purpose of this project was to determine what impact the improvements made on Canyon Lake have on the water-quality of the lake and to determine if the problems of flow were solved. Electrical models of water flow predicted the improvements would not provide adequate flow to flush the bacteria from the lake. Analysis of ice cover patterns on the lake support this prediction. Water flow rate and direction measurements confirm this prediction. Previous experimentation has shown poor lake water-quality in Canyon Lake in the form of coliform and enterrococci bacteria in higher that EPA recommendations for a recreational lake. This research shows bacteria levels to still be nearly double the recommended levels, although a minor (3-5%) decrease in levels of these bacteria was seen. Despite the improvement, it seems Canyon Lake is still,

STUCK WITH DUCK YUCK!