proceedings of the south dakota academy of science volume 85 … · 2017. 7. 12. · proceedings of...

Proceedingsof the

South Dakota Academy of Science

Volume 852006

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

EditorSteven R. Chipps

Terri Symens, Wildlife & Fisheries, SDSUSecretarial Assistant

Tom Holmlund, Graphic Designer

TABLE OF CONTENTS

Consolidated Minutes Of The Ninety-first Annual Meeting Of South Dakota Academy Of Science, Dakota Wesleyan University, Held At Cedar Shore Resort, Oacoma, SD, 7-8 April 2006 ................................................ 1

Report Of The Resolutions Committee ......................................................................... 6Resolution Of The South Dakota Academy Of Science ................................................ 6South Dakota Academy Of Science 2006-07 Executive Council ................................... 72006 Regional Science Fairs .......................................................................................... 8South Dakota Academy Of Science Treasurer’s Report 14-Oct-06 ................................ 8Executive Summary, Proceedings Editor ........................................................................ 9Presidential Address—The End Of Cheap Oil: Positive Or Negative

Effects For South Dakota? Presented By James Sorenson ..................................... 11

Complete Senior Research PapersPresented at the 91st Annual Meeting of the


Eclipse Of The Inner Satellite Of Jupiter. Perry H. Rahn and Jeffrey T. Rahn ..................................................................................................... 21

Nitrate In Rapid City’s Water Supply. Perry H. Rahn ................................................. 31New Information On The Ree Heights Fossil Site, Hand County, South

Dakota. David C. Parris, Doreena M. Patrick and Jared Williams ....................... 43Descriptive Analysis Of Aquatic Invertebrate Communities In Wadeable

And Non-wadeable Streams Of The Northern Great Plains Network. Jill D. Rust and Nels H. Troelstrup, Jr. ................................................................ 49

A Cluster For Clusters: High Performance Computing For Molecular Dynamics Of Large Rare-gas Atomic Clusters. David T. Huebner and Brian G. Moore ............................................................................................. 63

Phosphorus Fertilization Impacts On Wheat Growth And Selenium Bioavailability. Sang H. Lee and J. J. Doolittle .................................................... 73

Vital Signs Monitoring In Our Parks: What To Measure? Nels H. Troelstrup, Jr. and Jill D. Rust ................................................................ 83

N And Water Stress Impact On Hard Red Spring Wheat Yield And Quality. R. Brunner, D. Clay and C. Reese .......................................................... 95

The Evaluation Of Operands And Its Problems In C++. Dan Day and Steve Shum ................................................................................................. 107

Effects Of Grazing On Small Mammal Abundance In Eastern South Dakota. Wesley W. Bouska and Jonathan A. Jenks ............................................. 113

Detection Of Bison/Cattle Hybridization In Custer State Park Breeding Bulls Using Microsatellite And Mitochondrial DNA Markers: Tools For Conservation Management. Cynthia M. Anderson, Traci L. Berger, Forrest Cain and Shane K. Sarver ............................................ 119

Dietary Yeast Culture Supplementation During Initial Rearing Of Three Salmonid Species. Michael E. Barnes, Brian Fletcher, Dan J. Durben and Stuart G. Reeves .......................................................................................... 129

Computer Aided Statistical Analysis Of Satellite Sensor Data Cross- Calibration. Stuart Ness and Daniel Swets ......................................................... 141

Rare And Declining Fishes Of South Dakota: A River Drainage Scale Perspective. Christopher W. Hoagstrom, Cari-ann Hayer, Jason G. Kral, Steven S. Wall and Charles R. Berry, Jr. ...................................... 171

Perennial-warmwater Fish Communities Of The Cheyenne River Drainage: A Seasonal Assessment. Christopher W. Hoagstrom, Austin C. Dewitte, Nathan J.C. Gosch and Charles R. Berry, Jr. ....................... 213

Recent Range Extensions, Name Changes And Status Updates For Selected South Dakota Fishes. Cari-Ann Hayer, Brandon C. Harland and Charles R. Berry, Jr. .................................................................................... 247

Economic Performance In Forestry-dominated Lincoln County, Montana: 1969-2003. Russell L. Stubbles .......................................................... 267

Abstracts of Senior Research PapersPresented at the 91st Annual Meeting of the


Bacterial Succession On The Leaves Of Ponderosa Pine, Pinus Ponderosa. Crystal Hostetter and David Bergmann ............................................................. 275

Synthesis Of 8-aza-1, 4-Oxidobicyclo[4.3.0]non-2-ene On A Solid Support. Nandeo Choony, Jeremiah J. Gums, Amanda Pekny and Mark Ranek ................................................................................................ 276

Assessment Of Random Metal Binding Peptides As Models For Allowable Mutations In Functionally Important Metal Binding Motifs. Carrissa Pietz and Robert Webb ......................................................................... 277

Fungi In The Diet Of Flying Squirrels (Glaucomys Sabrinus) Captured From The Northern Black Hills, June To October 2005. A. Gabel, C. Ackerman, M. Gabel, E. Krueger and S. Weins ............................................ 278

Variations In Rare Earth Element (REE) Signatures And Unit Cell Dimensions (UCD) For Purposes Of Stratigraphic Correlation In The Pierre Shale, South Dakota. Doreena Patrick, Paul N. Wegleitner and James E. Martin ........................................................... 279

The Effects Of Hypothyroidism On Spontaneously Hypertensive Heart Failure Rat Models. Brenda Simon, Bassel Kisso, and A. Martin Gerdes ........................................................................................ 280

Gene Expression In Ecologically Meaningful Contexts: Evolution Of Plant Defenses In Competitive Environments. David H. Siemens, Riston Haugen, Lexi Steffes and Richard Gayle ................................................. 281

The First Unequivocal North American Occurrence Of The Mosasaur Hainosaurus (Reptilia) From The Cretaceous Pierre Shale Of The Missouri River Trench, Southern South Dakota. James E. Martin, Wayne A. Thompson and David C. Parris ......................................................... 282

West Nile Virus Infection Rates And Culex Tarsalis Population Dynamics At A Farm Site In The James River Valley Of East-Central South Dakota. Ryan J. Beyer, Clayton J. Wulf, Jacob S. Schaeffer, Mitchel M. McKenzie, Michael B. Hildreth, Anne N. Rounds. Ragna A. Godtland, Rachel A. Hoffman and Christopher D. Carlson ............... 283

Archaea Associated With The Porcine Ileum Of Weaned Pigs. Clayton L. Scofield, V. Brözel, S. Lindblom, A. Rosa, S. Vilain, S. George, R. Kaushik and D. Francis ............................................................... 284

Lectin Binding Profile On The Small Intestine Of 5-Week Old Pigs In Response To Use Of Antibiotics As Growth Promotants. Sajan George, Yejin Oh, Sebastien Vilain, Volker Brözel, Stacy J. Lindblom, Artur J.M. Rosa, David Francis and Radhey S. Kaushik ...................................................................................... 285

Pathology And Dissemination Of Encephalitozoon Intestinalis In Gnotobiotic Piglets. Gopakumar Moorkanat, Aaron F. Harmon, Yejin Oh, Larry Hollar, David Francis, Michael B. Hildreth and Radhey S. Kaushik ...................................................................................... 286

Kinetics Of Quaternary Ammonium Methyl Carbonates With Differing Chain Lengths. Eric M. Villa, Jessica L. Paumen, Gary W. Earl and Duane E. Weisshaar .................................................................................... 287

Diabetes Population Studies In The Pine Ridge Native American Reservation. Patricia Hammond, Chance Weston, Deig Sandoval and Durga Mittinti ............................................................................................ 288

Statistical Software For Time Series Analysis. Paul Marshall and Daniel Swets ............................................................................................... 289

Polymerization And Copolymerization Of Linseed Oil With Styrene And Divinylbenzene. Jay D. Heeren and Timothy R. Hightower ...................... 290

Isomeric Series Of Novel Cyanate Esters And Polycyanurate Resins. Josiah Reams, Tsvetanka Filipova and David A. Boyles ...................................... 291

Polyetherimides Incorporating High Aspect Ratio Bisphenolate: Monomer Synthesis And Polymerization. Kalub Hahne, Tsvetanka Filipova and David Boyles ................................................................. 292

Effect Of Buckthorn (Rhamnus Cathartica) On The Tree Community Of An Eastern South Dakota Woodlot. Dale L. Droge and Javeria Javed ................................................................................................ 293

Growth And Production Of Different Varieties Of Bitter Gort Millon. Al Eastman, Brian Danner, Deig Sandoval and Durga Mittinti ........................ 294

Evidence For Enameloid In Xenacanthid Shark Teeth. Wendy Stiernagle and Gary D. Johnson ......................................................................................... 295

An Inventory Of Amphibians And Reptiles Of The Black Hills Of South Dakota. Laurelin Cottingham and Brian E. Smith ............................................. 296

Genetic Variance In The Smooth Green Snake, Opheodrys Vernalis, In South Dakota. Laurelin Cottingham, Brian E. Smith, Cynthia Anderson and Shane Sarver .................................................................. 297

Eubacterial Diversity Of The Porcine Ileum Of Weaned Pigs. L. Weyrich, C. Scofield, V. Brözel, S. Lindblom, A. Rosa, S. Vilain, S. George, R. Kaushik and D. Francis ............................................................... 298

A Study Of The Prion Protein (PrP) Gene: The Evolutionary History And Serial Transmission To Unrelated Species. Forrest Cain, Cynthia Anderson and Shane Sarver .................................................................. 299

Cost And Benefits Of Compounds Functioning As A Defense Against Herbivores In Boechera Stricta. Shane Ziegenbein, Lexi Steffes and David Siemens ............................................................................................ 300

Characterization Of Para-Assymetric Bisphenol A Polycarbonate By Gel Permeation Chromatography. Nathan Roark, Tsvetanka S. Filipova and David A. Boyles .......................................................................................... 301

Comparison Of Capture Methods And Home Range Of White-Tailed Jackrabbits In Selected Fields In Eastern South Dakota. Dustin Schaible and Charles Dieter ................................................................... 302

Geometric Layered Triangulation Of Lens Spaces. Luc Patry .................................... 303Phosphorylation Of eIF-4e Regulates p53 Protein Synthesis Following

DNA Damage. Ying Zhang and Da-Qing Yang ................................................. 304The Identification Of An Internal Ribosomal Entry Site In The

5’-Untranslated Region Of p53 mRNA Provides A Novel Mechanism For The Regulation Of Its Translation Following DNA Damage. Marie-Jo Halaby, Ying Zhang, and Da-Qing Yang .............................. 305

Big Foot Ride. Sandra Byrd, Brandon Ferguson and Sylvio Mannel ......................... 306Lakota Land—Mapping Culture, History And Recreation.

Charles Comes Killing and Sylvio Mannel ......................................................... 307Eagle Nest Butte. Elvin Returns and Sylvio Mannel .................................................. 308Adaptation Of Leaves In Cypripedium Candidum. Katie Krahn

and Carol Wake ................................................................................................. 309Ovary Anatomy Of Cypripedium Candidum. Joann McNally

and Carol Wake ................................................................................................. 310

Proceedings of the South Dakota Academy of Science, Vol. 85 (2006) 1

CONSOLIDATED MINUTES OF THENINETY-FIRST ANNUAL MEETING OF SOUTH

DAKOTA ACADEMY OF SCIENCE, DAKOTAWESLEYAN UNIVERSITY, HELD AT CEDAR

SHORE RESORT, OACOMA, SD, 7-8 APRIL 2006

The Executive Council met at 10:00 am Friday 7 April 2006 for a final check of plans for the meeting. President Bob Tatina opened the executive committee meeting and noted that a quorum was present. Proceedings Editor Steve Chipps noted that fewer copies of the Proceedings have been printed than in the past. The payment rate has been around 85-90% for articles and is somewhat lower for abstracts. The Proceedings accounts will need to be audited. A copy of the Treasurers Report was distributed by Treasurer Kristel Bakker. The Audit Committee will consist of Audrey Gabel.

Committee reports were as follows: President Bob Tatina handled publicity for the 2006 meeting. He sent an electronic photograph to Neal Reese for posting on the SDAS website, phoned local TV stations about the upcoming symposium on energy, e-mailed press releases to the Argus Leader, and coordinated external publicity on the web site. Bob volunteered to develop a contact list for the Academy. This list will serve as campus contacts for e-mails who can then forward to members on their respec-tive campuses. In addition, Bob has graciously volunteered to serve as the Public Relations member of the Executive Council. Nels Troelstrup reported that the 2007 SDAS annual meeting, hosted by SDSU, will be 13-14 April on the SDSU campus. Jim Sorenson noted that the 2008 meeting hosted by Mount Marty College will be held at Cedar Shore Resort. The following resolutions will be presented to the membership at the Busi-ness Meeting. Bob Tatina has authored a resolution to be forwarded to the South Dakota Board of Education on inclusion in the new set of administrative rules to include statements on the molecular basis of education and the scientific theory and biological principles of evolution. Mark Gabel will present two resolutions, a resolution on Biological Collections and a resolution on the SDAS 2006 An-nual Meeting. The Executive Council endorsed a recommendation from Kristel for the Non-game Teaming with Wildlife Coalition. An endorsement would support the concept of funding non-game outdoor education, but does not involve finan-cial commitment from the SDAS. Bob Tatina will head a committee for forwarding resolutions to Pierre or to University Presidents as appropriate.

2 Proceedings of the South Dakota Academy of Science, Vol. 85 (2006)

Bob Tatina reported the continuation of an opportunity through the AAAS to provide awards to students. For the undergraduate award, to be presented to one male and one female undergraduate student, Jim Sorenson and Steve Chipps will judge the poster session, and winners will be announced at the banquet. Krisma DeWitt graciously agreed to head the Fellows Committee. Nomina-tions for Fellows for 2007 will be sent to Krisma. The Nominating committee, Neil Reese from SDSU and Jim Sorenson from Mount Marty College will have the following positions to fill: Second Vice President, and two members-at-large. John Naughten and Dave Swanson will be going off as Members-at-Large. Donna Hazelwood and Kristel Bakker sent eight award checks for $25.00 each to five South Dakota Regional Science Fairs. The regional science fair co-ordinators are: 1) Jodie Ramsay, Northern South Dakota Science and Math Fair; 2) Madeline Rose Eastern South Dakota Science and Engineering Fair; 3) Brian T. Hemmelman, High Plains Regional Science and Engineering Fair; 4) Michael Nobel Farney, South Central South Dakota Science and Engineering Fair; and 5) Ms. Monica Mayer, Northwest Area Schools Regional Science and Engineering Fair. The SDAS presence at the Regional Science Fairs was discussed. It was de-cided that the Academy continue support and add a statement that each science fair will be notified that continuation of funding will necessitate an application by a given date. A proposal for Peer Review of papers will be presented to the membership. A request will be made for associate editors and a stringent three-week turnaround will be applied. The review would be in the form of a friendly review. To facili-tate the process, a form for review will be placed on the website. In addition, proposal to have abstracts pre-reviewed by the inclusion of two signatures at the time of submission will be proposed. Neil Reese proposed that the Proceedings be placed on the SDAS web site in a format searchable by author or title. He will present a motion at the Business Meeting to allocate an estimated $300.00 from the Proceedings budget for the purpose of formatting the papers from 1991 to the present in PDF format. President Bob Tatina opened the Business meeting 1:00 pm Saturday 8 April 2006, and introduced President Elect Jim Sorenson from Mount Marty College who gave the Presidential Address on “The End of Cheap Oil; Boom or Bust for South Dakota”. Registration for the Symposium and Annual Meeting began 8:30 a.m. Friday 7 April, continued again 8:00 a.m. Saturday 9 April. Steve Chipps brought Terri Symens to assist with registration. The Symposium on “Energy Resources in South Dakota: Issues and Global Concerns” was hosted by Perry Rahn and included eight presentations. The Plenary Speaker, Dr. Tad Patzek, Department of Civil and Environmen-tal Engineering, University of California – Berkeley gave a thoughtful and timely presentation on the “The Real Biofuels Cycle.”


The business meeting was opened by President Bob Tatina. President Elect Jim Sorenson gave a well received presidential address entitled “The End of Cheap Oil: Boom or Bust for SD”. On behalf of First Past President Andy Detwiler, Second Past President Miles Koppang presented to outgoing President Bob Tatina a plaque honoring his contributions to the Academy. In addition, Miles thanked the Symposium Committee Bob Tatina, Andy Detwiler and Perry Rahn. Steve Chipps reported on the Proceedings Account. The account is at the break even point. To stay solvent the membership will need to increase, or, as a last resort, the page charge will need to increase. Because the account will need to be audited, the report was accepted. The Treasurer report was provided by Donna Hazelwood for Kristel Bakker. Audrey Gabel served as Auditing Committee. Audrey reported that the account was in order. Jim Lefferts moved and Mark Gable seconded that the treasurer’s report be accepted. The motion passed by voice vote. The CD at Dakotah Bank will be allowed to rollover for another term. Mark Gabel and H.L. Hutchinson served as the resolution Committee and forwarded the following resolutions: A resolution authored by Bob Tatina to be forwarded to the South Dakota Board of Education on inclusion in the new set of administrative rules to include statements on the molecular basis of education and the scientific theory and biological principles of evolution. The resolution will be forwarded to the State Board of Education; the hearing date is set for 15 May. Neil Reese moved and Audrey Gabel seconded acceptance f the resolution. The resolution carried by voice vote. Mark Gabel presented a resolution on the need to protect Natural History Collections and the need for a one-half FTE curator position for the state. The resolution will be distributed to the SD BOR and the legislature. Steve Chipps moved and Mike Wanous seconded acceptance of the resolution. The motion carried by voice vote. The Resolutions Committee, Mark Gabel and H.L. Hutchinson proposed the following resolutions: 1) thanks to Dakota Wesleyan University for hosting the event and to Bob Tatina and the local planning committee Steve Chipps, Donna Hazelwood, and Terri Symens; 2) thank you Bob Tatina and the Sympo-sium Committee Andy Detweiler and Perry Rahn for arranging the Symposium 3) thank you to Keynote Plenary Speaker Dr. Tad Patzek, Department of Civil and Environmental Engineering, University of California – Berkeley who spoke on “The Real Biofuels Cycle”; 4) commend Past President Bob Tatina for his direction and leadership 2005-2006, 5) thanks to Terri Symens, Di Drake, and Nancy Presuhn for secretarial assistance and to Terri Symens for assisting with registration for the meeting, 6) thanks to Jim Lefferts for presidential address on “The End of Cheap Oil: Boom or Bust for SD” 7) thanks to Terri Symens for as-sisting with registration at the meeting; 8) thanks to Secretary Donna Hazelwood and Treasurer Kristel Bakker for continued service, and 9) a special thanks goes to Editor Steve Chipps for his oversight of timely publication of the Proceedings.


Neil Reese moved and Audrey Gabel seconded acceptance of the resolutions. The motion carried by voice vote. In order to remain financially solvent and to be able to continue generous support to of the Regional Science Fairs, The Executive Council forwarded a mo-tion to raise dues for regular members from $20.00 to $25.00 per year and for students from $5.00 to $10.00 per year. Bob Tatina moved and Audrey Gabel seconded acceptance of the motion. The motion carried by voice vote. The Executive Council forwarded a motion to upgrade the image of the Proceedings by requiring manuscripts to be peer reviewed. A lively discussion followed. The motion was amended to read: All manuscripts submitted for publication to the Proceedings of the South Dakota Academy of Science will be either pre-peer reviewed in the friendly sense or will be subjected to a more for-mal review process with forms completed by two reviewers. Bob Tatina moved and Mike Hildreth seconded acceptance of the motion. The motion carried. Bob reported on the conversion of the Proceedings to PDF files. Currently, the 2000-2005 Proceedings which have been converted to a PDF file, but are in a single file. The Academy has the opportunity contract for $300.00 the conver-sion of 1991-2004 data to page files which would be accessible on the web site. Bob Tatina moved and Mark Gabel seconded that the sum of $300.00 be granted for a contract to convert the 1991-2004 data to separate pages and to load the files separately on the database. The motion carried by voice vote. Bob addressed the question on life membership. Miles Koppang noted that in the past life membership was awarded to retirees how have been active and were nominated by the membership. The question raised was whether someone could purchase life membership before retirement. Mile Wanous moved and Jim Sorenson seconded that the Treasurer do a study on the benefits to the Academy for providing life membership to members for specific costs and ages of mem-bers. Winners of the Undergraduate AAAS awards were Kelley M. Vannatta and Cody Henriksen for the following presentations respectively; Kelley M. Vannatta (UG)* and Paul G. Egland, Augustana College-Biology. “Metabolic Interac-tions Between the Dental Plaque Bacteria Streptococcus gordonii and Veillonella atypica.”; and Cody Henriksen (UG)1*, Matt McDougall (UG)1, Mitch Weber (UG)1, Gina Furman1, Steve Matzner1, Neil Reese2, and Maureen Diggins1, 1AUGIE-Biology, 2SDSU-Biology and Microbiology.. “Effect of Acorus calamus Extracts on Glucose and Insulin Levels in Lethal Yellow (AY/A) and Black (A/A) Mice.” Kelly and Cody will each receive a one year membership in the AAAS. Bob Tatina addressed the question of Fellows and noted that senior mem-bers, especially retirees be honored as Fellows. Krisma Dewitt has graciously agreed to continue as Chair of the Fellows Committee. Bob has kindly agreed to continue as Chair of the Publicity Committee and will be sending out information on the need to recruit new members. Elections were held for officers for 2006-2007. The nominations committee, Neil Reese and Jim Sorenson presented the nominations. Bob Tatina moved and Mike Wanous seconded nominations cease and members cast a unanimous ballot in favor for Second Vice-President Krisma Dewitt, and for 2005-2007 member-


at-large for Dave Swanson and John Naughten. The motion carried by voice vote. On behalf of the membership committee, Nels Troelstrup recommended that the Academy accept as new members 120 individuals. Bob moved and Jim Lefferts seconded the motion. The motion carried by voice vote. Next year SDSU will host the 92nd annual meeting 13-14 April at SDSU. Following similar format, a keynote speaker will address the Academy and we will invite a symposium. Anyone interested in participating in the symposium is invited to visit with Nels, Steve, or Donna. Steve requested that the Academy provide honoraria for assistance in the following amounts to Terri Symens $500.00, Di Drake $100.00, and Nancy Pre-suhn $100.00. Neil moved and Donna seconded a motion to give the amounts requested. The motion carried by a voice vote.

Committee positions for 2006-2007 include Membership to be filled Fellows Krisma DeWitt Resolutions to be filled Nominations to be filled Publicity Bob Tatina

Outgoing President Bob Tatina handed the hammer of office to Incoming President Jim Sorenson. Neil Reese moved and Gary Earl seconded the meeting be adjourned. The motion carried by voice vote. Several items for consideration at the fall meeting of the Executive Com-mittee were discussed. 1) the 2007 meeting hosted by SDSU; 2) nomination of individuals for Fellow; 3) recruitment of new members; 4) the 2008 meeting hosted by Mount Marty College; 5) the locations of future meetings, 6) exposure and publicity of the academy, and 7) forwarding resolutions to Pierre. To Recap: the SDAS 2006 Annual Meeting hosted by Dakota Wesleyan University included a Symposium on “Energy Resources in South Dakota: Issues and Global Concerns” hosted by Perry Rahn and included eight presenta-tions. Saturday, 48 posters and 24 papers were presented. A total of 89 members attended the 2006 Annual Meeting.

Respectfully submitted,Donna Hazelwood, DSU

SDAS Secretary


REPORT OF THE RESOLUTIONS COMMITTEE

The membership of the South Dakota Academy of Science thanks the science faculty of Dakota Wesleyan University and the Program Committee including Bob Tatina, Steve Chipps, Donna Hazelwood and Terri Symens for arranging the 91st annual meeting in Chamberlain. We thank Dr. Tad Patzek for his keynote address entitled “The Real Biofuels Cycle.” Bob Tatina is commended for his service as President of the Academy for the past year. Donna Hazelwood deserves thanks for her job as Secretary of the Academy. Kristel Bakker’s service as Treasurer is greatly appreciated. Thanks are extended to Terri Symens for excellent secretarial help. Thanks to President-elect James Sorenson for his address to the member-ship. The Academy expresses their thanks to Steve Chipps, editor, and Ken Hig-gins, associate editor for their work on the Proceedings of the South Dakota Academy of Science.

RESOLUTION OF THE SOUTHDAKOTA ACADEMY OF SCIENCE

Whereas the mission of the South Dakota Academy of Science includes the stimulation of scientific research and education and the diffusion of scientific knowledge;

Whereas it is acknowledged that natural history research collections are the foun-dation of research in biological sciences and related fields;

Whereas natural history research collections are increasing in value due to ad-vances in biotechnology;

Whereas natural history research collections are important in understanding the history of our state and critical to comparing modern and historical distribu-tions of species by providing baseline data critical to future studies;

Whereas natural history research collections are important in monitoring the presence of insect pests, fungi and weedy species;

Whereas the biodiversity of the state of South Dakota is represented by major natural history collections of plants and fungi in the herbarium at Black Hills State University, plants in the herbaria of South Dakota State University and the University of South Dakota and by a major collection of insects at South Dakota State University;

Whereas increasing demands of teaching, research and service on the curators result in less time for management and improvement research collections;

Whereas there is an urgent need to implement measures to protect the specimens from insect and fungal pests to ensure the very survival of these collections for posterity;


Be it resolved that the South Dakota Academy of Science supports the active maintenance, development and protection of the natural history research collections of the state;

Be it further resolved that the South Dakota Academy of Science recommends the assignment of 1⁄2 FTE by the South Dakota Board of Regents to each of the aforementioned research collections;

Be it finally resolved that the 1⁄2 FTE will be permanently assigned to the re-search collection rather than to the university to preserve the collections in perpetuity to ensure that these valuable collections will remain a viable source of data for future generations of scientists.

Respectfully submitted,

Mark GabelH.L. Hutcheson

SOUTH DAKOTA ACADEMY OF SCIENCE2006-07 EXECUTIVE COUNCIL

President James Sorenson, MMC, Biology, 668-1581 [email protected] Michael Wanous, Augustana College, Biology, 274-4712, [email protected] Vice-President Nels H. Troelstrup, Jr. SDSU, Biology and Microbiology, 688-5503 [email protected] Second Vice-President Krisma DeWitt, MMC Chemistry, 668-1530 [email protected] Secretary Donna Hazelwood, DSU, Biology, 256-5187 [email protected] Kristel Bakker, DSU, Biology, 256-5182 [email protected] Proceedings Editor Steve Chipps, SDSU, Wildlife and Fisheries, USGS, 688-5467 [email protected] Past President Robert Tatina, DWU, Biology, 995-2712 [email protected] Past President Andrew Detwiler, SDSM&T, IAS, 394-1995 [email protected] Robert Tatina, DWU, Biology, 995-2712 [email protected] 2005-2007 Dave Bergmann, BHSU, Biology, 652-2420 [email protected] 2005-2007 Bill Soeffing, USF, Biology, 331-6759 [email protected] Member-at-Large 2006-2008 John Naughten, NSU, Biology, 626-2456 [email protected] 2006-2008 David Swanson, USD, Biology, 677-6175 [email protected]


2006 REGIONAL SCIENCE FAIRS2006 Regional Science Fair Location: Timber Lake, South DakotaSDAS Representative: NONE

STUDENT SCHOOL YEAR TEACHER TITLE OF PROJECT

Zach Bunn& Lane Alley Isabel 8th Merretta Kahl I Spy: Does Age and Gender Affect

Visual Crime Scene Recall?

Abbey Thiel Isabel 7th Merretta Kahl

What is the correlationbetween IMF% score and scrotal

size in Charolais and BlackAngus yearling Bulls?

Kourt Starr Dupree 7th Zach Davis Purinna Verses Oats

Danielle Alley& Kali Wiedmer Isabel 8th Merretta Kahl

Does Upper Body Fatigue orLower Body Fatigue Have More

Effect on Shooting Ability?

Kelsey Lindskov Isabel 8th Merretta KahlWhich is More Effective on

Alfalfa growth: Natural Compostor Brand Name?

Teacher: Zach Davis, Dupree

SOUTH DAKOTA ACADEMY OF SCIENCETREASURER’S REPORT 14-OCT-06

Balance as of 25-October-05 $9962.68

Debits: 2006 SDAS meeting: Ted Patzek 500.00 Terri Symens (room) 49.52 SDSMT 36.00 Printing 21.00 Di, Nancy, Terri (thank yous) 700.00 Cedar Shores 3045.77 U-Haul (Miles K) 77.78 Science Fair 1000.00 US Post Office 16.40 Total $5446.47

Credits: 2006 SDAS meeting $4695.00Balance as of 14-October-2006 $9211.21

Respectfully,Kristel K. BakkerSDAS Treasurer


EXECUTIVE SUMMARY,PROCEEDINGS EDITOR

Volume 84 for 2005 totaled 369 pages and the production cost was $6,362.25. Our production editor, Tom Holmlund, converted Proceedings issues and individuals papers to PDF format for the years covering 1992 to pres-ent; PDF files are now available on-line and can be accessed through our web site. The account balance for Proceedings as of April 6, 2006 was $5,315.25. Copies of Proceedings have been mailed to all current members, all life members, all State libraries, and to abstract indexing services.

Respectfully submitted by:Steven R. Chipps, Editor

For the South Dakota Academy of Science ProceedingsAugust 24, 2006


PRESIDENTIAL ADDRESS

The End of Cheap Oil:Positive or Negative Effects for South Dakota?

Address to the South Dakota Academy of ScienceDakota Wesleyan University, Mitchell, SD

April 8, 2006

Presented by James SorensonMount Marty College

Yankton, SD

How wonderful it is to have a meeting like this: scientists (teachers, re-searchers, and students) from all corners of the state getting together at a central location, exchanging ideas, renewing friendships, and introducing students to the benefits of such a meeting. Now I left from Yankton at 7:00 AM yesterday morning, traveled about 150 miles and arrived in time for a 10:00 AM meeting of the Executive Committee, and in plenty of time for the start of the Special Symposium on Energy Resources in South Dakota. While traveling that dis-tance I sometimes wonder what it must have been like to travel from Yankton to Fort Randall 150 years ago, or even 100 years ago. I’m sure it took longer than three hours and was considerably more perilous. Conversely, I seldom marvel at how wonderful automobiles and the highway system have become – because usually I take them for granted: I can traverse 200 miles at 70 – 80 mph in three hours in temperature-controlled comfort with music and other creature comforts. This has come to be the norm, it’s expected. Also taken for granted is that when I arrive the facilities will be warm and dry (or cool and dry, depending on the season), with hot and cold running water, ample food supplies that can be quickly prepared, entertainment in my room, motor boats and other recreational facilities just a short stroll away. This too has become the norm, it’s expected, in fact it’s demanded. So that when I arrive after a three hour drive I can complain about the long journey and demand warm food and a hot shower. I believe all of this was unimaginable slightly more than 100 years ago. Today I’d like to point out that all of this – this meeting, getting here quickly and easily, having warm comfortable accommodations, indeed the way of life we consider the norm, the lifestyle we have come to expect has been made possible by the ready availability of cheap oil (and natural gas), and that the era of cheap oil is coming to an end such that what was unimaginable 100 years ago may be-come unimaginable 100 years in the future. I’d also like to point out that as oil becomes increasingly scarce and expensive, the lifestyle that we consider to be the norm will have to undergo some difficult changes at a fundamental level, unless we can develop the alternatives – many of which were discussed at the sympo-sium yesterday – in a relatively short period of time before oil becomes scarce and too expensive. From a situation where getting from Yankton to Ft. Randall took a day or more, traversing that distance has become a two-hour trip; from


a society where (and some of the younger folks in the audience may think this gross and disgusting) family members lined up and took a weekly bath (whether you needed it or not) because producing that much hot water in a household was a major undertaking, we have gone to a situation where a daily shower is a required aspect of personal hygiene. The question is whether or not this “quick-trip, daily-shower lifestyle” can be maintained. I believe the urgency of the situation is under-appreciated. This is an erudite crowd, so let’s have a brief demonstration through discussion. Let’s pretend that we know with 99% certainty that oil will be in very limited supply (completely unavailable) 20 years from now. Turn to your neighbor and ask about their level of concern, and why?(Pause). Now I’ve asked a few people the same question, and based on that I would guess that few people here are really seriously concerned and the reason given is because “they” will come up with something else, an alternative, in the mean time. However, the problem that is under-appreciated is that the develop-ment and maintenance of the viable alternatives might require an underlying oil-economy, i.e that certain constructs made possible by cheap oil (and now considered to be the norm) may be necessary for the alternatives to be able to provide the energy to continue our lifestyle. So today I’d like to discuss the theory predicting the end of plentiful and cheap oil (and natural gas), the purported promise of several of the alternatives, the effects (both positive and negative) of the loss of cheap oil on South Dakota, and the potential role of SDAS in ameliorating the negative effects and accentu-ating the positive effects.

The end of cheap oil: Hubbert’s Curve. M. King Hubbert was a petro-leum geologist who taught at Columbia University, worked for the USGS, and in the mid-50s while working as chief of research for Shell Oil developed math-ematical models which predicted that the supply of oil produced in the United States would peak in 1969 and decline inexorably after that. Peak U.S. produc-tion occurred in 1971, and led to numerous problems, most notably the 1973 oil embargo, and our subsequent dependence on oil from the Middle East. Hub-bert began to expand his models to consider the global supply of oil, which he predicted would peak in the year 2000. Hubbert died in 1989, and subsequent tweaking of his model by Kenneth Deffeyes of Princeton, among others, put the global peak at sometime between 2000 and 2010 – in other words – about right now. The meaning of the peak is sometimes unappreciated: At peak it means that half of the Earth’s endowment of oil has been extracted and burned. But here’s the kicker – the easily extracted (translation = inexpensive) half has been burned; the remainder is either in remote locations or in other ways is not readily accessible. For example, the oil in northern Alaska (ANWR) has been repeat-edly debated as an important source to exploit to benefit the US. In the 1920s through the 1950s extracting oil from such a remote location would have been a ridiculous proposal. Now, in the big picture, some of the remaining oil will remain in the ground for economic reasons. Large corporations do not like to lose money, do not


exist for long if they consistently lose money, and are perhaps the only entities with the technology and the resources to continue the business of extracting, refining, and distributing oil. When the amount of energy required to extract a barrel of oil from the ground starts to equal the energy content of a barrel of oil, corporations would be engaged in an act of futility; when the energy required for extraction exceeds the energy extracted they would be engaged in an act of foolishness (and corporate suicide). This index of energy-return compared to energy-invested will subsequently be referred to as Energy Return over Energy Invested (ER/EI). Based on Hubbert’s curve and the realities of capitalism, it appears that in the near future supplies of oil and natural gas will decline, and this will result in making them much more expensive.

Alternative energy sources. Thus the need for alternative fuels. One major problem is that none of the alternatives can match oil (and natural gas) in terms of the combination of power, versatility, and transportability. Several of the alternatives can only be used to generate electricity – e.g. solar, wind, nuclear – and with no commercially viable electric automobile available (or even on the horizon) those alternatives can heat and cool our living spaces and power our lights, TVs, appliances, and computers, but they won’t help transport us. Bio-fuels will have to provide that power, but given how they are produced it would seem counter-productive (certainly counter-intuitive) to burn them to generate electricity. Since no one alternative has the versatility of oil, it would seem that to maintain our lifestyle, many of them need to be developed simultaneously. However, all of the alternatives are currently feasible because they are built on a “platform” made possible by cheap oil. Can the materials and energy need-ed to manufacture wind turbines or solar panels be provided by electricity? Can ethanol and biodiesel be used to maintain our roads? Can lightweight plastics for vehicles and consumer goods be manufactured without oil? 95% of nitrogen fertilizers are manufactured using natural gas as the feedstock. Can the crops to make ethanol and biodiesel be grown without fertilizers and oil-powered farm machinery? Perhaps. Perhaps not. The point is that whereas a combination of the alter-natives might be able to replace cheap oil and natural gas, to develop them and put them in place will require cheap oil and natural gas, and they are becoming less readily available and more expensive every day. Thus, developing the alter-natives requires that we as a society get the process started before it becomes too expensive to even contemplate. In short, we are in trouble and if we want to avoid serious problems, we better get started with the change-over. However, as noted earlier, most people do not appreciate the short window of time that is available to accomplish the change-over.

How four selected alternative energy sources are either propped up by cheap oil, or cannot feasibly replace oil. 1) Ethanol – With ethanol, the bottom line is that the production of ethanol is highly dependent on the use of oil and natural gas. Applying ER/EI may make


it a net energy loser unless systems for its production are modified to get away from using oil and natural gas to grow the crops. 2) Wind – can only generate electricity, so may not be capable of provid-ing for transportation. Must be situated in windy areas which will require oil-powered heavy equipment for installation of the turbines and for installation of transmission lines. 3) Solar –(active photovoltaics) can only generate electricity, so may not be capable of providing for transportation (at least not at the present scale). Solar panels and storage batteries take much energy to produce, and have many plastic and metal components that may be difficult to obtain and manufacture without fossil fuels. -(passive) can be used to heat buildings if they are designed for it, but most buildings have been built with designs that assume cheap oil and natural gas, so do not attempt to capitalize on passive solar heating, and would be difficult to restructure to do so. 4) Hydrogen – repeatedly I have heard politicians advocate the development of a “hydrogen economy” and how this will rescue us from our oil addiction and allow us to maintain our lifestyle. Somehow these “hydrogen economy” advo-cates think we can switch over directly from a fossil fuel economy to a hydrogen economy in the blink of an eye, but it ain’t gonna happen that fast and perhaps never. (And I think the hydrogen economy advocates either never took a chem-istry course or forgot what the periodic table looks like). A hydrogen economy involves fuel cells, a proven technology in which hy-drogen is combined with oxygen to provide an electric current which then does useful work. It is attractive from an environmental point-of-view because the by-product is water. Fuel cells could power automobiles. However, although hydrogen is the most abundant element in the universe (at least nearby) it is not so readily available in pure form. For automobiles, most proposals are to strip the hydrogen off of hydrocarbons, so we’re back to the problem of the depletion of oil and natural gas (and pollution). It also takes more energy to manufacture the pure hydrogen than the hy-drogen can produce in a fuel cell. Another way to get hydrogen is from water by running an electric current through the H20 to split the hydrogens from the oxygen and then collect them separately. The electricity would have to come from another source (perhaps wind, hydroelectric, or nuclear) but the process of freeing hydrogen from another compound is an energy loser: the average ER/EI is about 1:1.4, so what’s the point. Hydrogen presents other problems related to its storage and distribution, and these problems are a direct outcome of its chemistry. Molecular hydrogen is a very small molecule, especially compared to methane (natural gas) or gasoline. Thus it is very light and very difficult to contain. It is also highly corrosive, as it likes to combine with most other elements or molecules. Storage tanks, valves, and seals would have to be engineered with these characteristics in mind or they would quickly become leaky. In any event, the existing oil or natural gas pipeline cannot simply be switched over to carrying hydrogen – i.e. if fuel cell powered


automobiles are going to be the wave of the future, a separate, parallel hydrogen distribution system will have to be built (and I don’t see that happening). Also because hydrogen is so light, in order to provide enough of it to power an automobile over any distance it would have to be compressed. Compared to fuels that are liquids at air temperature like gasoline, transfer of compressed gases from one container to another is much trickier. Hydrogen-air mixtures are flammable and the heat of compressed hydrogen escaping may be enough to cause spontaneous combustion. Moreover, the logistics of providing hydrogen to conveniently located refuel-ing stations (“gas stations”) by truck would be daunting at best. Gasoline is fairly easily and safely distributed by unpressurized trucks. A loaded gasoline tanker can deliver about 25 tons of gasoline at a time, which is transferred to under-ground storage tanks by gravity. A comparably-sized hydrogen delivery tanker truck would only be able to carry about half a ton of hydrogen, in a high pressure tank. The energy consumption by the truck would be greater than the payload energy at anything more than a short distance. Also, to deliver the equivalent of a gasoline tanker (a day’s supply) would require about 25 hydrogen tankers to fuel the same number of automobiles. Thus, a hydrogen powered vehicle system is not going to replace gasoline-powered vehicles for chemical and logistical reasons. This is an illusion that will probably never become a reality. In talking about this to Dr. Patzek after his talk last night, he agreed that the advocacy of a hydrogen economy is an indication of how delusional we have become about being able to replace oil. Thus in review: a hydrogen economy is an illusion, the other alternatives are viable only on the current platform of cheap oil, and the era of cheap oil is coming to an end. This means that we have a short time-window to build up the alternatives before oil becomes prohibitively expensive. I think, rather I hope, that the various electricity-generating alternatives – some combination of wind power, hydroelectric, and (dare I say it) nuclear power – will be able to meet our needs for electricity, and thereby (hopefully) meet the heating, cooling, and appliance needs of hundreds of millions of homes, office buildings, and commercial establishments (not to mention educational in-stitutions) in this country. But we need to get it started. Meeting our transportation needs is another story. Biofuels seem to be the best answer, but currently, they too are heavily propped up by oil. It just might be that the end of cheap oil spells the end of America’s love affair with the auto-mobile.

How will South Dakota fare at the end of the era of cheap oil? Yesterday I learned that in wet years, South Dakota is already an energy exporter. It certainly may be possible, given our small population and the great, largely untapped potential for wind generators and the presence of hydroelec-tric facilities, that South Dakota could become an exporter of electricity to the surrounding, more populous states in all years. But we need to get wind farms constructed (soon!) and the dams unsilted (soon!).


However, I have my doubts that South Dakota can become a net energy exporter in the arena of biofuels. That is, South Dakota may be able to export some liquid fuels, but if they continue to be produced in the same way that they are currently produced, a bottom-line energy balance sheet, I’m afraid, would indicate a net energy loss. But what other alternative is there if we wish to keep motoring about? Perhaps if we could convert all of our vehicles, not just auto-mobiles but also trucks and tractors and highway equipment to be able to run on biofuels so that we can get to the point where we can grow the crops needed to produce biofuels without using fossil fuels, then South Dakota would become a net exporter of biofuel energy.

The role of SDAS in enhancing the production of energy in South Dakota. The South Dakota Academy of Science is in an excellent position to advocate a sensible but rapid change-over to alternative fuels. It is made up of scientists from all institutions of higher education in the state, most of whom are involved in teaching or research or (typically) both. (As the nature of scientists is to be skeptical, I trust that you won’t simply accept what I have presented here, but will check on the evidence and the arguments –pro and con – for yourselves, as one of my ulterior motives is to start a discussion on these issues among the SDAS membership). SDAS also has strong contacts among K-12 educators. The SDAS membership has to educate their students, and the public at large, that the end of cheap oil is upon us and that – like it or not – things are going to change. By being proactive we can perhaps shape the change to our benefit. However, being reactive will make it so the change shapes our lifestyle for a long time to come. The other area where the SDAS can play a pivotal role is in bringing together biologists, chemists, agronomists, genetic engineers, chemical engineers, and fuel specialists to optimize the production of dedicated energy crops, distinct from food crops. Increasing the biomass yield, but more importantly, the energy yield of corn and soybeans and other potential energy crops destined for ethanol and biodiesel production would be one goal, and should be attainable with a combi-nation of modern breeding techniques and transgenic techniques coupled with a willingness to modify the processing to improve fuel production. Another overall goal would be to shift production towards a self-sustaining system able to function with no input at all from fossil fuels. The benefits to the people of South Dakota of such developments would be manifest. Recall that last November, in East River, a major ice-storm and blizzard cut power to thousands of people, making some areas of the state temporarily uninhabitable. The severity of that storm, and the damage it incurred were not anticipated. Likewise, the end of cheap oil has the potential to cut power to many areas of the state and to make those areas uninhabitable. However, the damage potentially incurred by the loss of cheap oil can be anticipated, and we would be doing a disservice to the State of South Dakota if we were NOT to make an effort to minimize the damage and possibly, with foresight and plan-ning, to benefit.


REFERENCES

Appenzeller, T. (2004). The End of Cheap Oil. National Geographic. http://magma.nationalgeographic.com/ngm/0406/feature5/#top

http://en.wikipedia.org/wiki/Peak_oilhttp://www.ameinfo.com/71519.htmlZijlstra, T. http://www.zylstra.org/wordpress/index.php?cat=6Campbell, C.J. and Laherrere, J. The End of Cheap Oil. Scientific American.

http://dieoff.org/page140.htmThe Coming Global Oil Crisis: The Hubbert Peak for World Oil http://www.

hubbertpeak.com/summary.htmSimonds, T. A Pollution-Free Hydrogen Economy? Not So Soon. http://www.

groupsrv.com/science/about22839.htmlEnergy Bulletin. http://www.energybulletin.net/14012.htmlEthanol Fuel from Corn Faulted as “Unsustainable Subsidized Food Burning”

http://healthandenergy.com/ethanol.htmStraight Talk on Clean Energy. http://www.cleanenergyfuels.com/Quickfacts.

html

Complete Senior Research Papers

presented at

The 91st Annual Meeting

of the



ECLIPSE OF THE INNER SATELLITE OF JUPITER

Perry H. Rahn Department of Geology & Geological Engineering

South Dakota School of Mines & TechnologyRapid City, SD 57701

Jeffrey T. RahnInfinera, Inc.

Sunnyvale, CA 94089

Keywords

Jupiter, Io, light

ABSTRACT

The calculation of the velocity of light by Ole Roemer in 1675 was an amaz-ing astronomical achievement. Roemer timed the period of revolution of Io about Jupiter and noticed the time was not the same when Earth was advancing towards Jupiter, as it was when Earth was receding from Jupiter. He correctly at-tributed this to the fact that the transmission of light was not instantaneous, and made the first determination of the speed of light. We repeated Roemer’s experiment using a six-inch reflecting telescope. Five observations of the beginning of an Io eclipse were made during the winter of 2004-05. The orbital period of revolution averaged 42h 28m 13.3s over an in-terval of 18 orbits. During the summer of 2005 three observations were made of the ending of an Io eclipse. Over 13 orbits the period of revolution averaged 42h 28m 47.6s. During the winter Earth was advancing into the light emanat-ing from Io and the orbital time was perceived to be 34.3 seconds shorter than the summer when Earth was retreating from Io. Using these data, and assum-ing Earth’s orbital velocity relative to Io, the speed of light was calculated to be 266,000 km/s. This is within 11% of the published speed of light.

BACKGROUND

During the Renaissance there was heated philosophical discussion concern-ing the properties of light, e.g., the question whether light travels instantaneously or has some measurable speed. The invention of the telescope provided an answer to the question. Galileo (1564-1642) was the first to turn a telescope to the sky, and in 1610 he saw the eclipses of the four inner satellites of Jupiter. The “inner satellite of


Jupiter” is now known as Io; it has an amazingly fast period of revolution about Jupiter, only 42h 28m 30s. [Note: for clarification, an “eclipse” of Io refers to the obfuscation of light as Io falls into Jupiter’s shadow. An “occultation” of Io means it is concealed, or hidden from view, as it passes in back of Jupiter as seen from Earth. A “transit” means that, as viewed from Earth, Io goes in front of Jupiter; hence is no longer visible to the amateur astronomer.] Galileo noted that the eclipses were reliable enough to be used as timepieces (Sobel, 1995). He recognized that the problem of locating a ship’s longitude depended on an accurate clock, and visualized that sailors would someday find their location using timetables from astronomical data. He pondered the ques-tion of the speed of light and in 1638 devised an experiment to measure its speed. He proposed that two persons “…stand opposite each other at a distance of a few cubits and practice until they acquire such skill in uncovering and occulting their lights that the instant one sees the light of his companion he will uncover his own” (Cohen, 1944). The experiment did not work, but it demonstrates that Galileo felt that light was not transmitted instantaneously. During 1671 Ole Roemer (1644-1710) was an astronomer in Uraniborg, the name Tycho Brahe (1546-1601) gave to his observatory on the Danish island of Hveen. Roemer observed eclipses of Io, while Jean Picard (1620-1682) and later Cassini (1675-1712) made similar measurements in Paris. Cassini, Huygens and Descartes and other scientists were aware that eclipses were so reliable that they could provide an accurate timing device. From 1671 to 1677 Roemer carefully observed 70 eclipses and divided them into two groups. Beginning in the spring of the year during 1671, he noted that Earth was generally receding from Jupiter (“K” in Figure 1). He called the start of these events an “emersion”, i.e. Io suddenly exited from Jupiter’s shadow. Then, during the nights when visibility permitted, he noted the time of another emersion and from this determined the average period of revolution. The second group of timed events started in the fall of the year 1671 when Earth was ap-proaching Jupiter (“G” in Figure 1). Roemer used an “immersion” as the start of this timed event, i.e. the moment when Io became eclipsed. Roemer noticed a consistent difference in the times between his two groups. He found the orbital period was greater during emersion. In September, 1676, Roemer announced some of his findings to the mem-bers of the Paris Academie des Sciences. Roemer said the speed of light was of such magnitude that it would require about 22 minutes to traverse the distance of the diameter of the annual orbit of Earth. In 1676, no one knew the Earth’s orbital diameter. Today we know the Earth is approximately 93 million miles from the Sun, equivalent to 149.6 million km or 8.3 light-minutes. Using these data, the speed of light using Roemer’s data would be: v = d/t = 299.2 X 109 m / 22 min = 299.2 X 109 m /1320 sec = 2.3 X 108 m/s.

Since the true speed of light is 3.0 X 108 m/s, Roemer had only a 23% error.


THEORY

Figure 2 is a sketch of a portion of the solar system showing Earth’s obit about the Sun and a portion of Jupiter’s orbit. The interval of time between Io’s occultation when observed at E1 is slightly different than at E2. Earth is receding from Jupiter at E2, so the light traveling from Io has to catch up to the Earth. [Note: the reason for the difference in time is not simply because Earth is closer or farther to Jupiter.] Figure 1 (Roemer’s original sketch) more accurately depicts the technique employed. The timed event is an eclipse, whereby Io falls into Jupiter’s shadow. An immersion begins when Earth is at G and the light from Io becomes extin-guished as it falls into Jupiter’s shadow at C. [Note: the website “Calsky” calls this “Io eclipse begin”.] Io exits Jupiter’s shadow after approximately 2h 14m, but this emersion cannot be seen from Earth because it occurs behind Jupiter at D (Figure 1). When Earth is at G only an immersion can be seen. Roughly six months later when Earth is at K an emersion can be seen, but not an immer-sion. Noting Io’s orbital period as perceived at G, and roughly six months later at K, the difference can be used to calculate the speed of light. Interestingly, there are references relative to Roemer’s experiment that are misleading. For example, the Figure 2 sketch (by Sears and Zemansky, 1955)

Figure 1. Roemer’s original sketch showing the Sun (A), Jupiter (B), and the Earth at different times of the year (G, H, L, K, and E) as Io begins to eclipse at C and emerge from Jupiter’s shadow at D. Counterclockwise revolution is illustrated. From Cohen (1944).


shows an occultation rather than an eclipse. Roemer actually timed eclipses, not occultations. Sobel (1995) describes Roemer’s experiment as follows:

“The eclipses of all four Jovian satellites would occur ahead of schedule when the Earth came closest to Jupiter in its orbit around the sun. Similarly, the eclipses fell behind the predicted schedules by several minutes when the Earth moved farthest from Jupiter. Roemer concluded, correctly, that the explana-tion lay in the velocity of light.”

As described by Sorbel, the experiment would not be possible. At Earth’s closest approach to Jupiter, the eclipses are not visible as Jupiter is behind the Sun. This experiment also relies on measuring the absolute time of the eclipse and comparing to a periodic schedule, requiring an absolute time reference over a year-long time period. The experiment should involve measurements of Io’s period of revolution when Earth is advancing towards Jupiter and comparing these times to when Earth is receding from Jupiter. Sorbel’s explanation misses the point in that the time difference in an Earthling’s perception of the period of Io’s revolution is not because of Earth’s distance to Jupiter. The difference in time is due to Earth’s movement, i.e., the fact that Earth is advancing or receding from the light being transmitted by Io.

OUR OBSERVATIONS

During 2004 and 2005 we used a six-inch reflecting telescope (Orion “Intel-liscope”) and timed events concerning Io’s orbit. The senior author is fortunate to live in a relatively dark place in the Black Hills, seven miles northwest of Hill City, South Dakota. There is no interference of city lights and the sky is usually free of clouds.

Figure 2. Sketch showing Earth’s orbit and how Io’s orbit can be used to determine the velocity of light. Clockwise revolution is illustrated. From Sears and Zemansky (1955).


Io events can be accurately obtained from numerous sources. We used the website “Calsky” to get an advanced warning of when these events would occur. Figure 3 shows the intervals between events during December 5 to 7, 2004, and helps explain the terminology of the events. At first a transit, as Io passed in front of Jupiter, was used to time Io’s period of revolution. For example, from Figure 3, on December 6, 2004, a transit began at 3h 48.2m. But an amateur astronomer, peering through a telescope, cannot determine this time with much accuracy because Io seems to gradually merge with Jupiter over one or two minutes. The beginning or ending of an eclipse, on the other hand, seems to occur almost instantaneously, and this time could be

Figure 3. Sketch of Io orbiting Jupiter during December 5-7, 2004, showing events listed in Calsky. Approximately to scale.


determined to within a second or two. As explained above, an immersion would be observed during the winter and an emersion during the summer. NASA’s Voyager spacecraft show remarkable images of Io as a tiny reddish-brown sphere suspended in front of Jupiter. Volcanoes can be seen erupting from Io’s surface. Jupiter’s diameter (71,398 km) greatly exceeds Io’s diameter (3,632 km). Figure 3 is a sketch of Io revolving about Jupiter. Io’s sidereal period is 42h, 28m 30s, and it’s semi-major axis of orbit is 422,000 km (Menzel and Pasachoff, 1985). From these data it can be shown that it takes 2h 14m for an Io transit, occulation, or an eclipse. Figure 3 shows Io during December 5, 2004, to December 7, 2004. The sketch explains why “Io eclipse begin” (Roemer’s im-mersion) is visible, but not “Io eclipse end” (Roemer’s emersion). From December 21, 2004, to January 22, 2005, five observations were made of “Io eclipse begin” and hence four intervals whereby Io’s period of revolution could be determined:

OBSERVATION DIFFERENCE FROM FIRST OBSERVATION

Date Time Revolutions of Io Time (s)

12/21/04 3:40:50 0 012/28/04 5:31:30 4 6114401/6/05 1:54:55 9 1376045

1/20/05 5:39:57 17 25991471/22/05 00:08:50 18 2752080

The last two measurements were from one orbit, yielding a period of 42h 28m 53s. For these winter measurements the orbital period averaged 42h 28m 13.3s over the entire interval of 18 orbits. This is equivalent to 152,893.3s. Roughly 6 months later the senior author timed the eclipses again. But, as explained above, for these summer observations, an immersion (“Io eclipse begin”) could not be seen. Only an emersion (“Io eclipse end”) could be seen. It’s more difficult to pick the exact moment of the ending of an eclipse than the beginning of an eclipse. The summer observations are the times when Io was first seen as I was peering intensely through the telescope, knowing (from Calsky) that the emersion was about to happen. I was staring at Jupiter, concentrating on the blackness just to the right of Jupiter, about one-half Jupiter diameter away. The first glimmer of light is the time I recorded. [Note: these times are all about 1 minute prior to the official Calsky time. Apparently Calsky calls “Io eclipse end” as the time when Io is completely illuminated.] I couldn’t accurately tell when Io became fully illuminated; Io just got brighter and brighter until after a minute or two it looked about as bright as the other 3 moons of Jupiter. I feel the best time to use is the first glimmer of light, the moment when I was certain I could see Io. I think Roemer used this, too. Another problem with summer observations is that the nights are short and so the opportunity to see any astronomical event is reduced. As explained above,


there is the problem of determining the time of the first glimmer of light. At-mospheric conditions vary due to turbulence related to temperature and winds aloft. For example, on occasion Jupiter’s two dark rings would fade in and out, and the brightness of the four Jovian satellites varied. The time recorded for im-mersions are probably only accurate to plus or minus 2 seconds; this is mainly due to my reaction time to get my eye from the telescope to my wrist watch. But during emersion, I estimate there could be a 10 or 20 second difference between my seeing the first glimmer due to variations in atmospheric conditions. Summer observations are also handicapped because of weather. June in the Black Hills is the rainy season, and it was cloudy on the few nights when an event was to occur. But eventually three good observations were made:

OBSERVATION DIFFERENCE FROM FIRST OBSERVATION

Date Time Revolutions of Io Time (s)

6/26/05 21:13:00 0 07/3/05 23:08:12 4 6117127/19/05 21:27:19 13 1988059

From the timed events for the summer, the entire interval is 23d 0h 4m 19s for 13 revolutions. Thus the orbital period is 42h 28m 47.6s (equivalent to 152,927.6s).

CALCULATION OF THE SPEED OF LIGHT

During the winter of 2004-2005 the orbital period was perceived to be 42h 28m 13.3s. During the summer of 2005 the orbital period was perceived to be 42h 28m 47.6s. The orbital period in the winter was perceived to be 34.3 sec-onds less than the summer. As a departure from the average orbital period, in the winter it took 17.15 seconds less time, and in the summer 17.15 seconds more time. For the purpose of this paper, it is assumed that, in the summer, Earth is receding directly away from Jupiter, and roughly six months later Earth is ad-vancing directly towards Jupiter. This can be visualized as, in the winter, Earth is moving towards Jupiter and plunges headlong into the train of light crossing through space from Io. In the summer Earth is receding, and the light catches up with Earth. In keeping with Roemer’s original experiment, the orbits of the planets are assumed to be circular and lying in one plane. The Earth’s “semi-major axis” is 149.6 million km (Menzel and Pasachoff, 1983). Earth’s orbital speed equals its circumference divided by one year:


v = 2 (π) (149.6 X 109 m) / 365.24 d = 9.3997 X 109 m / 3.1557 X 107 s = 29,786 m/s.

As a first approximation, in the summer, Earth retreats away from Jupiter at this rate. Hence, the distance traversed during the time of one Io revolution would be:

d = v x t = 29,786 m/s x 152,910.45 sec = 4.55 X 109 m.

We can say that, as perceived from Earth, when Earth was receding from Io the light took 17.15 seconds longer than it should have if light was transmitted instantaneously. Thus the speed of light (c) is the distance Earth traversed during this additional time divided by 17.15 seconds:

c = 4.55 X 109 m / 17.15 s = 2.66 X 108 m/s = 266,000 km/s.

This value is reasonably close to the value Roemer obtained, and within 11% of the published value of the speed of light (300,000 km/s). [Note: Sears and Zemansky (1955) show c = 299,790 km/s.] A major source of error in these calculations is that Earth is not traveling directly towards (or away from) Jupiter as assumed above. In other words, Earth’s velocity vector is not simply 29,786 m/s oriented exactly towards (or away from) Jupiter as assumed above. Further, Jupiter has moved during this time as it orbits the Sun. To be more accurate, one should determine the exact distances from Earth to Jupiter instead of using the distances as calculated from a tangential orbital velocity. A more accurate approach would entail Earth’s (and Jupiter’s) exact position on their elliptical traverses during the utilized interval of time. [The position of every planet in the solar system at any moment are published in “Sky and Telescope”. A program called “solar system live” calculates the distance to all planets at any time.] The orbital mechanics of Jupiter and its moons also vary the time between eclipses for Io. For example, the shadow of Jupiter advances slightly each time Io orbits, adding enough time to each orbit to total one period every Jovian year. This added time is not constant. The time between eclipses is longer at perihe-lion. Realization of the immense distances in space (even within the solar system) is a humbling experience. The Sun is approximately 8.3 light-minutes from the Earth, while Jupiter is approximately 43 light-minutes away. This experiment utilizes the timing of the light emanating from Io to the nearest second. But that light arriving at Earth left Io 43 minutes ago. In the summer this light catches up to Earth as Earth orbits around the Sun. In the winter Earth collides head-on into the oncoming light. It’s a merry-go-round of time and space of staggering dimensions.


SUGGESTIONS TO OTHER AMATUER ASTRONOMERS

Someone who wishes to determine the speed of light using Io’s period of revolution could follow the general procedure outlined in this paper. The follow-ing suggestions are offered to assist in that endeavor: • Location: Anyone living in a big city today is aware of lights reflecting off a smoggy overcast sky. [I lived in Chicago for a year and never saw a star or planet.] The observations described in this paper were taken from the Black Hills of South Dakota, one of the darkest areas in the U.S. as evidenced by NASA nighttime satellite imagery. Consequently most Americans are not going to be able to pursue this experiment at all. Ole Roemer, in Denmark in 1675 did not have to contend with city lights. • Weather: The spring of the year tends to be cloudy throughout much of North America; this precludes astronomical observations. Further, in the late spring of the year the nights become short and hence the likelihood of seeing an Io eclipse is not as great as those nights in the fall of the year. We were lucky to have made three observations during the spring of 2005. For example, on July 19 a cloud was floating by; but fortunately at the precise moment of emersion the cloud had not yet obscured the view and I was able to see the termination of Io’s eclipse. • Time: One must have an accurate watch. (Roemer must have had an amazing pendulum clock, one that was good to within a second over the period of time of one Io revolution.) I used a radio-controlled wrist watch that was ac-curate to the second. I double checked its accuracy against a Direct TV signal. The limiting factor for timing is that the time of an emersion is not completely evident; it’s a subjective call. Further, a few seconds may be lost between the time one looks through a telescope and sees the event and the time it takes to glance at one’s watch. Timing is best accomplished with a second person noting the time.

ACKNOWLEDGEMENTS

We thank Donald A. Teets, Department of Mathematics and Computer Science, and Thomas V. Durkin, Graduate Education and Sponsored Programs, South Dakota School of Mines and Technology, for review of this paper.

REFERENCES CITED

Cohen, I.B. 1944. Roemer and the first determination of the velocity of light. The Burndy Library, Inc., New York, 63 p.

Menzel, D.H., and J.A. Pasachoff. 1983. A field guide to the stars and planets. Houghton Mifflin Co., Boston. 473 p.

Sears, F.W., and M.W. Zemansky. 1955. University physics. Addison-Wesley Publ. Co., Cambridge, MA, 1031 p.

Sobel, D. 1995. Longitude. Penguin Books, New York, 184 p.


NITRATE IN RAPID CITY’S WATER SUPPLY

Perry H. RahnDepartment of Geology & Geological Engineering

South Dakota School of Mines & TechnologyRapid City, SD 57701

ABSTRACT

Nitrate concentrations are increasing in Rapid City’s municipal wells. Wells #5, #6, #8, #9, #10, and #11 were drilled to the Madison Limestone in 1991-92, and yearly samples for 1993-2005 show nitrate (as nitrogen) concentration in these wells fairly consistently increasing from roughly 0.15 to 0.35 mg/L. Well #8 is farthest from the Madison outcrops and has the lowest concentration. The nitrate concentration is still below the EPA drinking water limit of 10 mg/L. Nevertheless, the increasing concentration is disconcerting because it is clearly anthropogenic but its cause is not clear. Meadowbrook Gallery and Girl Scout Gallery obtain water from alluvium, and by induced infiltration from Rapid Creek. These two water sources show slightly declining nitrate concentrations, roughly from about 1.3 mg/L in 1993 to 1.0 mg/L in 2005. These relatively high values probably reflect fertilizers used in the Meadowbrook golf course and other places and/or high nitrate in Rapid Creek. Jackson Spring shows nitrate concentration increasing from roughly 0.25 to 0.35 mg/L. This water originates as part of the Jackson Spring/Cleghorn Spring complex; the nitrate probably reflects the general composition of the ground water in the Madison aquifer as well as Rapid Creek. The nitrate in the Madison wells and Jackson Spring could come from a number of sources: (1) Streams recharging the Madison aquifer at the sinkhole “loss zones” along Rapid, Boxelder, and Spring creeks. (2) On-site wastewater systems upgradient from the city wells. More than 1,000 upgradient on-site wastewater systems exist within three miles of the city wells. (3) Fertilizers from home sites and/or agricultural areas. Commercial agriculture is probably not the main reason for increasing nitrate in the city wells because there are no feedlots and practically no farmlands (where fertilizers would be used) on the recharge areas of the Madison aquifer. (4) Explosives used for mining. This is probably not the cause of nitrate in Rapid City’s water supply.

Keywords

Nitrate, Madison Limestone, septic tanks, Rapid City water supply, ground-water recharge


NITRATE IN KARST TERRAINS

Although igneous rocks may contain small amounts of soluble nitrate minerals, most nitrate found in the nation’s ground water comes from organic sources or agricultural chemicals (Hitt and Nolan, 2005). High nitrate problems are typically associated with excessive applications of nitrate fertilizer, feedlots, or runoff from barnyards. Organic materials also release nitrate by bacterial de-composition. Nitrates near the surface can be leached by percolating water and eventually reach ground water (Davis and DeWiest, 1966). Nitrate contamina-tion of karstic aquifer systems from agricultural sources has been documented at numerous sites (Katz et al., 2005, and references contained therein). For example, there has been a steady increase in nitrate concentration in the past 40 years in Fannin Springs, northern Florida. The Upper Floridan aquifer consists of Eocene/Oligocene carbonates, and nitrate (as nitrogen) concentration, up to 4.7 mg/L, originates from inorganic sources (fertilizers) as well as organic (manure spreading and waste disposal). Fertilizers are applied at an average rate of 19 kg/ha-year based on 50% agricultural use of the Fannin Springs basin. Solution-enhanced conduits occur in the Biscayne aquifer, a karst aquifer sup-plying water for Miami; the municipal well field, the largest in Florida, is near the Everglades wetland area. This karst area has also been studied for pathogen contamination including the presence of Cryptosporidium parvum (Renken et al., 2005).Anthropogenic (man-made) nutrient sources and resulting eutrophication of Florida’s waterways result from elevated nitrate. According to Bacchus and Barile (2005) invasive alien species include aquatic plants such as water hyacinth and cyanobacteria (blue-green algae); these plants grow vigorously if stimulated by nitrate. Nitrate components were indentified using nitrogen isotopes (δ15N), and originate from citrus groves, industrial dairies (manure from animal wastes), municipal effluent spray, and leachate from septic tank drainfields.

NITRATE IN SOUTH DAKOTA WATER

Typical nitrate (as nitrogen) concentration in municipal sewage ranges from 0-10 mg/L, whereas septic tank system concentration ranges from 0-50 mg/L, and barnyard/feedlot concentrations range from 0-200 mg/L (Meyer, 2000). In South Dakota, Meyer (1987) documented nitrate composition in South Dakota. Sources of nitrate include fertilizers, feedlots, landfills, domestic septic systems, and mining. Meyer found that for 1,037 private wells in the Big Sioux River basin, 284 wells had N03-N above the 10 mg/L drinking water standard. He es-timated that, for the State as a whole, several thousand private wells are probably above this standard. Fifty cases of methemoglobinemia have been noted. Water containing more than 10 mg/L has been linked to methemoglobinemia in infants, a condition in which the blood is deprived of oxygen and the skin turns blue. In South Dakota one fatality is known (Jeanne Goodman, SD Department of Environment and Natural Resources, 2005, pers. comm.). In the Black Hills, high nitrate is typically associated with domestic waste. Coker (1981) reported up to 19.3 mg/L nitrate (as nitrogen) in shallow alluvial


wells in Rapid Valley, downgradient from Rapid City. On-site wastewater sys-tems were responsible for the high nitrates as well as fecal and coliform bacteria. Musa (1984) found high nitrate and other contaminants in alluvial deposits in the eastern part of Rapid City. Hafi (1984) modeled high nitrate associated with domestic waste (on-site wastewater systems) in the flood plain area below Rapid City. A high nitrate (as nitrogen) concentration, 13 mg/L, was found in alluvium under the campus of SDSM&T (Rahn and Davis, 1986); this high value is probably the result of fertilizers used at the campus of South Dakota School of Mines and Technology, but also could include nitrate contributions from downtown Rapid City, which at one time had no municipal sewer system. Davis (1979) found coliform and fecal bacteria associated with domestic sewage in the Belle Fourche infiltration galleries along Spearfish Creek. Johnson (1975) determined nitrate concentration in shallow wells in Keystone; many of the private wells had 10 to 30 mg/L nitrate (as nitrogen), most likely due to private on-site wastewater systems in operation at the time. Schwickerath (2004) studied the Spring Creek watershed above Sheridan Lake. He reported fecal bacteria exceeded water quality criteria for immersive recreation (swimming). Nitrate concentrations up to 2.8 mg/L occurred in Palmer Gulch, a tributary to Spring Creek. Sawyer (in prep.) studied bacteria and nitrate in ground and surface waters at various Black Hills locations. He found that Boxelder Creek at Norris Peak Road generally had less than 0.2 mg/L nitrate (as nitrogen). Rapid Creek in Dark Canyon had 0.1 or less mg/L. Spring Creek at the Stratobowl ranged from 0.2 to less than 0.1 mg/L. A well at Rocky Knolls golf course in Custer had high nitrate (4.2 mg/L). Observation wells below the Hill City sewage lagoons were typically 1 mg/L, but one observation well had 10.1 mg/L. Bear Butte Creek near Galena had 4.8 mg/L, possibly due to mining activities on this watershed. Downstream, Sturgis well #2 had 1.2 mg/L, and Bear Butte Creek at Sturgis had concentra-tions ranging from approximately 2.6 to 3.9 mg/L (Williamson, 2000). Driscoll et al. (1996) showed two samples of Spring Creek above Mitchell Lake had 0.100 and 0.110 mg/L nitrogen (NO2 + NO3). Two samples of Rapid Creek above Victoria Lake had 0.15 and less than 0.05 mg/L, but below Farm-ingdale reached 3.40 and 0.75 mg/L. They found the highest nitrate in the Black Hills, 4.10 and 9.30 mg/L, to be from Annie Creek near Lead; this is a recently active gold mine area. Carter et al. (2002) reported the elevated nitrate at Annie Creek is caused by a breakdown of blasting agents and cyanide. Heap-leaching at surface gold mines generates nitrate in the leachate from the spent ore (Davis et al., 1996). Driscoll et al. (2002) showed a median value of nitrate + nitrite nitrogen in the Madison Limestone is less than 1 mg/L.

RAPID CITY WATER SUPPLY

Long and Putnam (2002) described the hydrogeology of the western part of Rapid City, and included an analysis of the ground-water flow patterns in the bedrock aquifers utilized in Rapid City’s water supply.


Figure 1 is a map of the western part of Rapid City showing the location of the primary municipal water supply consisting of nine wells, two infiltration galleries and one spring. In the summer, when demand is great, water is also obtained directly from Rapid Creek. The nine wells include three older wells in the Minnelusa Formation (#1, #3, and #4), and six newer wells (#5, #6, #8, #9, #10, and #11) in the Madison Limestone. Table 1 shows the dates of construction, the total depths, and ap-proximate pumping rates.

Figure 1. Map of Rapid City showing water supply sources. Modified from Greene (1999) and Anderson et al. (1999). The Madison Limestone is highlighted; its potentiometric surface (feet above sea level) and general ground water flow direction is shown. MB = Meadowbrook gallery. GS = Girl Scout gallery. JS = Jackson Spring.


The Meadowbrook and Girl Scout galleries consist of collection pipes laid horizontally into the alluvium along Rapid Creek. Most of this water is obtained by induced infiltration from Rapid Creek, although some is derived from ground water flowing downvalley through the alluvium. Jackson Spring is part of the “Jackson-Cleghorn complex” (Anderson et al., 1999). Ground water discharges into Jackson Spring and nearby Cleghorn Spring, emanating from the Madison Limestone via a breccia pipe extending through the Minnelusa Formation (Long and Putnam, 2002). Jackson Spring is along the bank of Rapid Creek; a pump installed into the spring also induces infiltration from Rapid Creek.

Table 1. Nitrogen concentration (mg/L) in Rapid City water sources. Data shows “Nitrogen (Ni-trate as N)” concentration for annual collection dates: from 1993 to 1996 by Maxim Labs and from 1997 to 2005 by Energy Labs. NA indicates no water sample was taken. [Note: the EPA maximum recommended allowable limit for drinking water is 10 mg/L nitrogen.] The discharge (Q) refers to average withdrawals from 1988 to 1997 (from Long and Putnam, 2002).

7/12 8/5 12/5 8/14 8/18 8/25 8/13 7/17 8/7 6/25 8/15 6/24 8/3

1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005

Jackson Springs(1) 0.14 0.19 0.28 0.4 0.36 0.33 0.35 0.33 0.27 0.4 0.36 0.34 0.31

Meadowbrook(2) 1.20 0.74 1.52 1.4 2.02 0.85 1.25 NA 1.07 0.6 0.93 1.0 0.83

Girl Scouts(3) 0.68 0.74 1.16 1.2 1.40 0.77 1.64 1.07 0.92 0.9 0.72 0.63 0.64

Well #1(4) 0.16 0.30 NA NA NA 0.27 0.34 0.27 0.24 0.3 0.35 0.33 0.31

Well #3(5) NA NA NA NA NA 0.28 0.28 NA NA NA NA NA NA

Well #4(6) 0.58 NA 0.42 0.5 0.44 0.51 0.43 0.42 0.42 0.5 0.56 0.60 0.53

Well #5(7) NA NA NA 0.2 0.15 0.19 0.2 0.20 0.17 0.2 0.30 0.28 0.28

Well #6(8) 0.14 0.22 0.26 0.2 0.26 0.25 0.27 0.23 0.21 0.2 0.32 0.29 0.28

Well #8(9) <0.10 0.13 0.14 0.1 0.10 0.15 0.15 0.17 0.14 0.2 0.27 0.24 0.23

Well #9(10) 0.25 0.26 0.27 0.3 0.28 0.36 0.38 0.31 0.32 0.4 0.42 0.40 0.38

Well #10(11) 0.12 0.21 0.21 0.2 0.19 0.26 0.28 0.26 0.22 0.3 0.33 0.31 0.30

Well #11(12) NA 0.26 0.25 0.3 0.27 0.32 0.33 0.26 0.27 0.3 0.39 0.36 0.39(1) Jackson Springs has a 22 ft deep “spring box”. Built in 1943. Water source is alluvium, Minnelusa Formation and Madison Limestone.(2) Meadowbrook is a 20 ft deep infiltration gallery. Built in 1950. Water source is alluvium.(3) Girl Scouts is a 20 ft deep infiltration gallery. Built in 1960. Water source is alluvium.(4) Well #1 is 1,460 ft deep. Drilled in 1933. Q = 0.13 cfs. Water source is Minnelusa Formation, Madison Limestone, and Deadwood Formation.(5) Well #3 is 902 ft deep. Drilled in 1935. Q = 0.22 cfs. Water source is Minnelusa Formation, and possibly Madison Limestone.(6) Well #4 is 1,080 ft deep. Drilled in 1938. Q = 0.70 cfs. Water source is Minnelusa Formation.(7) Well #5 is 1,272 ft deep. Drilled in 1991.Q = 0.77 cfs. Water source is Madison Limestone.(8) Well #6 is 1,300 ft deep. Drilled in 1991. Q = 0.32 cfs. Water source is Madison Limestone.(9) Well #8 is 2,680 ft deep. Drilled in 1991. Q = 0.39 cfs. Water source is Madison Limestone.(10) Well #9 is 1,051 ft deep. Drilled in 1991.Q = 1.16 cfs. Water source is Madison Limestone.(11) Well #10 is 1,790 ft deep. Drilled in 1992. Q = 0.91 cfs. Water source is Madison Limestone.(12) Well #11 is 1,280 ft deep. Drilled in 1992. Q = 0.15 cfs. Water source is Madison Limestone.


NITRATE DATA

Nitrate data (Table 1) are collected annually for all the Rapid City water sources. Complete water chemistry analyses including radiological data are col-lected less frequently. Figure 2 is a graphical plot of the nitrate concentration for the six Madison wells. The rising trends from 1993 to 2005 in all the wells are clearly evident. A best-fit linear line is shown for each well, using an EXCEL spreadsheet. The coefficients of determination (R2) are all high, indicating a good correlation.

POSSIBLE SOURCES OF NITRATE The increasing nitrate concentration in Rapid City’s wells is believed to be anthropogenic, i.e., caused by man’s influence on the hydrologic regime. There are many possible causes, including mining, agriculture, domestic waste, and lawn fertilizers. Mining impacts in the northern Black Hills have been shown to cause in-creased nitrate. There are no active mines in the three watersheds contributing to the Rapid City water supply; however, in the western part of Rapid City there are three active quarries in the Minnekahta Limestone. The rock is used for road base and for the manufacture of cement. The 100 ft thick Opeche Formation underlies the Minnekahta Limestone and probably prevents any appreciable wa-

Figure 2. Nitrate plots for Rapid City’s six Madison wells.


ter from seeping to the deeper Minnelusa Formation and Madison Limestone. The potentiometric head of these two deeper formations is approximately 3,420 to 3,500 ft (Long and Putnam, 2002; Carter et al., 2002); this is near the land elevation in the general area of these three quarries. Thus movement of contami-nants from the surface down into the Madison aquifer would be unlikely. These quarries do not appear to be affecting Rapid City’s wells because the nitrate concentration in the wells near the quarries is not greater than other wells. Three streams recharge the Paleozoic carbonate aquifer in the vicinity of Rapid City’s wells. Boxelder, Rapid, and Spring creeks lose much of their water to the Madison Limestone and Minnelusa Formation. The nitrate in Rapid City’s wells could simply reflect an increasing nitrate in these streams. This ni-trate probably stems from domestic on-site wastewater systems along the three streams and from the Hill City sewage lagoons along Spring Creek. Dye injected into the loss zone of Boxelder Creek showed up in Rapid City well #6 within 30 days and well #10 within 41 days (Greene, 1999). Spring Creek recharges wells downgradient from its loss zone; Putnam and Long (2002) found that water lev-els in wells near stream recharge zones rise with increased stream loss rates. Long and Putnam (2004) used stable isotopes of oxygen, δ18O, to estimate water ve-locity recharged at Spring Creek to a nearby well at Highland Hills. They found “conduit flow” velocity through the Madison Limestone was 540 m/d, resulting in a five-day response time. In November 5, 1993, a nearby well at Copper Oaks had an incident with Giardia bacteria, presumably from water infiltrating from Spring Creek (Miller, 2005). Long and Putnam (2005) concluded that public water supply wells downgradient are very sensitive to potential contamination from Spring Creek. For each of Rapid City’s Madison wells, Greene (1999) estimated the proportion of the water that originates from these three streams. He assumed recharge on local outcrops of the Madison Limestone were minor, and used potentiometric maps, dye test data, and δ2H and δ18O isotopes to estimate the percentage of water derived from these three streams as follows:

• Wells #5, #6, #8, and #10 “…obtain approximately 75 percent of their water from Rapid Creek and about 25 percent from Boxelder Creek”,

• Well #9 is 40% Rapid Creek water and 60% Spring Creek water,• Cleghorn Spring is 55% Rapid Creek water and 45% Spring Creek

water, and• Well #11 is primarily recharged by Spring Creek.

If these percentages are accurate, the rising nitrate in the wells simply reflects rising nitrate in the three streams. The existing data for nitrate concentration in these three streams is not complete enough to evaluate the possibility that they are the source of nitrate in Rapid City’s wells. While these streams might be the primary source, there is also local recharge to the Madison Limestone and Minnelusa Formation. Carter et al. (2002) show that between 1 and 2 inches annually recharge these units in the area within a few miles of Rapid City. Therefore the water reaching Rapid


City’s wells would also include water infiltrating down from the surface at the outcrop areas near Rapid City as well as water lost by the streams in the sinkhole zones. In a study of recharge to the Madison aquifer just west of Rapid City, extending roughly from Elk Creek to Battle Creek, Long and Putnam (2002, Table 7) estimated 38.8 cfs streamflow recharge and 16.1 cfs infiltration recharge from the surface. Using these numbers, the ratio of stream recharge to infiltra-tion recharge = 38.8/16.1 = 2.41/1. In other words, the water that is found in Rapid City wells originates as 71% stream recharge and 29% infiltration from nearby outcrops. This proportion would vary depending on the well location. For example, well #6 and well #10 would logically contain a greater percentage of stream recharge from Boxelder Creek than the other Madison wells. The prolonged Black Hills drought (1999 to 2005) may have some influ-ence on the nitrate concentration of stream water in that lower stream discharge typically has higher dissolved constituents. However, the nitrate concentrations were increasing prior to 1999, so the nitrate concentration is not simply due to the drought. The increasing urbanization surrounding Rapid City is of concern because most of these new houses have on-site wastewater systems. Sawyer and Cow-man (2000) inventoried wastewater systems in the Black Hills. In year 2001, approximately 9,000 on-site wastewater systems were identified in the central Black Hills (Sawyer and Lindquist, 2003). The U.S. Environmental Protection Agency (2002) estimates that 10 to 20% of on-site wastewater systems are “fail-ing”. Some areas west of Rapid City have very thin soils and are underlain by Paleozoic carbonate rocks that show karst features and contain commercial caves. Sawyer (in prep.) documented the movement of pathogens and contamination in some of these areas. On-site wastewater systems only remove about 10 to 20% of nitrogen, generally in the soil horizon (Anderson, 2003). Natural recharge occurs by precipitation falling directly on areas underlain by Paleozoic rocks. Rahn and Gries (19763) found that most water discharged by large springs around the perimeter of the Black Hills originates in this way. Carter and Driscoll (2006) estimate that the annual recharge from precipitation on the Madison Limestone and Minnelusa Formation within a few miles of Rapid City ranges from 2.5 to 5 inches. Infiltrating water from septic leach fields would be added to this natural recharge. Figure 3 shows the location of on-site wastewater systems near Rapid City that were known to exist in 2001. Many new houses are upgradient and are directly on the Madison Limestone and/or Minnelusa Formation. The ap-proximate number of on-site wastewater systems for areas upgradient of the new Madison wells (located west of the contact of the Spearfish Formation and the Minnekahta Limestone as shown in Figure 3) can be summarized into three jurisdictional areas:

205 within the city limits,170 within 1 mile of the city limits,652 within 3 miles of the city limits.

The total from these three areas is 1,027 sites. This represents a minimum num-ber because some sites were undoubtedly missed during the initial inventory, and


many new sites have been developed since 2001. The recharge area for the Rapid City wells, as defined by the South Dakota Source Water Assessment and Pro-tection Program, includes the drainage basins of Boxelder, Rapid, and Spring Creeks; an estimated 5,000 on-site wastewater systems exist in this recharge area. There are numerous on-site wastewater systems along Rapid Creek in the Hisega area and Spring Creek below Hill City; these could lead to higher nitrate in these streams.

Figure 3. Map of the western part of Rapid City showing the city limits and location of on-site wastewater systems (modified from South Dakota Department of Environment and Natural Re-sources, unpublished data, 2001). Small circles show the location of on-site wastewater systems; these systems are only shown upgradient of the contact between the Minnekahta Limestone and the Spearfish Formation (heavy line). The outcrop area of the Madison Limestone is highlighted. Large circles show the location of Rapid City’s primary water supply.


The nitrate concentration for the two infiltration galleries is somewhat ir-regular, but generally is greater than the other sources. The concentration seems to be decreasing over the past 13 years. The nitrate concentration probably re-flects fertilizers applied to the Meadowbrook golf course. Fertilizers used in agricultural lands probably contribute little nitrate to Rapid City wells because there are few farms in the recharge area. However, lawn fertilizers used in the expanding urban areas could be a significant nitrate contribution. Range cattle, horses and pets would contribute some nitrate but these are probably not significant.

CONCLUSION

The increasing nitrate in Jackson Spring and Rapid City’s six Madison wells is most likely anthropogenic, but its origin is not known with certainty. It prob-ably reflects contamination by water recharging the Madison Limestone and Minnelusa Formation within a few miles upgradient of the wells. The cause of this contamination is most likely a combination of: (1) increasing nitrate in the three streams recharging these aquifers, and (2) water seeping into these aquifers from local on-site wastewater systems and lawn fertilizers. The general increasing concentrations of nitrate are similar for all the Madison wells. This indicates the nitrate source is widespread rather than a point source such as recharge by Spring Creek, the primary source of recharge for well #11. This supports the hypothesis that on-site wastewater systems are the major source of nitrate. While the concentration of nitrate is still much less than the drinking water limit, there is concern because the contamination is getting worse every year. Treated sewage wastewater and rural septic systems have been shown to contain pharmaceutical and endocrine-disrupting compounds that make their way into the nation’s waters (Stone and Heglund, 2005). If the increasing nitrate in Rapid City’s water supply originates from domestic waste, pathogens and pharmaceuti-cal products may eventually reach the water supply. Future research using nitrogen isotopes could help identify the nitrate sources.

REFERENCES CITED

Anderson, M.T. 2003. On-site waste disposal: septic tanks and the alternatives (Abstract): Western South Dakota Hydrology Conference, South Dakota Department of Environment and Natural Resources, p. 38.

Anderson, M.T., D.G. Driscoll, and J.E. Williamson. 1999. Ground-water and surface-water interactions along Rapid Creek near Rapid City, South Dakota: U.S. Geological Survey, Water-Resources Investigations Report 98-4214.

Bacchus, S.T., and P.J. Barile. 2005. Discriminating sources and flowpaths of anthropogenic nitrogen discharges to Florida springs, streams, and lakes: Environmental and Engineering Geoscience: Vol. 11, No. 4, p. 347-369.


Carter, J.M., D.G. Driscoll, and J. E. Williamson. 2002. Atlas of water resources in the Black Hills area, South Dakota: U.S. Geological Survey, Hydrologic Investigations Atlas HA-747.

Carter, J.M. and D.G. Driscoll, 2006, Estimating recharge using \relations be-tween precipitation and yield in a mountainous area with large variability in precipitation: Journal of Hydrology, vol 316, p. 71-83.

Coker, D. 1981. Shallow ground water resources of a portion of Rapid Valley, Pennington County, South Dakota: M.S. Thesis, South Dakota School of Mines and Technology, 96 p.

Davis, A.D. 1979. Hydrogeology of the Belle Fourche water infiltration gallery area, Lawrence County, South Dakota: M.S. Thesis, South Dakota School of Mines and Technology, 59 p.

Davis, A.D., A. Heriba, and C.J. Webb. 1996. Prediction of nitrate concentra-tions in effluent from spent ore: Mining Engineering, Vol. 48, No., 2, p. 79-83.

Davis, S.N., and R.J.M. DeWiest. 1966. Hydrogeology: John Wiley & Sons, New York, 463 p.

Driscoll, D.G., W.L. Bradford, and K.M. Neitzert. 1996. Selected hydrologic data, through water year 1994, Black Hills Hydrology Study, South Dakota: U.S. Geological Survey, Open-File Report 96-399.

Driscoll, D.G., J.M. Coker, J.E. Williamson, and L.D. Putnam. 2002. Hydrol-ogy of the Black Hills area, South Dakota: U.S. Geological Survey, Water-Resources Investigations Report 02-4094.

Greene, E.A. 1999. Characterizing recharge to wells in carbonate aquifers using environmentally and artificially recharged tracers: U.S. Geological Survey, Water-Resources Investigations Report 99-4018C, p. 803-807.

Hafi, Z.B. 1983. Digital-computer model for nitrate transport, Pennington County, South Dakota: M.S. Thesis, South Dakota School of Mines and Technology, 45 p.

Hitt, K.J., and B.T. Nolan. 2005. Nitrate in ground water using a model to simulate the probability of nitrate contamination of shallow ground water in the conterminous United States: U.S. Geological Survey, Scientific Inves-tigations Map 2881.

Johnson, T.J. 1975. Water quality of a shallow aquifer at Keystone, South Da-kota: M.S. Thesis, South Dakota School of Mines and Technology, 135 p.

Katz, B., R. Copeland, T. Greenhalgh, R. Ceryak, and W. Zwanka. 2005. Using multiple chemical indicators to assess sources of nitrate and age of ground-water in a karstic spring basin: Environmental and Engineering Geoscience, Vol. 11, No. 4, p. 333-346.

Long, A.J., and L.D. Putnam. 2002. Flow-system analysis of the Madison and Minnelusa aquifers in the Rapid City area, South Dakota---conceptual model: U.S. Geological Survey, Water-Resources Investigations Report 02-4185.

Long, A.J., and L.D. Putnam. 2004. Linear model describing three components of flow in karst aquifers using 18O data: Journal of Hydrology, Vol. 296, p. 254-270.


Meyer, M.R. 1987. A summary of groundwater pollution problems in South Dakota: Office of Drinking Water Quality, South Dakota Department of Water and Natural Resources, Pierre, SD, 18 p.

Meyer, M.R. 2000. Are chlorides a good indicator for western South Dakota aquifers?: in Strobel, M.L., et al., eds., Hydrology of the Black Hills: South Dakota School of Mines and Technology, Bulletin 20, p. 157-163.

Miller, S.L. 2005. Influence of geologic structures and stratigraphy on ground-water flow paths in the karstic Madison aquifer in the Rapid City area, South Dakota: Ph.D. Thesis, South Dakota School of Mines and Technology, 191 p.

Musa, N.S. 1984. Hydrogeology of the alluvial aquifer in eastern Rapid City, Pennington County, South Dakota: M.S. Thesis, South Dakota School of Mines and Technology, 96 p.

Putnam, L.D., and A.J. Long. 2005. Tracing ground-water flow with fluorescent dyes injected into Spring, Rapid and Spearfish creeks in the Black Hills, South Dakota (Abs): Western South Dakota Hydrology Conference, South Dakota Department of Environment and Natural Resources, Rapid City, SD, p. 33.

Rahn, P.H., and A.D. Davis. 1986. An educational and research well field: Jour. Geol. Ed., Vol. 44, p. 506-517.

Renken, R.A., et al. 2005. Assessing the vulnerability of a municipal well field to contaminate a karst aquifer: Engineering and Environmental Geoscience: Vol. 11, No. 4, pp. 319-331.

Sawyer, J.F. in prep. Water quality near selected wastewater treatment systems in alluvial hydrogeologic settings and in selected public water supplies, Black Hills, South Dakota: Ph.D. Thesis, South Dakota School of Mines and Technology.

Sawyer, J.F., and T.C. Cowman. 2000. Source water assessment and protection in the Black Hills region, South Dakota: in Strobel, M.L., et al., eds., Hy-drology of the Black Hills: Bulletin 20, South Dakota School of Mines and Technology, p. 214-221.

Sawyer, J.F., and V.A. Lindquist. 2003. Identification of onsite wastewater treat-ment systems in the Central Black Hills, South Dakota (Abstract): Western South Dakota Hydrology Conference, South Dakota Department of Envi-ronment and Natural Resources, p. 39.

Schwickerath, P. 2004. Analysis of fecal coliform bacteria in Spring Creek above Sheridan Lake, in the Black Hills of South Dakota: M.S. Thesis, South Da-kota School of Mines and Technology, 89 p.

Stone, J., and D. Heglund. 2005. Pharmaceuticals in water supplies (Abstract): Western South Dakota Hydrology Conference, South Dakota Department of Environment and Natural Resources, p. 12.

U. S. Environmental Protection Agency. 2002. Onsite wastewater treatment systems manual: U. S. Environmental Protection Agency, Office of Water, Washington, D.C. EPA/625/R-00/008, 308 p.

Williamson, J.E. 2000. Streamflow and water-quality data for Bear Butte Creek downstream from Sturgis, South Dakota: U.S. Geological Survey, Open-File Report 00-430.


NEW INFORMATION ON THE REE HEIGHTSFOSSIL SITE, HAND COUNTY, SOUTH DAKOTA

David C. ParrisNew Jersey State Museum

Trenton, NJ 08625

Doreena M. PatrickUniversity of Pennsylvania

Philadelphia, PA 19104

Jared WilliamsNew Jersey State Museum

Trenton, NJ 08625

ABSTRACT

Application of new analytical methods has refined interpretations of the Ree Heights Site, a fossil fish locality known for more than a century. Recovery of additional specimens, made possible by the generous permission of the Fawcett family, owners of the site, enabled Rare Earth Element (REE) testing of the sediments and fossils. Utility of REE analysis, previously established for marine sediments, can now be extended to non-marine environments, such as lacustrine settings. Results have been compared to the Fossil Lake site in Oregon and the Dungannon Site in Virginia, and much different REE signatures were deter-mined among these three freshwater Pleistocene deposits. These REE signature variations reflect differences in the environment during diagenesis and can be used to interpret differences in the paleoenvironment. Although previously correlated stratigraphically to the Cary Substage of the Wisconsinan Glacial Stage, the Ree Heights deposits have not as yet received an absolute date determination. Significant alteration of the fossil fish bones has precluded radiocarbon testing. Most of the species reported from Ree Heights are extant taxa. Renewed investigations are likely to add new records of species that are still living within the region.

REE HEIGHTS FOSSIL LOCALITY

Known for more than a century, the unique deposit at Ree Heights, South Dakota, continues to produce Pleistocene fish fossils of considerable interest. Previous investigations have emphasized the identification of the fish fauna, pa-leoenvironmental interpretations of the lacustrine environment, and the general age determination. While all of these aspects remain of interest, our renewed investigations seek to provide detailed geochemistry of the deposit in compari-


son to other known lacustrine sediments of the Pleistocene epoch. We also seek to refine the age determinations, in order to place the site accurately within the spectrum of glacial advances and retreats. New techniques of instrumentation make these analyses possible, most notably the testing of the fossil bone and sediment for Rare Earth Elements (REE). The Ree Heights fossil site is located in Hand County, South Dakota on the property of Leonard and Dorothy Fawcett, who for many years have generously permitted access to the site for the sake of scientific investigations. The site is in the southwest quarter of Section 21, Township 111 North, Range 70 West. The specific deposit is a remnant of Pleistocene sediment on top of a hill. The geol-ogy was described in detail by Ossian (1973), who performed the most extensive investigations of the site as a doctoral dissertation. However, the first report of fossil fish specimens from the site was by Cope (1891). It was then thought that the site was of Tertiary age, and inaccuracies and misinterpretations resulted. Various other expeditions have since collected at the site, as reviewed by Ossian (1973). By the middle of the Twentieth Cen-tury, many specimens were in repository in various collections, and the age was determined to be Pleistocene. The investigation of Ossian (1973) resulted in a detailed faunal list and geological interpretations of the deposit itself. It was cor-related to the Cary Substage of the Wisconsinan Stage on stratigraphic evidence. No subsequent investigation has challenged this determination, although the site has been reviewed various times (Pinsof, 1985).

CURRENT INVESTIGATIONS

Our investigations began in 2004, with additional collecting in 2005. We seek to perform detailed geochemical testing, to increase the faunal list for the site, and also hoped to obtain a radiocarbon date for the deposit. The latter has not yet been achieved, because preliminary testing indicated that the fragile fossil fish bone had been contaminated with carbon during diagenesis. We continue to seek ways in which an absolute date may be obtained. Preparation and identifica-tion of the newly collected specimens continues. This report deals with the Rare Earth Element (REE) testing, which was successful.

REE IN FOSSIL BONES: GENERAL PRESUMPTIONS

Recent bones have concentrations of <20ppm REE; fossil bones can have concentrations as high as 20,000ppm. There is no evidence to suggest that concentrations in recent bone are affected by diet, species, phylogeny, or type of bone. REE are introduced during early diagenesis within a few thousand years post mortem. The stable REE signatures therefore seem to reflect the composi-tion of pore waters during early diagenesis. REE in fossils record average sig-natures of water in depositional/early diagenetic environment and can be used in the interpretation of paleoenvironments. REE signatures in vertebrate fossils within lithostratigraphic members/units generally are characteristic of each unit


and can thus be used to identify lithostratigraphic units for the purposes of pro-venience determination and stratigraphic correlation.

REE HEIGHTS RARE EARTH ELEMENT RESULTS

REE signature results are depicted in Figure 1 in comparison with test results from other sites determined to be of lacustrine origin. REE signatures indicate different depositional environments. Differences in the REE geochemistry of the fossil vertebrate material reflect differences in the availability of the elements in water mass during early diagenesis (Patrick et al., 2004). The samples from Ree Heights are from the same unit and thus show similar signatures. Further testing of sediments at the site may show distinctions among strata. Concentrations are lower in the Ree Heights sample than for the other sites shown by comparison. This is an artifact of sampling. The interpretation of the signatures is by compari-son of the elements with one another. Using three representative rare earth elements, the concentration signatures may also be shown graphically for the freshwater sediments thus far tested. The three end members, gadolinium, neodymium, and yterrbium, are transition lanthanide series elements. The Ree Heights samples show equal mixing of them (Figure 2). This is believed to correspond to the field of circum-neutral waters and/or limestone dissolution. Fossil Lake samples show a mixing of two waters, indicating an evolution of waters during early diagenesis.

Figure 1: NASC-normalized REE signatures from lacustrine samples graphed in standard form as described by Patrick et al. (2004). The y-axis is log (REE/NASC), Rare Earth Element Content/North American Shale Composite.


ACKNOWLEDGEMENTS

We gratefully recognize the hospitality of the Fawcett families (Leonard, Dorothy, Dennis, and Connie) and their generosity to scientific investigations. Our field work took place during course work under the auspices of the South Dakota School of Mines and Technology (S.D.S.M.T.), which has provided ad-ditional facilities. Comparative studies at Fossil Lake, Oregon were conducted by Dr. James E. Martin (S.D.S.M.T.), with the cooperation of the Oregon Bureau of Land Management and support from Wayne Harold, Bess Harold, and Lena Martin. Susan C. Parris, Elizabeth Beitel, and Richard Large assisted us in the production of this report. The New Jersey State Museum (N.J.S.M.) will be the repository for specimens resulting from our investigations. Some of the data included herein are used by permission from copyrighted reports by GeoChemical Solutions LLC. (See Patrick and Wegleitner, 2003, 2005)

Figure 2. Triangular diagram of NASC-normalized values of Yb (a heavy rare earth), Gd (a middle rare earth), and Pr (a light rare earth), in standard form as described by Patrick et al. (2004). Axes are relative percentages, and relative enrichments are at each vertex.


REFERENCES CITED

Cope, E. D. 1891. On some new fishes from South Dakota. The American Naturalist 25: 654- 659.

Martin, J. E., Patrick, D., Kihm, A.,Foit, F., and Grandstaff, D. E. 2005. Lithostratigraphy, tephrachronology, and rare earth geochemistry of fossils at the classic Pleistocene Fossil Lake area, south central Oregon. Journal of Geology 114: 139-155.

Negrini, R. M., Erbes, D., Faber, K., Herrera, A., Roberts, A., Cohen, A., Wigand, P., and Foit, F. A paleoclimate record from the last 250,000 years, from Summer Lake, Oregon. Paleoclimatology 23: 125-149.

Ossian, C. R. 1973. Fishes of a Pleistocene lake in South Dakota. Publications of the Museum: Michigan State University 1 (3): 101-126.

Patrick, D., Martin, J. E., Parris, D. C., and Grandstaff, D. E. 2004. Paleoen-vironmental interpretations of rare earth element signatures in mosasaurs (Reptilia) from the Upper Cretaceous Pierre Shale, central South Dakota, U. S. A. Palaeogeography, Palaeoclimatology, and Palaeoecology 212: (3-4): 277-294.

Patrick, D. M. and Wegleitner, P. N. 2003, 2005. Mineral/Rare Earth Elements (REE) Interval Delineations, Correlations and Resulting Database. U.S. Copyright TX-6-098-093. (all rights reserved). GeoChemical Solutions LLC.

Pinsof, J. D. 1985. The Pleistocene vertebrate localities of South Dakota. Dako-terra 2 (2): 233-264.


DESCRIPTIVE ANALYSIS OF AQUATICINVERTEBRATE COMMUNITIES IN WADEABLE

AND NON-WADEABLE STREAMS OF THENORTHERN GREAT PLAINS NETWORK

Jill D. RustBiology and Microbiology Department

South Dakota State UniversityBrookings, SD 57007

Nels H. Troelstrup, Jr.Biology and Microbiology Department


ABSTRACT

The National Park Service has initiated the Inventory & Monitoring Pro-gram to identify and monitor vital signs of park conditions throughout the United States. Initial assessments and inventories are required to facilitate this monitoring program. This effort provided preliminary and methodologically consistent descriptions for wadeable and non-wadeable streams of the North-ern Great Plains Network (NGPN). Sweepnet samples were collected from 41 reaches of 7 non-wadeable streams and 23 wadeable streams during the summers of 2004 and 2005 using modified U.S. EPA EMAP protocols. Wadeable stream samples contained 219 taxa (77 families, 188 genera), comprised primarily of insects (86%). Non-wadeable samples contained 179 taxa (62 families, 148 gen-era), also mostly insects (85%). Diptera and Coleoptera contributed the greatest number of genera and species to wadeable and non-wadeable stream communi-ties. However, Ephemeroptera and Diptera were most numerically abundant from both habitats and the cumulative percent contribution of Ephemeroptera, Plecoptera and Trichoptera averaged 33.9% and 37.3% in wadeable and non-wadeable streams, respectively. Wadeable stream Shannon-Weiner H’ (SW) averaged 1.71 while non-wadeable SW averaged 1.51. Hilsenhoff Biotic Index values in wadeable sites ranged from 3.1 to 9.6 while non-wadeable site values ranged from 0.9 to 9.0. Swimming and clinging taxa, shredding large or gather-ing fine organic detritus were most abundant. Feeding and habit guild diversity was roughly similar between wadeable and non-wadeable stream classes. Results of this effort provide some of the first inventories and descriptions of aquatic invertebrates allowing comparisons among all 13 parks. These data provide a baseline for future monitoring of wadeable and non-wadeable streams within the network.


Keywords

National Park Service, invertebrates, invertebrate metrics, streams, vital signs

INTRODUCTION

The National Park Service (NPS) is known for its conservation efforts to-wards maintaining and managing minimally impacted ecological systems. Many parks are subjected to negative human impacts such as urban encroachment, air and water pollution, and excessive visitation. These impacts can threaten the quality or existence of many natural resources and ecosystems in the parks. To help prevent the loss or impairment of natural resources, Congress appropriated funds for the National Park Service to establish the Natural Resource Inventory and Monitoring (I&M) Program (NPS 1996). One of the main goals of the NPS I&M Program is to develop inventories of park resources and ecosystems and to determine their status. Baseline data will be used to monitor park resources into the future. Park monitoring will focus on changes in “vital signs” established for each park. Vital signs serve as measurable signals that indicate changes that may impair the long-term health of natural resources or ecosystems. Aquatic macro-invertebrate community structure has been rated high by park staff and collabo-rators for monitoring park conditions. Baseline descriptions are essential for monitoring future resource changes, whether those changes stem from management decisions, natural fluctuations, or anthropogenic disturbance. The objectives of this effort were to provide initial assessments of non-wadeable and wadeable streams within the Northern Great Plains Network (NGPN) and provide descriptions of macroinvertebrate com-munities within aquatic habitat of the NGPN.

STUDY AREA

All sampling sites were located within wadeable and non-wadeable streams of the Northern Great Plains Network (NGPN) of the NPS (Figure 1, Table 1). Parks comprising the NGPN are located in South Dakota, North Dakota, Wyoming, and Nebraska, falling within six different ecoregions (Table 2). Con-sequently, natural differences in habitat, physical, chemical parameters and inver-tebrate communities were expected. Each site was divided into different system types (wadeable and non-wadeable streams) and had one to three reaches.

METHODS

Habitat and biological assessments were conducted at all reach lengths within each study site and were adapted from EPA’s Environmental Monitoring and Assessment Program (EMAP). Modifications were made to several sampling


procedures as constraints in money and time prevented us from fully adopting the methodology (Rust 2006). Reach lengths were 40 times the wetted width of the channel (Lazorchak et al. 2000; Peck et al. 2006) with 10 transects at each reach. A hand held Global Positioning System (GPS) was used to determine latitude and longitude of all transects within a sampled reach. Invertebrate samples were taken from five randomly chosen transects within a reach. A D-frame net (350 um mesh) was used for invertebrate sampling at wadeable and non-wadeable sites. The base of the net was positioned against the substrate and the stream bottom was sufficiently disturbed to dislodge stream organisms for three minutes while the organisms were carried by the current into the net. Five sweepnets were combined to generate one composite sample. The sample was placed into a labeled container and preserved with 70% ethanol. Invertebrate samples were subsampled and sorted in the laboratory (Barbour et al. 1999). Samples were rinsed thoroughly in a 250 um mesh sieve to remove

Figure 1. Map depicting the NGPN of the National Park Service and the respective parks (NPS 2003).


ethanol and fine sediment. Large organic material (i.e. leaves, twigs, macrophytic mats) were visually inspected for invertebrates and discarded. After washing, the sample was spread evenly in a gridded pan and re-suspended in water. Four random grid cells were selected. The four grid cells were extracted from the whole sample and placed in another gridded pan. If more than 300 organisms (+/- 20%) were found within this sample, then subsampling was complete. If the invertebrate density of the four rings had many more than 300 organisms, ran-domly selected rings were extracted from that subsample and placed in another gridded pan for sorting until 300 (+/- 20%) organisms were found. If, in the original gridded pan, 300 organisms (+/- 20%) could not be found within the four rings, another ring was chosen until 300 organisms (+/- 20%) are found or the entire pan was sorted. After invertebrate sorting, the tray was scanned for large, rare, and voucher specimens. Ten percent of the samples were randomly recounted for quality control during the first year of invertebrate sorting. Major taxa were separated into separate vials to be identified to the lowest possible taxonomic level (genus, species) using several identification keys (e.g.

Table 1. Parks, their alpha codes, sites within the parks and classification of sites as either wade-able or non-wadeable streams for the NGPN.

PARKPARK

ALPHACODE

SITES SITECLASSIFICATION

Agate Fossil Beds AGFO Niobrara River StreamBadlands BADL Sage Creek StreamDevils Tower DETO Belle Fourche River River

Fort Laramie FOLANorth Platte River

Laramie RiverDeer Creek

RiverRiver

StreamFort Union Trading Post FOUS Missouri River River

Knife River Indian Villages KNRI Missouri RiverKnife River

RiverRiver

Missouri National Recreation River MNRR Missouri River

Mount Rushmore MORUBeaver Dam Creek

Lafferty GulchGrizzly Creek

StreamStreamStream

Niobrara National Scenic River NIOB

Niobrara RiverBerry FallsFort Falls

Smith Falls

RiverStreamStreamStream

Scott’s Bluff SCBL North Platte River River

Theodore Roosevelt THRO Little Missouri RiverBeaver Creek

RiverStream

Wind Cave WICA Cold Spring CreekHighland Creek

StreamStream


Merritt and Cummins 1996, Thorp and Covich 1991, Wiggins 1997, Weider-holm 1983). Invertebrate identifications were randomly checked by capable staff for quality control. Voucher specimens of each taxa were also kept for future reference.

RESULTS

Insects composed 86% of the 219 invertebrate taxa collected from wadeable streams. Overall, familial and generic diversity was high with 77 families and 188

Table 2. The ecoregions of the NGPN, average summer temperature and average rainfall.

ECOREGIONSUMMER

TEMPERATURE (MIN/MAX ºC)

ANNUALAVERAGE

RAINFALL (CM)

PARKSWITHINREGION

NorthwesternGreat Plains 13/29 38 BADL, FOUS,

KNRI, THRONorthwesternGlaciated Plains 16/32 53.3 MNRR, NIOB

Middle Rockies 13/28 43.2 DETO, JECA, MORU, WICA

NorthernGlaciated Plains 16/31 48.3 MNRR, NIOB

Nebraska Sand Hills 14/33 41.9 NIOB

Western High Plains 14/33 41.9 AGFO, FOLA, NIOB, SCBL

Western CornBelt Plains 17/31 61.0 MNRR

Table 3. Class Insecta familial and generic richness from wadeable streams of the NGPN.

ORDER FAMILIES GENERA

Diptera 12 68Coleoptera 13 33Trichoptera 9 16

Ephemeroptera 6 10Hemiptera 8 15Odonata 6 12

Plecoptera 5 8Collembola 1 1Lepidoptera 0 0Megaloptera 1 1


genera. Diptera and Coleoptera contributed the greatest number of families and genera for wadeable streams (Table 3). Fifty different taxa were unique to wade-able stream habitats within the NGPN. Most of the unique taxa were insects. Feeding guilds consisted primarily of collector-gatherers (58%), collector-filterers (17%), and predators (14%). Clingers (30%) and swimmers (29%) were the most common invertebrate habit guilds. Sprawlers (17%) and burrowers (14%) were moderately abundant (Figure 2). Wadeable systems within the Black Hills and the Niobrara basin all exhib-ited similar invertebrate metric characteristics. Systems within the Nebraska panhandle and eastern Wyoming (i.e. FOLA’s Deer Creek) displayed higher HBI values. Wind Cave wadeable streams also had higher HBI values (Figure 3). Those areas with fewer intolerant organisms were generally more accessible to livestock or bison (i.e. BADL, WICA, FOLA). Black Hills and Niobrara sites had the highest Shannon-Wiener Diversity (H’), while Sage Creek (BADL) displayed the lowest H’ of all sites (Figure 4). Percent non-insect abundance was lowest at Grizzly Creek (MORU), Fort Falls (NIOB), Highland Creek (WICA), and Sage Creek (range=0%-66.5%). High-est percentages of invertebrates other than insects were found from the Niobrara River (AGFO). The Ephemeroptera, Plecoptera, Trichoptera (EPT): EPT + Chi-ronomidae ratio averaged about 1.00 in Berry Falls (NIOB), Highland Creek (WICA), and Smith Falls (NIOB). Percent EPT richness was also over 40% at Fort Falls (NIOB), Smith Falls (NIOB), Berry Falls (NIOB), and Grizzly Creek (MORU). A total of 179 different invertebrate genera and species were found in non-wadeable streams of the NGPN (19 orders, 62 families, and 148 genera). Most of these were insects (85%). Coleoptera contributed the highest familial rich-ness, while Diptera contributed the largest number of genera (Table 4). Thirty-three invertebrate genera collected were only found in non-wadeable streams.

Figure 2. Functional feeding guilds and habit guilds of wadeable streams in the NG-PN. PerCF=Percent Collector-Filterers, PerCG=Percent Collector-Gatherers, PerScr=Percent Scrapers, PerShr=Percent Shredders, PerPred=Percent Predators, PerSwm=Percent Swimmers, PerClg=Percent Clingers, PerClb=Percent Climbers, PerBur=Percent Burrowers, PerSka=Percent Skaters, PerGL=Percent Gliders, PerSpr=Percent Sprawlers.


Ephemeroptera contributed the most to the unique taxa (9 genera), while Dip-tera contributed 8 genera. Feeding guilds primarily consisted of collector-gatherers (45%, Figure 5). Collector-filterers and predators were also well represented (18% and 19%). Habit guilds were dominated by clingers (33%) and swimmers (31%), with burrowers and sprawlers contributing moderately (15% and 13%) to guild structure. Overall, functional feeding guild (FFG) diversity equaled 0.82 while the Niobrara River (NIOB) and the Laramie River (FOLA) had the highest FFG

Figure 3. Mean (+/- 1 SE) intolerant taxa richness for wadeable streams for the NGPN. Site abbre-viations: BerFls=Berry Falls, BvDmCr=Beaver Dam Creek, CoSpCr=Cold Spring Creek, DrCrk=Deer Creek, FrtFls=Fort Falls, GrzCrk=Grizzly Creek, HldCrk=Highland Creek, LftGlch=Lafferty Gulch, NioRiv=Niobrara River, SagCrk=Sage Creek, SmthFls=Smith Falls.

Figure 4. Mean (+/- 1 SE) Shannon-Wiener Diversity (H’) for all wadeable streams within the NGPN. Site abbreviations: BerFls=Berry Falls, BvDmCr=Beaver Dam Creek, CoSpCr=Cold Spring Creek, DrCrk=Deer Creek, FrtFls=Fort Falls, GrzCrk=Grizzly Creek, HldCrk=Highland Creek, LftGlch=Lafferty Gulch, NioRiv=Niobrara River, SagCrk=Sage Creek, SmthFls=Smith Falls.


diversity (both with H’=1.02). The Missouri River (KNRI) had the highest FFG evenness (0.77), while overall FFG evenness was 0.66. Habit guild diversity was highest at the Niobrara River (H’=1.18) and overall habit diversity was 1.02. Habit evenness was highest within the Missouri River (KNRI, 1.00), while over-all habit evenness reached 0.73. Percent EPT varied considerably among non-wadeable sites (Figure 6). The Central and Gering irrigation canals (SCBL) displayed the highest percentage of EPT while Knife River and the Missouri River (KNRI) had the lowest numbers of these three insect orders. The North Platte River (FOLA) had the highest percentage of Chironomidae and total richness was greatest at the Niobrara River (NIOB), followed by the Laramie River (FOLA). Non-Insecta taxa were

Table 4. Class Insecta familial and generic richness from non-wadeable streams of the NGPN.

ORDER FAMILIES GENERA

Diptera 7 43Coleoptera 12 29Trichoptera 5 9

Ephemeroptera 8 20Hemiptera 5 9Odonata 3 7

Plecoptera 4 5Collembola 2 2Lepidoptera 1 1Megaloptera 1 1

Figure 5. Functional feeding guilds and habit guilds of non-wadeable streams in the NG-PN. PerCF=Percent Collector-Filterers, PerCG=Percent Collector-Gatherers, PerScr=Percent Scrapers, PerShr=Percent Shredders, PerPred=Percent Predators, PerSwm=Percent Swimmers, PerClg=Percent Clingers, PerClb=Percent Climbers, PerBur=Percent Burrowers, PerSka=Percent Skaters, PerGL=Percent Gliders, PerSpr=Percent Sprawlers.


also most abundant from the Laramie River. HBI averaged approximately 5.0, but HBI values over 6.0 were found from the Knife River and Missouri River (KNRI, Figure 7).

DISCUSSION

The insect orders Ephemeroptera, Plecoptera and Trichoptera were com-monly observed among NGPN wadeable and non-wadeable sites. Taxonomic richness of these orders and their contribution to total abundance have been selected as metrics for use in other lotic systems (Hannaford and Resh 1995, Maxted et al. 2000, Zweig and Rabeni 2001). EPT metrics do well for monitor-ing programs because these insect groups are sensitive to water quality changes related to dissolved oxygen. Mayflies, stoneflies and caddisflies all have gills that allow them to absorb oxygen from the water (Resh and Rosenberg 1984). If oxygen is depleted from organic pollution or temperature increases, numbers of EPT taxa and relative abundance decrease. Mayflies are also generally intolerant of low pH (Mackie 2004). Plecoptera reside in cool, well-oxygenated waters and have lower diversities in eutrophic streams (Resh and Rosenberg 1984, Mackie 2004). Caddisflies (Trichoptera) display a large diversity of habitat and func-tional feeding behavior (Mackie 2004, Mandaville 1999). They inhabit cool, running waters and their presence reflects good water quality (Mackie 2004). The order Diptera contributed a large percentage of total abundance and number of genera from sampled wadeable and non-wadeable stream sites. While generally not noted as good water quality indicators, members of the order Diptera are a very diverse group of insects and they respond differently to anthropogenic and natural disturbances (Lenat 1993, Gronke 2004, Foley 1997, Yoder and Rankin 1994). The Chironomidae, in particular, may be both very

Figure 6.. Mean (+/- 1 SE) percent EPT for all non-wadeable sites within the NGPN. Site abbre-viations: BelFrc=Belle Fourche River at DETO. FlPlt=North Platte River at FOLA, FuMis=Missouri River at FOUS, KniRiv=Knife River at KNRI, KrMis=Missouri River at KNRI, LarRiv=Laramie River at FOLA, LitMis=Little Missouri River at THRO, MisRiv=Missouri River at MNRR, NioRiv=Niobrara River at NIOB, and SbPlt=North Platte River at SCBL.


abundant and also very diverse (taxonomically and functionally) within prairie stream systems. Because members of this group may inhabit systems over a wide range of physical and chemical conditions (Mackie 2004), identification and enumeration of this group may facilitate discrimination among prairie stream sites. Functional guilds are examined to evaluate how organisms utilize their en-vironment. Guilds are defined based on evolved morphological and behavioral adaptations (Rosenberg and Resh 1996). Functional guilds are known to vary depending on system type, land-use, and landscape attributes. For example, invertebrate clingers and climbers were most prevalent in wadeable and non-wadeable stream sites with good water and habitat quality (Grizzly Creek, Smith Falls, Laramie River). These same habit guilds are often depressed when habitat becomes impaired (Maxted et al. 2000, Merritt et al. 2002, Mackie 2004). Glider abundance was especially high at the Niobrara River at AGFO. The community of this sampled reach included four gastropod genera and one bi-valve genus, together contributing up to 16% of the total abundance. The high abundance of gliding gastropods is probably associated with periphyton and detrital materials on and below macrophyte beds within the stream (Thorp and Covich 1991). Wadeable stream invertebrate metrics varied in association with geomorphic landscape features. Invertebrate diversity was greatest from Black Hills streams, but some prairie streams also displayed high diversity. For example, Shannon-Wiener diversity was high from the Niobrara River (AGFO) and Deer Creek (FOLA) relative to Black Hills streams. Both of these systems flow through prairie landscapes upstream from sampled reaches. However, while diversity was high in most of the sampled plains streams, these communities tended to have higher abundances of Diptera and Mollusca (e.g. Niobrara River at AGFO, and

Figure 7. Mean (+/- 1 SE) HBI for all non-wadeable sites within the NGPN. Site abbreviations: BelFrc=Belle Fourche River at DETO. FlPlt=North Platte River at FOLA, FuMis=Missouri River at FOUS, KniRiv=Knife River at KNRI, KrMis=Missouri River at KNRI, LarRiv=Laramie River at FOLA, LitMis=Little Missouri River at THRO, MisRiv=Missouri River at MNRR, NioRiv=Niobrara River at NIOB, and SbPlt=North Platte River at SCBL.


Deer Creek at FOLA), while most Black Hills streams and NIOB streams had higher numbers of EPT taxa. Plains streams have been shown to display smaller numbers of genera and families than mountain streams (Tate and Heiny 1995), but have also been shown to have higher production than forested streams dur-ing times of stable flow (Stagliano and Whiles 2002). Wadeable streams with higher numbers of tolerant organisms were found primarily in plains areas, particularly those with bison or livestock using the stream upstream or within sampled reaches. Plains streams sampled within this study were primarily used for livestock grazing upstream from sampled reaches. WICA streams were utilized by bison, and Beaver Dam Creek (MORU) had a heavily used horse trail along a portion of the stream. Livestock (or bison) graz-ing along the riparian area can also influence the structure of aquatic invertebrate communities by physically altering the habitat through trampling or foraging or by adding nutrients (through manure). These changes in habitat may lead to species replacement and a shift to greater numbers of tolerant taxa (Zomora-Munez and Alba Tercedor 1996; Del Rosario et al. 2002). The Missouri River communities in North Dakota had slightly fewer total taxa than the lower reaches in South Dakota, but had 10% more intolerant taxa, greater EPT abundance and more EPT taxa. The Missouri River in MNRR had higher percentages of collector-gatherers and clingers, but smaller percentages of swimmers. HBI was nearly equal for all Missouri River sites. Plecoptera was not collected from MNRR sites, but added over 6% to the total abundance for North Dakota Missouri River sites. Higher numbers of intolerant taxa may in part, be due to lower average temperature, higher dissolved oxygen, lower con-ductivity and dissolved solids from North Dakota sites (Weiel et al. 2003). An ideal temperature range for maximum growth of stoneflies falls between 5ºC and 22ºC (Brinck 1949; Heiman and Knight 1975). Summer temperatures of MNRR approached upper limits of that range while those from North Dakota sites were at the low end of the range. Data collected from this effort characterize invertebrate communities of NGPN streams using consistent methodology. These data may be used to iden-tify optimal invertebrate metrics for the development of an index of biotic in-tegrity and serve to facilitate future biological monitoring with NGPN streams. Limited invertebrate data are available within many of the parks sampled during this study. Thus, we recommend additional invertebrate inventories and more comprehensive sampling efforts to further describe and inventory these unique and diverse biotic communities.

ACKNOWLEDGEMENTS

Funding for this project was provided by the USDI National Park Service through a CESU cooperative agreement. Thanks are extended to Jill Anderson, Regina Cahoe, Jessica Meisenhoelder and Bret Winterfeld for their assistance in the field and laboratory. Thanks are also extended to park staff of the Northern Great Plains Network (NPS), North Dakota Department of Health, Nebraska Department of Environmental Quality, South Dakota Department of Environ-


ment and Natural Resources and Wyoming Department of Environmental Quality for their cooperation and assistance.

CITATIONS

Barbour, M,T., J. Gerritsen, B.D. Synder, and J.B. Stribling. 1999. Rapid bio-assessment macroinvertebrates and fish. 2nd Edition, EPA 841-B-99-002, United States Environmental Protection Agency, Office of Water, Washing-ton, D.C.

Brinck P. 1949. Studies on Swedish stoneflies. Opusc. Ent. Suppl. 11:1-126.Buss, D.F., D.F. Baptista, M.P. Silveira, J.L. Nessimian, L.F. Dorville. 2002.

Influence of water chemistry and environmental degradation on macroin-vertebrate assemblages in a river basin in south-east Brazil. Hydrobiologia 481(1-3): 125-136.

Carr, G.M., P.A. Chambers, A. Morin. 2005. Periphyton, water quality, and land use at multiple spatial scales in Alberta rivers. Canadian Journal of Fisheries and Aquatic Sciences 62(6):1309-1319National Park Service. 2006a. Water planning program http://www.nature.nps.gov/water/planning.cfm

Del Rosario, R.B., E.A. Betts, V.H. Resh. 2002. Cow manure in headwater streams: tracing aquatic insect responses to organic enrichment. Journal of the North American Benthological Society 21(2): 278-289.

Foley, J. 1997. A biological assessment of landscape disturbance potential on a northern prairie pothole lake. M.S. Thesis, South Dakota State University Brookings, South Dakota, USA.

Gronke, A.L. 2004. Development of an integrated index of biotic integrity for prairie pothole lakes of eastern South Dakota. M.S. Thesis, South Dakota State University Brookings, South Dakota, USA.

Hannaford, M.J. and V.H. Resh. 2000. Variability in macroinvertebrate rapid-bioassessment surveys and habitat assessments in a northern California stream. Journal of the North American Benthological Society 14(3): 430-439

Heiman, D.R. and A.W. Knight. 1975. The influence of temperature on the bioenergetics of the carnivorous stonefly nymph Acroneuria californica Banks (Plecoptera: Perlidae). Ecology 56: 105-116.

Lenat, D.R. 1993. Chironomidae taxa richness: natural variation and use in pol-lution assessment. Freshwater Invertebrate Biology 2 (4): 192-198.

Lazorchak, J.M., B.H. Hill, D.K. Averill, D.V. Peck and D.J. Klemm (eds). 2000. Environmental Monitoring and Assessment Program – Surface wa-ters: Field operations and methods for measuring the ecological condition of non-wadeable rivers and streams. U.S. Environmental Protection Agency, Cincinnati, OH.

Mackie, G.L. 2004. Applied Aquatic Ecosystem Concepts. 2nd Edition. Kend-all/Hunt Publishing Co.

Mandaville, S.M. 1999. Bioassessment of freshwaters using benthic macroinver-tebrates-a primer. First Ed. Project E-1, Soil & Water Conservation Society of Metro Halifax. viii, Chapters I-XXVII, Appendices A-D. 244p.


Maxted, J.R., M.T. Barbour, J. Gerritsen, V. Poretti, N. Primrose, A. Silvia, D. Penrose, R. Renfrow. 2000. Assessment framework for Mid-Atlantic coastal plain streams using benthic macroinvertebrates. Journal of the North Ameri-can Benthological Society 19(1): 128-144.

Merrit, R.W. and K.W. Cummins [eds.]. 1996. Insects of North America. Ken-dall/Hunt Publishing Co. Iowa, USA.

National Park Service. 1996. Natural resource information division. Inventory and Monitoring Program. Annual Report. http://www.nature.nps.gov/publi-cations/i&mann96/96i&mtxt.htm

Peck, D.V., A.T. Herlihy, B.H. Hill, R.M. Hughes. P.R. Kaufmann, D.J. Kl-emm, J.M. Lazorchak, F.H. McCormick, S.A. Peterson, P.L. Ringold, T. Magee, and M. Cappaert. 2006. Environmental Monitoring and Assessment Program-Surface Waters Western Pilot Study: Field Operations Manual for Wadeable Streams. EPA/620/R-06/003. U.S. Environmental Protection Agency, Office of Research and Development, Washington, D.C.

Resh, V.H. and D.M. Rosenberg. [eds]. 1984. The Ecology of Aquatic Insects. Praeger Publishers, New York.

Rosenberg, D. M. and V. H. Resh. 1993. Freshwater biomonitoring and benthic Macroinvertebrates. New York, New York: Chapman and Hall.

Rosenberg, D.M. and V.H. Resh. 1996. Use of aquatic insects in biomonitor-ing. In R.W. Merrit and K.W. Cummins [eds.], Insects of North America. Kendall/Hunt Publishing Co. Iowa, USA.

Rust, J. D. 2006. Establishing baseline data for aquatic resources in National Parks of the Northern Great Plains Network. M.S. Thesis, South Dakota State University Brookings, South Dakota, USA.

Stagliano, D.M., and M.R. Wiles. 2002. Macroinvertebrate production and trophic structure in a tallgrass prairie stream. Journal of the North American Benthological Society 21(1): 97-113.

Tate, C.M., J.S. Heiny. 1995. The ordination of benthic invertebrate commu-nities in the South Platte River Basin in relation to environmental factors. Freshwater Biology 33(3): 439-454.

Thorp, J.H., and A.P. Covich (eds). 1991. Ecology and classification of North American freshwater invertebrates. Academic Press, Inc., New York.

Weiderholm, T. 1983. Chironomidae of the Holarctic Region: Keys and diagno-ses. Part 1. Larvae. Entomologica Scandinavica No. 19.

Wiggins, G.B. 1977. Larvae of the North American caddisfly genera (Trichop-tera). University of Toronto Press, Toronto, Ontario, Canada.

Yoder, C.O. and Edward T. Rankin. 1994. Biological response signatures and the area of degradation value: new tools for interpreting multimetric data. In: Biological Assessment and Criteria. Tools for Water Resource Planning and Decision Making, Davis, W.S. and T P. Simon, [eds.] Boca Raton: Lewis Publishers.

Zweig, L.D. and C.F. Rabeni. 2001. Biomonitoring for deposited sediment us-ing benthic invertebrates: a test on 4 Missouri streams. Journal of the North American Benthological Society 20(4): 643-657

Zamora-Munoz, C. and J. Alba-Tercedor. 1996. Bioassessment of organically polluted Spanish rivers, using a biotic index and multivariate methods. Jour-nal of the North American Benthological Society 15(3): 332-352.


A CLUSTER FOR CLUSTERS: HIGH PERFORMANCE COMPUTING

FOR MOLECULAR DYNAMICS OF LARGERARE-GAS ATOMIC CLUSTERS

David T. Huebner and Brian G. MooreChemistry Department

Augustana CollegeSioux Falls, SD 57197

ABSTRACT

A rare-gas atomic cluster in the size range of 10-200 atoms can be prepared in either a liquid or solid state, an effect clearly seen in computer simulations and also to some extent in experiments. Larger clusters have not been studied in detail, especially the liquid state. Larger clusters prepared computationally initially in a liquid state seem to show a spontaneous, sharp transition to the solid. To study this effect in detail, one must be able to efficiently simulate the dynamics of clusters with thousands or tens of thousands of atoms. In order to meet these demands, high performance computing solutions and methods of improving and parallelizing the molecular dynamics program were investigated. A rack-mounted four node dual-processor cluster running the Clustermatic software package (developed at Los Alamos Labs) was designed to meet the high processing demands. This cluster has achieved a performance of 21.5 Gi-gaFLOPS during initial testing using HPL, a benchmark based on matrix calcu-lations. The molecular dynamics program was parallelized using MPI (message passing interface). We have run clusters up to 10,000 atoms, seeing a roughly 7.7 factor speed increase, with results confirming the presence of the spontane-ous transition from liquid to solid, which now appears to be ubiquitous for large clusters.

Keywords

Clusters, high performance computing, Beowulf, molecular dynamics, un-dergraduate

INTRODUCTION

The study of the properties of atomic clusters has entertained much research through both simulation and through experimentation. These clusters, experi-mentally formed in a vacuum by a jet and held together by dispersion forces, have many properties far different from that of bulk matter (Harris et al. 1984). Some of the more interesting properties of these clusters are the presence of liq-


uid state at temperatures and pressures far below the triple point and evidence for a liquid-solid transition (Hahn, et al. 1988). Previous work suggests that the stability of the liquid phase of a cluster may be critically dependent upon its size. In a constant energy simulation, a large liquid cluster may display a phase change as atoms desorb into the surrounding vacuum and the cluster cools. In Figure 1, the temperature over time of two simulations is shown, consisting of 133 and 650 atoms. The large temperature jump at approximately t*=3200 in the 650 atom run is an indication that a phase change has occurred, as the cluster changes from a cool liquid to a warmer solid. However, the smaller 133 atom cluster does not display a phase change and instead remains a stable liquid throughout the run (Moore, et al. 2000). All simulations were performed using the

Lennard-Jones force potential, shown below, to calculate the pairwise potential

Figure 1. Temperature of the entire system as a function of time for two runs of differing sizes. The larger, 650 atom cluster displays a phase change from liquid to solid at t*≈3200 while the smaller cluster remains a liquid throughout.


energy. In the force potential, r is the interatomic distance, σ is the atomic diameter, and ε is a parameter describing the depth of the attractive forces. In addition, all simulations were performed in dimensionless reduced units. Re-duced time, t*, is equivalent to (ε/mσ2)1/2t, reduced temperature is T* = kT/ε, and reduced distance is r* = r/σ. Under the temperatures the simulations are run, bulk matter cannot exist in a liquid state. However, some clusters can exist as a liquid under these same con-ditions. In order to study how the properties of a cluster map onto those of the bulk, we have begun to turn the focus of our study towards much larger clusters. We have studied clusters of sizes ranging up to tens of thousands of atoms that display a liquid state. Unfortunately, our computing requirements for studying these larger clusters scales quadratically as a function of the number of particles, leading to unfeasibly long calculation times.

METHOD

One of the first methods investigated for minimizing the calculation times for large clusters was that of high performance computing. This was due to the prevalence of high performance computing in molecular dynamics simulations as well as the raw, economical compute power offered by clusters. As the cost to performance ratio of the cluster was of primary concern, we chose to construct a Linux based Beowulf computing cluster (Sterling, 2001). A cluster consisting of five 1.8GHz dual Opteron nodes with 1GB of RAM and networked with gigabit Ethernet was designed with one node acting as master and the others as compute slaves. The master node runs Fedora Core 3 Linux with Clustermatic 5, an open source collection of cluster software based upon BProc packaged by Los Alamos National Laboratories (Hendriks, 2002). As with all BProc based clusters, the slave nodes run a minimal Linux kernel obtained from the master node over the network at boot time. While not required, as the kernel and entire process space is located in RAM, each slave node is provided with a hard drive to be used as swap space. Interprocess communication is handled by an implementation of MPI (Message Passing Interface), primarily MPICH (Dongarra , 1995, Gropp et al. 1997). In order to eliminate the need for any file transfer from master to slave, the master’s /home directory is mounted on each slave node via NFS (Net-work File System). Numerous mature programs such as NAMD and LAMMPS already exist that can be used for the simulation of atomic clusters (Phillips et al. 2005, Plimp-ton, 1995). However, our simulations are run on an object oriented molecular dynamics code written in C++ which has been developed in-house (Schneider-man et al. 2004). The use of in-house code has many benefits over other readily available packages. Due in part to the streamlined nature of the code and lack of unnecessary features, one such advantage is that it has been designed to en-able undergraduate students to quickly grasp the molecular dynamic calculations (Schneiderman et al. 2004). In addition, the relatively simple and flexible code base has enabled students with no prior programming experience to become


contributors in computational chemistry research by tailoring the code to their needs through simple modifications. The molecular dynamics code used follows the Verlet velocity algorithm over the course of many thousand steps (Verlet, 1967). A simplified primary loop structure of the non-parallelized code is shown below (Allen et al. 1987).

for(int i = 0; i < numSteps; ++i){ move1(); for( int j = 0; j < numAtoms; ++j) { for( int k = 1; k < numAtoms; ++k) { calculateForce(j,k); } } move2();}

The outermost loop breaks the simulation up into a number of steps and, in combination with the time step dt, dictates the length of simulated time for which the simulation will be performed. The method move1() updates the posi-tions of the atoms based on the pair potentials from the previous step and then to advances the velocities from t to t + dt/2. Then, for each atom, the force between it and all other atoms in the system must be calculated. Following that, move2() advances the velocities from t+dt/2 to t+dt. Since each atom is frozen in time during the force call, any calculation upon a particle may be calculated independently of all others provided that the positions of all atoms in the system are properly updated when the force is cal-culated. As such, the algorithm can be easily divided into equal partitions to be distributed amongst multiple processors, while allowing for the basic algorithm to remain unchanged. The basic structure of the parallelized molecular dynamics code is as follows.

for( int i = 0; i < numSteps; ++i ){ move1(); MPI::COMM_WORLD.Allgatherv( &posVect[ partitionStartIndex ], // memory location of data to send numAtomsInPartition, customDataType, &posVect[ 0 ], // memory location to receive data arrayPartitionLengths, arrayPartitionMemDisplacements, custom-DataType ); for( int j = partitionStartIndex; j < partitionEndIndex; ++j ) { for( int k = 0; k < numAtoms; ++k ) { calculateForce(j,k); } } move2();}


The primary change to the algorithm is the partitioning of the pair potential calculation loop and the addition of a MPI communications call. While every process in the cluster runs the same code and indeed the same binary, each pro-cess is assigned a unique integer identifier between zero and the number of pro-cesses in the communications world minus one. With this and in combination with the total number of processes that are in the communications world, it is trivial to partition a list of items over a number of processes. The variables par-titionStartIndex and partitionEndIndex, unique to each process, are integer values that describe the subset of atoms that the pair potentials will be calculated for. The methods move1() and move2(), have also been parallelized using the same constraints, thus parallelizing the entire algorithm. Since the positions of all atoms in the system is essential, each process must communicate to all other processes the updated positions of its subset of atoms. While other methods of communication are available, the MPI function Allgatherv() is especially well suited for this task as it gathers together partitions of variable length and then broadcasts the gathered data to all processes. With these changes, the code can then be run on most any high performance computing system implementing MPI.

RESULTS

With the parallelization of the molecular dynamics code, it has become feasible for clusters of a much larger size to be simulated. As shown in Figure 2, the benefit of parallelization even when run in a small parallel environment is apparent. For a cluster of 8788 atoms over 1000 steps, the compute time for a single process was 35m34s. However, when utilizing all eight processors of the dedicated, rack mounted cluster, the compute time was reduced to 4m38s, a 7.68-fold speedup with an efficiency of 96%. An example of the benefits of the speed increases can be seen in Figure 3. For the 13500 atom cluster, an equivalent, non-parallel simulation would have taken over 80 days to complete, whereas the parallelized version completed execution in less than 11 days. While larger clusters initially display a liquid state, our simulations indicate that only the smaller clusters such as the 133 atom cluster shown in Figure 1 have a liquid state that is indefinitely stable. The three runs for larger clusters shown in the potential energy graph Figure 3 all display a sharply occurring phase transition as indicated by the drop in potential energy. We have observed that the likelihood that a cluster will not have a stable liquid state increases with cluster size. Clusters on the order of thousands of atoms may initially display a liquid state however, our results indicate that a phase transition to the solid is inevitable.

DISCUSSION

A frequent barrier to simulations run in high performance computing envi-ronments is the cost of the hardware. With the prevalence of computer labs in


0

500

1000

1500

2000

2500

3000

3500

0 2 4 6 8 10 12

Rea

l tim

e (s

ec)

N (number of processes)

N=8788, 1000 step test runs

Rack Mount Cluster 1000Mbps NetworkComputer Lab 100Mbps Network

Figure 2. Time for job completion on two different Beowulf clusters for a varying number of processors. The rack mounted cluster on a 1Gbps network scales well as more processors are added. However, the computer lab cluster quickly reaches a saturation point due to network limitations.

-5.3

-5.2

-5.1

-5

-4.9

-4.8

0 1000 2000 3000 4000 5000 6000 7000

Pote

ntia

l Ene

rgy

Reduced time t

Phase Transitions in Large Clusters

n=8788n=10976n=13500

Figure 3. Potential energy as a function of time for three clusters of differing sizes. All three clusters begin as a liquid and make a spontaneous phase transition to the solid as evidenced by the sharp drop in the potential energy of the system.


educational environments, it is tempting to harness the computing power of an unused lab. This can be achieved with no risk of damage to existing software due to BProc and hence Clustermatic’s ability for slave nodes to operate fully without a hard drive. In order to harness unused computer labs on our campus, we used a generic Linux computer running Fedora Core 3, installed Clustermatic 5, and configured it to be a master node. When a lab becomes available for use, we place the master node on the same network subnet as the computer lab. Then for each lab computer, we insert the Clustermatic installation CD, which doubles as a boot disc, and reboot. Since each computer in a lab is already connected to the same network, no communication setup is required. The nodes then regis-ter with the master node and obtain the functional kernel. Once all the nodes have booted, the cluster operates identically to that of our rack mounted cluster. When issued a reboot command from the master node with the CDs no longer in the machines, the lab computers will boot back to their normal operating state having had no data written or read from their hard drives. For a lab of 12 com-puters, the cluster can be up and running in less than 15 minutes and the return to normal operation takes even less time. While the use of computer labs has offered us a great resource, it should be noted that the slower speed and higher latency of the computer lab’s network inhibits the scalability of the program, as seen in Figure 4. In the cluster of 8788 atoms, the saturation point, or point be-yond which processor utilization is not optimal, is quickly reached for the paral-lel code (Saavedra-Barrera et al. 1990). Better performance is achieved however, with clusters of a larger size. As the number of atoms in a simulation increases, the communication to computation ratio decreases, lifting the bottleneck on the slower network. As shown in Figure 4, for each successively larger simulation, the scalability of the code is increased. The ease of utilization and especially the cost effectiveness of the use of computer labs as a high performance computing resource has enabled us to utilize them as a valuable tool for processing large amounts of data with little to no initial costs. While parallelization has decreased drastically the required computation time required for a large system, the simulation still scales as O(N^2/P) with N as the number of particles in the system and P as the number of processors used in the calculation. In order to combat this, we have recently added the common practice of neighbor listing to our molecular dynamics code. Neighbor listing considers all forces beyond a certain radius to be zero and conserves energy by scaling the potential energy down to zero at that radius rather than discontinu-ously dropping the energy to zero. As such, computation time for large systems can be decreased to O(N log(N)/P) or even O(N/P) (Plimpton, 1995).

ACKNOWLEDGEMENTS

Partial funding for this work was supplied by the Augustana Research and Artists Fund (ARAF), the NASA South Dakota Space Grant Consortium, and the Northern Plains Undergraduate Research Center (NPURC). We would like to thank Dan Steinwand and EROS for helpful conversations pertaining to parallel computing. We also would like to acknowledge the previous work done


by John Schneiderman and Colin Taphorn in addition to the ongoing work by Benjamin Bomstad, Jacquelyn Strey, and Joe Coppock.

LITERATURE CITED

Allen, M.P., and D.J. Tildesley. 1987. Computer Simulation of Liquids. Ox-ford University Press.

Dongarra, J., and S.W. Otto. 1995. An Introduction to the MPI Standard. Technical report CS-95-274, University of Tennessee.

Gropp, W., and E. Lusk. 1997. Sowing MPICH: A case study in the dis-semination of a portable environment for parallel scientific computing. The International Journal of Supercomputer Applications and High Performance Computing. 11:103-114.

Hahn, M.Y., and R.L. Whetten. 1988. Rigid-Fluid Transition in Specific Size Argon Clusters. Physical Review Letters. 61:1190-1193

Harris, I.A., R.S. Kidwell, and J.A. Northby. 1984. Structure of Charged Argon Clusters Formed in a Free Jet Expansion. Physical Review Letters. 53:2390-2393.

Hendriks, E. 2002. BProc: The Beowulf Distributed Process Space. In: Pro-ceedings of the 16th Annual ACM International Conference on Supercom-puting. ACM Press. 129-136.

0

0.2

0.4

0.6

0.8

1

1.2

1 2 3 4 5 6 7 8 9 10

rela

tive

time

t/t1

N (number of processes)

Computer Lab Cluster 100Mbps Network

Ideal (1/n)N=8788

N=16384N=32000N=55296

Figure 4. Timing results for various sized clusters run on the computer lab cluster with a 100Mbps network. As the number of atoms increases, parallel efficiency increases due to the simulation scaling linearly with respect to communication and exponentially with respect to computation.


Moore, B.G., and A.A. Al-Quaraishi. 2000. The Structure of Liquid Clusters of Lennard-Jones Atoms. Chemical Physics. 252:337-347.

Philips, J.C., R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E. Villa, C. Chipot, R.D. Skeel., L. Kalé, and K. Schulten. 2005. Scalable Molecular Dynamics with NAMD. Journal of Computational Chemistry. 26:1781-1802.

Plimpton, S.J. 1995. Fast Parallel Algorithms for Short-Range Molecular Dy-manics. Journal of Computational Physics. 117:1-19.

Saavedra-Barrera R.H., D.E. Culler, and T. von Eicken. 1990. Analysis of mul-tithreaded architectures for parallel computing. 2nd Annual ACM Sympo-sium on Parallel Algorithms and Architectures. 169-178 pp.

Schmidt, M.W., K.K. Baldridge, J.A. Boatz, S.T. Elbert, M.S. Gordon, J.H. Jensen, S. Koseki, N. Matsunaga, K.A. Nguyen, S. Su, T.L. Windus, M. Du-puis, J.A. Montgomery Jr. 1993. General Atomic and Molecular Electronic Structure System. Journal of Computational Chemistry. 14:1347-1363.

Schneiderman, J.M., and B.G. Moore. 2004. Molecular Dynamics of Atommic Clusters: An Object Oriented Approach. Proceedings of the South Dakota Academy of Science. 83:101-114

Sterling, T. 2001. Beowulf Cluster Computing with Linux. The MIT Press.Verlet, L. 1967. Computer ‘Experiments’ on Classical Fluids. I. Thermody-

manical Properties of Lennard-Jones Molecules. Phys. Rev. 165:201-214.


PHOSPHORUS FERTILIZATIONIMPACTS ON WHEAT GROWTH

AND SELENIUM BIOAVAILABILITY

Sang H. Lee and J. J. DoolittleDepartment of Plant Science


ABSTRACT

The objective of this study was to determine the effect of phosphorus (P) fertilization on selenium (Se) uptake by wheat (Triticum aestivum L. var. Oxen, Granger, Arapahoe, and Wendy) and the changes in soil Se fractionation. Soil (A horizon) from Presho, South Dakota, was used in a greenhouse experiment. The soil Se was consecutively fractionated into 0.07 mg Se kg-1 soluble Se (F1-Se); 0.08 mg Se kg-1 ligand exchangeable Se (F2-Se); 1.56 mg Se kg-1 acid extractable Se (F3-Se); 1.17 mg Se kg-1 oxidative acid decomposable Se (F4-Se); and 1.36 mg Se kg-1 residual Se (F5-Se). The main treatment was P fertilization at three different rates (0, 100, and 250 mg P kg-1). Total biomass of spring and winter wheat was affected by a wheat variety X P application rate interaction. Phospho-rus fertilization increased the available P in the soil and total P in stems and grain tissues of all wheat varieties. However, there was no statistical difference in total P concentration in wheat tissues between 100 and 250 mg P kg-1 fertilization rate. After wheat harvest, the only significant change in soil Se fraction measured was in the F2-Se. F2-Se significantly decreased with increasing P fertilizer application in both spring and winter wheat. However, total Se in the soil was unaffected by wheat varieties or P fertilizer application. Phosphorus fertilization increased the total absorbed Se in stems and grain in all wheat varieties.

Keywords

Selenium, Fractionation, Wheat, Phosphorus, Bioavailability

INTRODUCTION

Global interest in selenium (Se) has increased over the past few years due to its anti-oxidant and potential anti-cancer attributes in animal and human health (Xu et al., 2004; Yoshizawa et al., 1998). The main source of Se to humans and animals is through plants, although plants do not require Se to complete their life cycle. Wheat (Triticum aestivum) is considered one of the most efficient Se accumulators of the common cereal crops and is one of the most important Se sources for humans (Lyons et al., 2003). Global demand for food products naturally rich in Se is increasing due to consumer interest in reducing their


cancer risk. Management of Se uptake by plants is necessary in high Se areas of the world (e.g., South Dakota) in order to obtain a steady supply of certifiable Se-rich food and feed products. Therefore, understanding crop management influence on Se uptake by wheat is a critical need if producers with seleniferous soil are to capitalize on this natural resource. The level of Se accumulation by plants is governed by environmental fac-tors such as amount and form of Se, plant species, and other nutrients (e.g. Cl-, SO4

2-, and PO43-) (Mantgem et al., 1996; Johnson et al., 2000; Brown et al.,

1982; Hooper et al., 1999). The successful management of Se uptake by plants requires a wide range of knowledge related to sustainable agronomic practices, for example, fertilizer application, irrigation, and crop rotation. Selenium can be adsorbed by the surface of soil humus or clay and can cause low Se availability for plant uptake (Neal, 1995). Selenium uptake by plants or transportation in the soil is highly related to its association with soluble and available Se rather than the total amount of soil Se. Chao et al. (1989) separated soil Se into five fractions: 1) soluble Se (available to plants), 2) ligand exchangeable Se (available to plants), 3) acid extractable Se (conditionally available to plants), 4) oxidative acid decomposable Se (unavailable to plants), and 5) residual Se (unavailable to plants). This procedure was used to determine changes in soil Se storage. In order to manage Se uptake by plants in high Se areas, the relationships between changes in plant available Se in soil and the total Se in plant tissues should be explained. Therefore, this study was conducted to determine the effect of P fertilizer on Se uptake by selected wheat varieties and to investigate the changes in Se fractionation when P fertilizer is applied in naturally seleniferous soil.

METHODS

Soil naturally high in Se was collected from the A horizon (depth 0 - 15 cm) near Presho, in Lyman County, South Dakota (100º07’W longitude and 44º03’N Latitude) and used in a greenhouse study. The collected soil was the Promise soil series (very-fine, smectitic, mesic Typic Haplusterts). Selected physi-cal and chemical properties are shown in Table 1. The soil sample was air dried, and crushed to pass through 6 mm screen for use in a greenhouse pot study. Nitrogen (200 mg N kg-1 using NH4NO3) and phosphorus (0, 100, and 250 mg P kg-1 using KH2PO4) fertilizer was pre-mixed with the soil before it was packed into pots. To improve the physical properties of the high clay soil, perlite was also pre-mixed with the soil at a 1:10 (v/w) perlite to soil ratio before packing into pots. Two different varieties of spring wheat (var. Oxen and Granger) and two varieties of winter wheat (var. Arapahoe and Wendy) were selected to investigate wheat variety differences. Pots were saturated with distilled water and kept in greenhouse for one day. Eight spring wheat seeds were directly planted in each pot, and later thinned to four seeds per pot. Winter wheat seeds were vernalized before planting. The greenhouse temperature was controlled at 25 ± 5 ºC. Soil-moisture content was adjusted regularly to reach approximate field capacity and the plants were harvested after 12 weeks of growth. A completely randomized

design (CRD) with four replications was used. Soil Se was sequentially fraction-ated into five fractions according to the stepwise procedure method described by Chao et al. (1989) (Table 2). The data were statistically analyzed using analysis of variance (ANOVA) with the JMP program. Spring and winter wheat data were treated separately in the statistical analysis due to different growth time periods (January to March for spring wheat, March to June for winter wheat). Least significant differences (LSD) were used to separate means when F-tests were statistically significant (P=0.05).

RESULTS

Phosphorus fertilizer application increased the grain yield in all wheat variet-ies tested (Table 3). Total biomass of both spring and winter wheat was affected by a variety X P application rate interaction. Total biomass and grain yield of wheat treated with P fertilizer increased significantly when compared to the no P application treatment. However, there was no difference between the 100 and 250 mg P kg-1 application rate at the 95% level. Available P in the soil planted with spring wheat significantly increased with P fertilizer application (Figure 1). However, wheat varieties did not show an effect on available P since the absorbed P in wheat tissues was not significantly different between the wheat varieties. This result was same in the soil planted with winter wheat. The P fertilizer application significantly increased total P in wheat stems and grain when compared to the treatment with no P in all varieties


Table 1. Physical and chemical properties of the Promise soil (A horizon) used in this Se bioavail-ability study.

pH EC† TC IC TN NO3-N Avail. P SO4-S CEC Particle size distribution (g kg-1)

(1:1) dS m-1 g kg-1 mg kg-1 cmolc kg-1 Sand Silt Clay

Promise 7.9 0.57 32.7 12.1 2.5 2.91 14.3 6.9 49.7 14 365 621†EC: Electrical Conductivity (1:1 extraction method), TC: Total Carbon, IC: Inorganic Carbon, TN: Total Nitrogen, Avail. P (extracted by sodium bicarbonate), SO4-S (extracted by calcium phosphate), CEC: Cation Exchange Capacity (NaOAC method).

Table 2. Selenium fractionation in the Promise soil (A horizon) used in this study.

F1-Se F2-Se F3-Se F4-Se F5-Se Total Se

mg Se kg-1

Promise 0.07 0.08 1.56 1.17 1.36 4.95 F1-Se: Soluble; Extracted by 0.25M KClF2-Se: Ligand exchangeable; Extracted by 0.1M KH2PO4

F3-Se: Acid extractable; Extracted by 4M HClF4-Se: oxidative acid decomposable; Extracted by KClO3 + conc. HClF5-Se: Residual; Extracted by conc. HF + HNO3 + HClO4


(Figure 2). However, phosphorus fertilizer application of more than 100 mg P kg-1 did not increase the total P concentration in plant tissues. The Se fraction distribution in the Promise soil (A horizon) sampled was: F3-Se (Acid extractable Se) > F5-Se (Residual Se) > F4-Se (oxidative acid decom-posable Se) >> F2-Se (Ligand exchangeable Se) > F1-Se (Soluble Se) (Table 2). Using a classification of plant-availability, the order was: unavailable Se (F4-Se and F5-Se) > conditionally available Se (F3-Se) >> available Se (F1-Se and F2-Se). Table 4 shows the impact of P fertilizer on the sequential Se fractionation of the Promise soil after spring/winter wheat harvesting. In this study, F1-Se (soluble Se) was not significantly affected by wheat variety or P fertilizer ap-plication. On the other hand, the ligand exchangeable Se (F2-Se) concentration significantly decreased with P fertilization regardless of wheat variety. However,

Table 3. Phosphorus (P) fertilization impacts on total biomass and grain yield of wheat (Triticum aestivum L. var. Oxen, Granger, Arapahoe, and Wendy) grown in a Promise soil (A horizon).

P application(mg P kg-1)

SPRING WHEAT WINTER WHEAT

Total biomass Grain yield Total biomass Grain yield

(g pot-1)Oxen Arapahoe

0 6.97 2.94 6.20 1.08100 12.72 5.35 11.97 2.72250 12.02 4.88 12.58 2.66

Mean 10.57 4.39 10.25 2.15Aϕ

Granger Wendy0 5.40 2.21 4.96 2.44

100 14.17 5.43 11.37 5.16250 14.33 4.50 10.98 5.22

Mean 11.30 4.05 9.10 4.27B

P Means0 6.19a 2.58a 5.58a 1.76a

100 13.45b 5.39b 11.67b 3.94b

250 13.18b 4.69b 11.78b 3.94b

LSD†

Varieties NS‡ NS NS 0.78P 1.30 0.87 1.54 0.95

Var. X P 1.83 NS 2.18 NS†LSD0.05 was used to separate means when F-tests were statistically significant. ‡NS = not significantly different at the 95% level.ϕMean comparisons between wheat varieties (capital letters), and between P applications (small letters). n=4 for each variety mean, n=8 for each P means.


Se concentration in F2-Se was not significantly different be-tween 100 and 250 mg P kg-1

fertilizer treatments in both spring and winter wheat culti-vation. The other Se fractions (F3-Se through F5-Se) and to-tal Se were not affected by P application or wheat variety. Selenium concentration in stems and grain of spring or winter wheat was not signifi-cantly affected by P fertilizer application or wheat variety at the 95% level (Table 5). How-ever, the total amount of Se absorbed by spring or winter wheat tissues was affected by P fertilizer application (Figure 3). The amount of Se absorbed by wheat in the treatments with P fertilizer was higher than the treatment with no P.

DISCUSSION

Total biomass and grain yield increased significantly with P fertilizer appli-cation. Wheat yield was increased 55% because P encourages root growth and stimulates both tillering and seedling emergence. However, P fertilizer of more

Figure 1. Available phosphorus (extracted by NaHCO3) in a Promise soil (A horizon) treated with 0, 100, and 250 mg P kg-1 fertilizer application after harvesting of spring wheat (Triticum aestivum L. var. Oxen and Granger). *Bars with the same letter are not significantly different at 95% level.

Figure 2. Phosphorus (P) fertilization impacts on total P content in stems mg P kg-1 and grain of wheat (Triticum aestivum L. var. Oxen, Granger, Arapahoe, and Wendy) grown on a Promise soil (A horizon). *Bars with the same letter are not significantly different at 95% level.


than 100 mg P kg-1 did not increase or decrease the wheat yield. Phosphorus application was applied as high as 1,000 mg P kg-1 (data not shown) without a significant change in total biomass or grain yield beyond the 100 mg P kg-1

rate. According to this study, total P concentration in wheat tissue was significant-ly increased the application of 100 and 250 mg P kg-1 of P fertilization. Phospho-

Table 4. Phosphorus fertilization impact on sequential Se fractionation in a Promise soil (A horizon) after harvesting of spring and winter wheat (Triticum aestivum L. var. Oxen, Granger, Arapahoe, and Wendy).


F1-Se F2-Se F3-Se F4-Se F5-Se Total-Se

mg Se kg-1

Oxen / Arapahoe0 0.04/0.11 0.06/0.14 1.36/1.36 0.87/1.18 1.49/1.43 4.73/4.72

100 0.04/0.11 0.05/0.09 1.45/0.14 0.92/0.31 1.43/0.157 4.93/4.87250 0.05/0.09 0.04/0.08 1.51/1.20 0.86/1.29 1.43/1.38 4.91/4.93

Mean 0.05/0.10 0.05/0.11 1.44/1.23 0.88/1.26 1.45/1.46Granger / Wendy

0 0.05/0.10 0.05/0.11 1.45/1.27 0.95/1.33 1.55/1.40 5.08/4.83100 0.04/0.10 0.03/0.13 1.46/1.25 0.77/1.29 1.53/1.45 4.96/4.88250 0.05/0.10 0.03/0.10 1.47/1.22 1.00/1.22 1.51/1.34 4.92/4.81

Mean 0.05/0.10 0.04/0.11 1.46/1.25 0.91/1.28 1.53/1.40Control

0 0.05/0.11 0.06/0.10 1.55/1.33 1.14/1.26 1.41/1.50 4.76/4.97100 0.06/0.11 0.04/0.10 1.54/1.30 0.92/1.07 1.29/1.51 4.72/4.9250 0.07/0.10 0.04/0.08 1.56/1.16 0.80/1.27 1.38/1.55 4.96/4.98

Mean 0.060.11 0.05/0.10 1.55/1.26 0.95/1.20 1.36/1.52P Means

0 0.05/0.11 0.06a/0.12aϕ 1.46/1.32 0.99/1.26 1.48/1.44 4.86/4.84100 0.05/0.11 0.04b/0.11b 1.48/1.23 0.87/1.22 1.42/1.51 4.87/4.88250 0.06/0.10 0.04b/0.09b 1.51/1.19 0.89/1.26 1.44/1.42 4.93/4.92

LSD†

Varieties NS/NS‡ NS/NS NS/NS NS/NS NS/NS NS/NSP NS/NS 0.02/0.02 NS/NS NS/NS NS/NS NS/NS

Var. X P NS/NS NS/NS NS/NS NS/NS NS/NS NS/NS†LSD was used to separate mean when F-tests were statistically significant (P=0.05).‡NS = not significantly different at the 95% level.ϕMean comparisons between P applications. n=4 for each variety mean, n=8 for each P means.


rus fertilizer application significantly increased plant-available P in soil, which would be available to interact with soil Se. An increase in plant-available P could cause an increase in Se uptake by wheat, since P and Se compete for the same soil adsorption sites and P might desorb or out-compete Se for the adsorption sites. This would make Se more available for plant uptake (Carter et al., 1972). Rajan et al. (1976) demonstrated that PO4

3- was adsorbed three times more than SeO3

2- at low concentrations, due to the PO43- displacing more aqua groups and

thus making the surface less positive. Therefore, more P in soil can be adsorbed on the surfaces of clay or humus, thus the P application increases the possibility for plant roots to absorb Se. In this study, F2-Se concentration decreased with increasing P fertilization. The decreased Se in F2-Se might be taken up by plants or move down with a percolating water front. Total Se concentration in soil was not significantly affected by wheat variety or P fertilizer application although F2-

Table Table 5. Phosphorus fertilization impact on Se concentrations in stems and grain samples of spring and winter wheat (Triticum aestivum L. var. Oxen, Granger, Arapahoe, Wendy) grown on a Promise soil (A horizon). n=4 for each variety mean, n=8 for each P mean.


SPRING WHEAT WINTER WHEAT

Stem Grain Stem Grain

mg Se kg-1

Oxen Arapahoe0 0.158 0.256 0.317 0.442

100 0.162 0.249 0.290 0.500250 0.207 0.280 0.330 0.450

Mean 0.175 0.262 0.312 0.464Granger Wendy

0 0.157 0.239 0.243 0.452100 0.168 0.242 0.260 0.418250 0.206 0.277 0.328 0.512

Mean 0.177 0.253 0.277 0.461P Means

0 0.158 0.248 0.280 0.447100 0.165 0.246 0.275 0.459250 0.207 0.279 0.329 0.481

LSD†

Varieties NS‡ NS NS NSP NS NS NS NS

Var. X P NS NS NS NS†LSD0.05 was used to separate mean when F-tests were statistically significant. ‡NS=not significantly different at the 95% level.


Se concentration decreases with P fertilizer application. This may be due to the relatively small portion of F2-Se compared to the total Se concentration. Selenium concentration in stems and grain of spring and winter wheat were not significantly affected by P fertilizer application or wheat varieties at the 95% level even though the F2-Se concentration in soil was significantly decreased. These data indicate that the amounts of Se absorbed by wheat were increased by P fertilizer application. However, Se concentrations in wheat tissues were not significantly different because the absorbed Se concentration was diluted by the wheat tissue. The Se concentration was higher in grain than stems in all varieties. This makes sense because most of the Se taken up by plant roots is associated with amino acids and proteins (Shrift, 1973), and protein content of grain is usually higher than in plant stems. According to this study, we have found that the P fertilization should be a part of a management plan to control Se uptake by wheat. However, the soil Se concentration was significantly different at the different growing times which suggest a strong environmental influence. Selenium may be transferred into available form as a result of chemical weathering and changes in pH and redox potential (Chao et al., 1989). Therefore, selenium movement in soils or uptake by plants appears to be governed by multiple environmental factors (e.g., tem-perature, irrigation, and nutrition) which should be investigated further.

ACKNOWLEDGEMENTS

This study was funded through a grant from the South Dakota Wheat Commission and the SD Agricultural Experiment Station at South Dakota State University. The authors wish to express their thanks to Mr. Talyor, Dr. Malo, Dr. Schumacher, Dr. Ibrahim, and Dr. Glover for their support in this study.

Figure 3. Phosphorus fertilization impacts on the total amount of Se absorbed in the stems and grain of winter wheat (Triticum aestivum L. var. Arapahoe and Wendy) grown on a Promise soil (A horizon). *Bars with the same letter are not significantly different (P=0.05).


LITERATURE CITED

Brown, T.A., and A. Shrift. 1982. Selenium: Toxicity and tolerance in higher plants. Biol. Rev. 57:59-84.

Carter, D.L., C.W. Robbins, M.J. Brown. 1972. Effects of phosphorus fertiliza-tion on the selenium concentration in alfalfa. Soil Sci. Soc. Am. Proc. 36, pp. 624-628.

Chao, T.T., and R.F. Sanzolone. 1989. Fractionation of soil selenium by sequen-tial partial dissolution. Soil Sci. Soc. Am. J. 53:385-392.

Hopper, J.L., and D.R. Parker. 1999. Plant availability of selenite and selenate as influenced by the competing ions phosphate and sulfate. Plant and Soil. 210: 199-207.

Johnson, G.C., F.M. Fordyce, X. Ge, K.A. Green, and X. Liu. 2000. Selenium distribution in the local environment of selected villages of the Keshhan disease belt, Zhangijakou district, Hebei Province, China. Appl. Geochem. 15:385-401.

Lyons, G., J. Stangoulis, and R. Graham. 2003. Nutriprevention of disease with high-selenium wheat. J. of Australian College of Nutritional & Environ-mental Medicine. 22(3):3-9.

Mantgem, P.J., L. Wu, and G.S. Bañuelos. 1996. Bioextraction of selenium by forage and selected field legume species in selenium laden soils under mini-mal field management conditions. Ecotoxicol. Environ. Saf. 34:228-238.

Neal, R.H. 1995. Selenium. pp. 260-283. In Alloway B.J. (ed.) Heavy metals in soils. Blackie Academic & Professional. Glasgow, UK.

Rajan, S.S.S., and J.H. Watkinson. 1976. Adsorption of selenite and phosphate on an Allophane clay. Soil Sci. Soc. Am. J. 40:51-54.

Shrift A. 1973. Metabolism of selenium by plants and microorganisms. pp. 763-814. In Klayman, D.L. and W.H. Gunther (ed.) Organic Selenium Com-pounds: Their Chemistry and Biology. John Wiley & Sons, New York, NY.

Xu, J., and Q. Hu. 2004. Effect of foliar application of selenium on the anti-oxidant activity of aqueous and ethanolic extracts of selenium-enriched rice. Journal of Agricultural and Food Chemistry. 52: 1759-1763.

Yoshizawa, K., W.C. Willett, S.J. Morris, M.J. Stampfer, D. Spiegelman, E.B. Rimm, and E. Giovannucci. 1998. Study of prediagnostic selenium level in toenails and the risk of advanced prostate cancer. J. Natl. Cancer Inst. 90: 1219-1224.


VITAL SIGNS MONITORING INOUR PARKS: WHAT TO MEASURE?

Nels H. Troelstrup, Jr.Department of Biology & Microbiology


Jill D. RustDepartment of Biology & Microbiology


ABSTRACT

The National Park Service (NPS) Inventory and Monitoring program seeks to define vital signs for the purpose of monitoring and managing park conditions throughout the United States. Aquatic macroinvertebrate biotic integrity ranks high as one potential vital sign of interest to park staff and partnering agencies. The objective of this effort was to identify discriminating measures of inverte-brate community structure which might be used to monitor aquatic biotic in-tegrity. Invertebrate sweepnet samples were collected from 58 large river, stream, spring and bison watering hole habitats during the summers of 2004 and 2005. Invertebrate counts were used to calculate 68 metrics of abundance, diversity, guild structure and pollution tolerance. A metric selection process was imple-mented to maximize between-site discriminatory power, reduce informational redundancy and maximize detection of anthropogenic disturbance. Two sets of 10 metrics each were selected using this process for future monitoring of wade-able and non-wadeable stream sites within the NGPN. Optimal sets consisted of metrics describing community structure, diversity, guild structure and pollution tolerance and all metrics displayed good discriminatory power between sampled sites. A total of 47 significant correlations were observed among wadeable stream metrics and measures of water quality, channel habitat and riparian condition. Only 19 significant correlations were observed for non-wadeable stream metrics. Wadeable stream metrics correlated poorly with stream size but 6 of 10 non-wadeable stream metrics were significantly correlated with drainage area. Several of the metrics selected from this process are currently in use by U.S. EPA, USGS and the states of Nebraska and Wyoming. Thus, the value of NPS monitoring data to partner agencies is high. Selected metrics will be incorporated into habi-tat specific indices of biotic integrity to facilitate vital signs monitoring by the National Park Service.

Keywords

Vital sign, biomonitoring, macroinvertebrates, metrics


INTRODUCTION

Monitoring is an important component of sustainable natural resource management and many state and federal agencies have devoted large investments toward the collection, management and use of monitoring data (Oakley et al. 2003). Monitoring programs are now carefully planned and designed to (1) provide cost-effective information for monitoring changes in natural resource conditions and (2) provide scientifically defensible data for monitoring changes over space and time. The National Park Service (NPS) has initiated its Inventory and Monitoring (I&M) Program to (1) inventory natural resources within park boundaries and (2) initiate the collection of data to monitor change in park conditions (Na-tional Park Service 2006a). Monitoring efforts designed as part of this program focus on “vital signs”, measurable signals that indicate changes that may impair the long-term health of natural resources or ecosystems (National Park Service 2006b). Vital signs are indicators. They tend to be both sensitive to a broad ar-ray of environmental changes and integrative of ecological structure and function across levels of biological organization. Aquatic macroinvertebrate community structure has been identified as one potential vital sign for monitoring NPS aquatic resources (National Park Service 2006c). However, there are many ways to characterize the macroinvertebrate community. Total and relative abundance, community composition, number of species, diversity, guild structure and dis-turbance tolerance measures all provide different perspectives on biotic integrity. A combination of several measures is recommended for development of an index of biotic integrity (Karr and Chu 1999). However, the question of what metrics to include is important as discriminatory abilities and relationships of different measures to environmental change are known to vary with stream size and geo-graphically (Bramblett et al. 2003; Karr and Chu 1999; King and Richardson 2002; Klemm et al. 2002; Larson and Troelstrup 2001). The objectives of this effort were to (1) define the discriminatory power of different measures of mac-roinvertebrate community structure among aquatic systems within parks of the Northern Great Plains Network (NGPN) and (2) define relationships between community measures and aquatic habitat features within the NGPN. Optimal community structure measures (or metrics) are recommended for future moni-toring of stream and large river systems within the NGPN.

STUDY AREA

All sampling sites were located within streams and rivers of the NGPN (Fig-ure 1, Table 1). In many cases, three reaches (40x channel width) were sampled from the mainstem of each system. In some cases, only one or two reaches could be sampled within the park boundary. Parks comprising the NGPN fall within six different ecoregions of Nebraska, North Dakota, South Dakota and Wyoming (Table 1). Consequently, natural differences in physical, chemical, channel habitat and riparian conditions exist among park locations.


METHODS

Methods for this effort were adapted from EPA’s Environmental Monitor-ing and Assessment Program (EMAP) (Lazorchak et al. 2000; Peck et al. 2006). Water quality, habitat and invertebrate assessments were completed twice during 2004 and once during 2005 over the period May 15 – August 1 from 10 cross-channel transects within each sampled reach (40x channel width). A D-frame net (350 um mesh) was used to sample invertebrates from five randomly chosen transects within each reach. These five sweepnet samples were pooled to generate one composite sample for each reach on each of three sam-pling dates. Composite samples were preserved with 70% ethanol and trans-ported to the laboratory for processing. Invertebrate samples were subsampled

Figure 1. Location of individual park units within the Northern Great Plains Network of the National Park Service.


and sorted in the laboratory (Barbour et al. 1999). Invertebrates were sorted into separate vials to be identified to the lowest possible taxonomic level (genus, species) (Merritt and Cummins 1996, Thorp and Covich 1991, Wiggins 1997, Weiderholm 1983). Invertebrate identifications were randomly checked by ca-pable staff and voucher specimens of each taxon were retained. Ten “optimal” community structure metrics were selected for future moni-toring of stream and river sites based on the results of an iterative screening procedure. Optimal invertebrate metrics were those with (1) high between-site discriminatory power, (2) low redundancy with other metrics, (3) low number of undefined values (<25%), (4) high data range among sites, (5) high correlation with water quality and habitat indicators of disturbance and (6) high value to partnering resource agencies. Kruskal-Wallis F-statistics were calculated to evalu-ate among versus within site variability (discriminatory power) of individual community metric attributes. Metric redundancy and relationship to water quality and habitat features were evaluated using Spearman rank correlations (Conover 1980).

Table 1. National park stream and river reaches sampled during 2004 and 2005.

PARK SYSTEM TYPE REACHES

Agate Fossil Beds National Monument Niobrara River Wadeable 1

Badlands National Park Sage Creek Wadeable 3

Devils Tower National Monument Belle Fourche River Non-Wadeable 1

Fort Laramie National Historic Site

North Platte River Non-Wadeable 1

Laramie River Non-Wadeable 2

Deer Creek Wadeable 3

Fort Union Trading Post National Historic Site Missouri River Non-Wadeable 1

Knife River Indian Villages National Historic SiteMissouri River Non-Wadeable 1

Knife River Non-Wadeable 1

Missouri National Recreation River Missouri Non-Wadeable 3

Mount Rushmore National Memorial

Beaver Dam Creek Wadeable 3

Lafferty Gulch Wadeable 3

Grizzly Creek Wadeable 3

Niobrara National Scenic River

Niobrara River Non-Wadeable 3

Berry Falls Wadeable 1

Fort Falls Wadeable 1

Smith Falls Wadeable 1

Scott's Bluff National Memorial North Platte River Non-Wadeable 1

Theodore Roosevelt National Park Little Missouri River Non-Wadeable 3

Wind Cave National Park

Beaver Creek Wadeable 3

Cold Spring Creek Wadeable 3

Highland Creek Wadeable 3


RESULTS

A total of 68 metrics were evaluated for monitoring wadeable and non-wadeable streams of the NGPN. Of the total pool, ten were selected for future wadeable (Table 2) and ten for non-wadeable (Table 3) stream monitoring. Se-lected metrics represent a mixture describing components of taxonomic compo-sition, diversity, functional organization and tolerance to organic pollution. Taxa richness and diversity index measures were among those frequently displaying optimal characteristics. All of the metrics selected displayed high discriminatory power among sampled sites (KW p<0.05, Tables 2, 3).

Table 2. Optimal invertebrate metrics for monitoring wadeable stream conditions within the NGPN. Values presented include minimums, medians, maximums, Kruskal-Wallace F statistics and probability values to evaluate discriminatory power for each metric.

METRIC MIN MED MAX KW F (p)

Percent Non-Insecta 0.0 10.3 100 3.94 (<0.01)EPT:Chironomidae Ratio 0.00 0.81 1.00 5.31 (<0.01)EPT Richness 0 3 11 2.89 (<0.01)Chironomidae Richness 0 3 14 3.06 (<0.01)Shannon H’ 0.00 1.85 2.80 3.15 (<0.01)Predator Richness 0 4 14 4.60 (<0.01)Feeding Guild H’ 0.00 0.93 1.29 2.09 (0.02)Percent Sprawlers 0.0 12.5 66.1 3.15 (<0.01)Habit Guild H’ 0.00 1.14 1.54 2.42 (<0.01)Modified HBI 3.07 5.05 9.60 5.28 (<0.01)

Table 3. Optimal invertebrate metrics for monitoring non-wadeable stream conditions within the NGPN. Values presented include minimums, medians, maximums, Kruskal-Wallace F statistics and probability values to evaluate discriminatory power for each metric.

METRIC MIN MED MAX KW F (p)

Percent EPT 0.0 24.5 93.5 3.98 (<0.01)Percent Chironomidae 0.0 9.6 100 2.95 (<0.01)Total Richness 2 9 26 3.60 (<0.01)Non-Insecta Richness 0 2 6 2.29 (0.02)EPT Richness 0 2 10 5.09 (<0.01)Collector-Filterer Richness 0 1 6 5.30 (<0.01)Collector-Gatherer Richness 1 4 13 2.81 (<0.01)Clinger Richness 0 2 10 4.70 (<0.01)Swimmer Richness 0 2 7 2.83 (<0.01)Modified HBI 2.92 5.23 9.00 2.73 (<0.01)


Optimal wadeable stream metrics (Table 2) displayed 47 significant (p < 0.1) rank correlations with water quality, channel habitat and riparian condition data. The Hilsenhoff Biotic Index (Figure 2a), EPT:Chironomidae ratio, percent sprawlers (Figure 2b), EPT richness, predator richness and feeding guild diversity metrics displayed the greatest number of significant rank correlations. Those water quality and habitat attributes most frequently correlated with invertebrate metrics included stream substrate embeddedness, percent silt-clay channel sub-strate, nitrate-nitrogen and total Kjeldahl nitrogen. None of our optimal stream invertebrate metrics were correlated with bank vegetation density, ammonia-ni-trogen or total suspended solids habitat and water quality data. Optimal non-wadeable stream metrics (Table 3) displayed only 19 signifi-cant (p < 0.1) rank correlations with water quality, channel habitat and riparian condition data. Of 18 water quality and habitat features, only fecal coliform counts, total dissolved solids, total Kjeldahl nitrogen, total suspended solids, spe-cific conductance, channel snag counts and water temperature were significantly (p < 0.1) correlated with invertebrate metrics. Clinger richness and Hilsenhoff Biotic Index values were most frequently correlated at a significant level while richness of collector-gatherers was not significantly correlated with any of the water quality or habitat measures. Those water quality and habitat attributes most highly correlated with invertebrate metrics were channel snag counts and water temperature. Some metrics displayed what appeared to be a threshold relationship with selected water quality and habitat features (Figure 2c). None of the optimal wadeable stream metrics were significantly correlated with stream size as indicated by stream discharge and all metrics except preda-

Figure 2. Relationships of selected invertebrate metrics to water quality and habitat features of wadeable (2a, 2b) and non-wadeable (2c, 2d) streams of the Northern Great Plains Network.


tor richness, percent sprawlers and habit guild diversity were positively related to discharge. However, six of ten optimal non-wadeable stream metrics were significantly correlated with stream size as indicated by drainage area (Figure 2d). In addition, all optimal non-wadeable stream metrics were negatively cor-related with drainage area except percent Chironomidae and Hilsenhoff Biotic Index values which increased as drainage area increased above the sampled reach. Percent Chironomidae, non-insect richness, swimmer richness and Hilsenhoff Biotic Index values displayed no significant relationship with stream size for non-wadeable streams.

DISCUSSION

All of the optimal wadeable and non-wadeable stream metrics selected in this study were capable of discriminating well among streams within their respec-tive classes. Metrics contributing to development of an index of biotic integrity (IBI) should be able to discriminate degraded sites from non-degraded sites (Barbour et al. 1999; Karr and Chu 1999). Because our sites were all located within the boundaries of National Parks and may be expected to be relatively undegraded, we used between-site discriminatory power as a measure of metric ability to detect site differences. Future comparison of these NPS metric values against those from truly degraded sites would further validate their effectiveness in detecting stream impairment (Bramblett et al. 2003; Larson and Troelstrup 2001; Klemm et al. 2002). Both of our optimal metric sets included measures of community composi-tion, diversity, guild structure and pollution tolerance. Representation among these metric categories is necessary to provide an integrated evaluation of bio-logical integrity within sampled streams (Barbour et al. 1999; Karr and Chu 1999). Most metrics within our optimal sets displayed significant correlations with paired measurements of water quality, channel habitat and/or riparian condition. Channel substrate conditions and nutrient enrichment appeared to be strong correlates with macroinvertebrate metrics from wadeable streams while woody snag densities and water temperature appeared to be stronger correlates from non-wadeable sites. These relationships are important to establish the sensitiv-ity of each metric to different possible sources of degradation (Barbour et al. 1999; King and Richardson 2003; Klemm et al. 2002). However, many more significant relationships were observed for wadeable than non-wadeable stream metrics. Better relationships between invertebrate community metrics and habi-tat features may reflect the tighter linkage normally found between water quality, channel habitat and riparian conditions for smaller streams (Vannote et al. 1980; Troelstrup and Perry 1990; Dovciak and Perry 2002). Optimal wadeable stream metrics in this study were not significantly cor-related with stream size. However, several of our non-wadeable stream metrics were significantly correlated with drainage area above the sampled site. Indices of Biotic Integrity are known to vary as a function of stream size even in the absence of degradation (Barbour et al. 1999; Karr and Chu 1999). Our observa-


tion of significant relationships for the non-wadeable stream group is probably a reflection of the greater range of stream sizes within this class. Some of our “non-wadeable” streams were reduced to smaller, shallower channels later in the growing season. Future IBI development by the NPS should account for natural variation in metric values with stream size. Many of the metrics selected from this analysis are currently in use by state and federal monitoring agencies (Table 4). North and South Dakota have not presently defined optimal metrics for monitoring wadeable and non-wadeable streams. However, three of the four optimal metrics selected for use by Nebraska also ranked high from our analysis (Bazata 2005), 5 of 12 metrics selected for Montana streams are members of our optimal sets (Bahls et al. 1992) and Wyo-ming currently reports 16 of the 18 metrics we selected as part of their state monitoring effort (Jeremy ZumBerge, Wyoming Department of Environmental Quality, personal communication). The United States Geological Survey and U.S. Environmental Protection Agency also utilize several of the metrics re-sulting from our optimization effort (Bramblett et al. 2003). Individual parks within the NPS network are unlikely to have resources sufficient to shoulder their entire monitoring burden. Costs associated with monitoring may be offset through collaborative partnering efforts as many of these groups would benefit from sharing data and associated site information. Table 4. Optimal metrics selected for NGPN streams and rivers and use by associated water quality agencies.

METRIC ND NE SD WY USEPA USGS

Percent EPT - - - X X -Percent Chironomidae - - - X - -Total Richness - X - X - XNon-Insecta Richness - - - X - -EPT Richness - X - X X -Collector-Filterer Richness - - - X X -Collector-Gatherer Richness - - - X - -Clinger Richness - - - X - -Swimmer Richness - - - X X -Modified HBI - X - X - -Percent Non-Insecta - - - X X -EPT:Chironomidae Ratio - - - X - XChironomidae Richness - - - X - -Shannon H’ - - - X - XPredator Richness - - - X - -Feeding Guild H’ - - - - - -Percent Sprawlers - - - X - -Habit Guild H’ - - - - - -


Biological monitoring is widely recognized as a necessary component of water resources management (Karr and Chu 1999). Of course, use of biological monitoring requires some knowledge of the flora and fauna. Many of the parks and systems sampled in this effort had no baseline description of their inverte-brate communities. While some community metrics appear to be robust across a number of ecoregions and system types, metric selection procedures are needed to identify those measures which are regionally sensitive, integrate ecosystem properties and correlate well with likely disturbance sources (Klemm et al. 2003; Larson and Troelstrup 2001). These “optimal” metric sets are those most likely to detect changes induced by the predominant disturbance types found within an ecoregion.

ACKNOWLEDGEMENTS

Funding for this project was provided by the USDI National Park Service through a CESU cooperative agreement. Support to the principle investigator was also provided by the South Dakota Agriculture Experiment Station. Thanks are extended to Jill Anderson, Regina Cahoe, Jessica Meisenhoelder and Bret Winterfeld for their assistance in the field and laboratory. Thanks are also ex-tended to Dan Licht and other members of the park staff of the Northern Great Plains Network (NPS), North Dakota Department of Health, Nebraska Depart-ment of Environmental Quality, South Dakota Department of Environment and Natural Resources and Wyoming Department of Environmental Quality for their cooperation and assistance.

LITERATURE CITEDBahls, L., R. Bukantis and S. Tralles. 1992. Benchmark biology of Montana

reference streams. Water Quality Bureau, Department of Health and Envi-ronmental Sciences, Helena, MT.

Barbour, M.T., J. Gerritsen, B.D. Snyder, J.B Stribling. 1999. Rapid bioassess-ment protocols for use in streams and wadeable rivers: periphyton, benthic macroinvertebrates and fish. Second Edition, EPA 841-B-99-002, United States Protection Agency, Office of Water, Washington, D.C.

Bazata, K. 2005. Nebraska stream classification using fish, macroinvertebates, habitat and chemistry evaluations from R-EMAP data, 1997-2001. Final Completion Report, Surface Water Section, Water Quality Division, Ne-braska Department of Environmental Quality, Lincoln, NE.

Bramblett, R.B., T.R. Johnson, A.V. Zale and D. Heggem. 2003. Development of biotic integrity indices for prairies streams in Montana using fish, macro-invertebrate and diatom assemblages. Project Completion Report, Montana Cooperative Fishery Research Unit, U.S. Geological Survey, Department of Ecology, Montana State University, Bozeman, MT.

Conover, W.J. 1980. Practical nonparametric statistics. Second edition, John Wiley & Sons, New York.


Dovciak, A.L. and J.A. Perry. 2002. In search of effective scales for stream man-agement: Does agroecoregion, watershed, or their intersection best explain the variance in stream macroinvertebrate communities? Environmental Management 30: 365-377.

Karr, J.R. and E.W. Chu. 1999. Restoring life in running waters: better biologi-cal monitoring. Island Press, Washington, D.C.

King, R.S. and C.J. Richardson. 2003. Integrating bioassessment and ecological risk assessment: An approach to developing numerical water-quality criteria. Environmental Management 31: 795-809.

Klemm, D.J., K.A. Blocksom, W.T. Thoeny, F.A. Fulk, A.T. Herlihy, P.R. Kaufmann, S.M. Cormier. 2002. Methods development and use of macro-invertebrates as indicators of ecological conditions for streams in the mid-Atlantic Highlands region. Environmental Monitoring and Assessment 78: 169-212.

Larson, A.M. and N.H. Troelstrup, Jr. 2001. Optimal macroinvertebrate metrics for the assessment of a northern prairie stream. Proceedings of the South Dakota Academy of Science 80: 173-183.

Lazorchak, J.M., B.H. Hill, D.K. Averill, D.V. Peck and D.J. Klemm (eds). 2000. Environmental Monitoring and Assessment Program – Surface wa-ters: Field operations and methods for measuring the ecological condition of non-wadeable rivers and streams. U.S. Environmental Protection Agency, Cincinnati, OH.

Merritt, R.W. and K.W. Cummins (eds). 1996. Insects of North America. Ken-dall/Hunt Publishing Co. Iowa, USA.

National Park Service. 2006a Apr 17. Inventory and monitoring: Discovering and protecting America’s natural heritage. http://science.nature.nps.gov/im/in-dex.cfm. Accessed 2006 Aug 15.

National Park Service. 2006b Apr 17. Prioritizing and selecting vital signs – What should be monitored? http://science.nature.nps.gov/im/monitor/Vital-Signs.cfm. Accessed 2006 Aug 15.

National Park Service. 2006c Apr 17. National framework for inventory & monitoring. http://science.nature.nps.gov/im/monitor/NationalFramework.cfm. Accessed 2006 Aug 15.

Oakley, K.L., L.P. Thomas and S.G. Fancy. 2003. Guidelines for long-term monitoring protocols. Wildlife Society Bulletin 31: 1000-1003.

Peck, D.V., A.T. Herlihy, B.H. Hill, R.M. Hughes. P.R. Kaufmann, D.J. Kl-emm, J.M. Lazorchak, F.H. McCormick, S.A. Peterson, P.L. Ringold, T. Magee, and M. Cappaert. 2006. Environmental Monitoring and Assessment Program-Surface Waters Western Pilot Study: Field Operations Manual for Wadeable Streams. EPA/620/R-06/003. U.S. Environmental Protection Agency, Office of Research and Development, Washington, D.C.

Thorp, J.H., and A.P. Covich (eds). 1991. Ecology and classification of North American freshwater invertebrates. Academic Press, Inc., New York.

Troelstrup, N.H., Jr. and J.A. Perry. 1990. Interpretation of scale dependent in-ferences from water quality data. Pages 64-85, W.S. Davis (ed). Proceedings of the 1990 Midwest Pollution Control Biologists Meeting, Chicago, IL.


Vannote, R.L., G.W. Minshall, K.W. Cummins, J.R. Sedell and C.E. Cushing. 1980. The river continuum concept. Canadian Journal of Fisheries and Aquatic Science 37: 130-137.

Weiderholm, T. 1983. Chironomidae of the Holartic Region: Keys and diagno-ses. Part 1. Larvae. Entomologica Scandinavica No. 19.

Wiggins, G.B. 1977. Larvae of the North American caddisfly genera (Trichop-tera). University of Toronto Press, Toronto, Ontario, Canada.


N AND WATER STRESS IMPACT ON HARDRED SPRING WHEAT YIELD AND QUALITY

R. Brunner, D.E. Clay and C. ReesePlant Science Department


ABSTRACT

Water and nitrogen stress impact wheat (Titicum Aestivum) yield and qual-ity. To minimize yield losses from N and water stress it is essential to developing corrective treatment options. The objective of this experiment was to determine the influence of N and water stress on hard red spring wheat, crop reflectance, yield, and quality(protein). A randomized split block experiment with each treatment replicated four times was conducted in 2003. Treatments were four N and two soil moisture regimes. Reflectance data was collected using a cropscan radiometer. 13C isotopic discrimination(∆) was used to asses N and water stress. Reflectance data was then compared to yield and ∆ values. Yields were increased by N rate and were not increased by supplemental irrigation. Reflectance mea-sured at the 5-6 leaf growth stage was highly correlated to N stress. These results indicated that remote sensing can be used to assess N stress. At the boot growth stage, protein content and reflectance were highly correlated. Results from these relationships suggest that corrective N treatment based on crop reflectance can be used to improve wheat quality characteristics. This information can be used to allow for corrective treatments and improve marketing decisions.

Acknowledgements

Funding was partially provided by a Griffith Award and the SD Wheat Commission.

INTRODUCTION

Water and Nitrogen stress are the two most limiting factors to crop growth and development. In production fields they interact to cause variability (Clay et al. 2001b). This variability is the direct consequence of different amounts of available water and N in summit, backslopes , toe slopes, and depressional areas. Matching N to available water is critical (Clay et al. 2001a). Bauer et al. (1965) reported that if stored water was <5.1 cm then wheat did not respond to N and if stored water was >15.2 then the grain fertilizer response was 10 kg grain kg-


1N. Over applying N can reduce wheat yields as well as advance environmental consequences while under applying N can reduce wheat quality and yields. Research conducted in corn (Zea mays) showed crop reflectance can be used to assess N and water stress (Barnes et al. 2000; Clay et al. 2006). This work showed nitrogen stress had a larger influence on reflectance in the green band than water stress, while in the near infrared band (NIR), water stress influenced reflectance more than N stress. These findings suggest that corrective N treat-ments to corn (Zea mays) based on reflectance in the green, red, and NIR bands can be developed. To improve profitability and water quality a similar approach must be developed in wheat. The objective is to determine the influence of N and water stress on wheat crop reflectance, yield, and quality (protein).

MATERIALS AND METHODS

Quantifying N and water stress

Plant 13C discrimination (∆) can be used to evaluate water, nutrients, dis-eases, and soil compaction stresses. Equations for determining nitrogen and wa-ter stress interactions with ∆ have been developed for wheat (Clay et al., 2001a) and corn (Clay et al. 2005). The approach is based on solving two equations: Optimum yield – measured yield =YLWS + YLNS d∆ = YLWS (δ∆/δyield WS) + YLNS (δ∆/δyield NS)where, d∆ is the difference between the ∆ value of a well fertilized plant under low water stress, δ∆/δyield NS is the partial derivation of the line relating ∆ and yield when N limits yield and water stress was constant (Clay et al. 2005; Clay et al. 2001b). The δ∆/δyield WS is the partial derivative relating ∆ and yield when N does not limit yield (Webb et al. 1972). The testing of this method showed that ∆-based YLNS and YLWS values were highly related to measured yield losses to N and water stress (Clay et al. 2005) and plants growing under high water stress had lower ∆ than plants growing under low water stress (Clay et al. 2001). Similar results relating ∆ to grass, durum wheat, and barley yields have been reported (Araus et al. 1999). The approach is based on more 13CO2 being fixed during photosynthesis under water stressed than non-water stressed conditions (Farquhar and Lloyd 1993).

Experimental Design

The field experiment was conducted in 2003 at the Aurora research farm (96º 40’ West and 44º 18’ North). The soil is a Brandt silty clay loam (fine-silty, mixed, frigid, Calcic Hapludoll.) The experimental design was a randomized split block design. Two water treatments and four N rates were used for the experiment. Each treatment was replicated four times. The plot size was 12 X 12 m. Hard red


spring wheat was planted at 90 lbs/acre on April 15th. Herbicides were applied to control weeds at 783mL/ha Puma, 37 mL/ha Harmony GT, and 890 mL /acre MCPA. The four N rates used were 0, 56, 140, and 224 kg N/ha. The two water treatments were natural precipitation and natural precipitation + irrigation. The natural precipitation received 38 cm of rain. The irrigated treatments received 1.9 and 4.4 cm of water on June 19 and July 15. The total natural precipitation and irrigation was 44.3 cm. Crop reflectance was collected on May 20th, June 4th, and June 29th. Growing degree days (base 10ºC) was 1171 GDD. The grain was harvested from an area 18.9 m2. Grain samples were analyzed for yield, protein, moisture, 13C discrimination(∆), N,δ15N using a 20-20 Europa Ratio mass spec-trometer (PDZ Europa, Chesire, UK; Clay et al. 2005). Reflectance was mea-sured on May 20th, June 4th, and June 29th. Anova was conducted using SAS. Relationships between parameters were determined using correlation analysis.

Reflectance

Crop reflectance in the blue (485 ± 68 nm), green (568 ± 70 nm), red (661 ± 57 nm), NIR (840 ± 151 nm), and MIR (1650 ± 195 nm) bands were measured 2 m above the plants at three dates (Clay et al. 2006). Based on these values the reflectance indices were calculated using the equations: NDVI = (NIR-Red)/(NIR + Red); GNDVIS = (NIR-Green)/(NIR+Green)/GNDVIr; NDWI=(NIR-MIR)/NIR+MIR); and NRI=(NIR/green)/(NIRr/Greenr), where GNDVIr, NIRr, Greenr, were taken from well fertilized and water controlled plots (Bausch and Duke, 1996; Shanahan et al. 2001; Jackson et al. 2004). The three other indices were Cgreen [(R800 nm/R700 nm)-1], Cred [(R800 nm/R550 nm )-1], CNIR [(R840-870/Rred(720—740) nm)-1] where R is reflectance (Gitelson et al 2005).

RESULTS AND DISCUSSION

Plant Characteristics

The two inputs in the study were nitrogen and water. The data in table 1 shows the results for each treatment individually and combined. Nitrogen had a significant impact on yield, protein, Yield Loss to N Stress (YLNS), and Yield Loss to Water Stress (YLWS). Increasing the N rate from zero to low increased the yield from 3180 kg/ha to 3680 kg/ha. Associated with this yield increase was reduced YLNS and YLWS. These results suggest that N and water can have additive effects on yields. Increasing the N rate from low to high reduced yields and increased YLWS. The protein values ranged from 11.8% in the 0 N treatments to 15.0% in the high N treatments. Protein concentrations were higher in the irrigated then the non-irrigated treatment. These results were attributed to supplemental irri-gation proving additional N. Supplemental irrigation did not increase yields or reduce YLNS.


Wheat eflectance was impacted by N rate and sampling date. On May 20th, N rate and reflectance had mixed results. Increasing the N rate from 0 to the medium level generally decreased reflectance in the blue, green, red, and NIR bands. (Table 2). Further increases in N reversed these relationships. Mixed re-sults were attributed to interactions among N, water, and the amount of bare soil exposed to the sensor. On June 4th (Table 3) and June 29th (Table 4) reflectance in the blue, green, red, and MIR decreased with N rate. Reflectance in the NIR band had opposite results.

Table 1. The influence of 4 N Rates and soil moisture regime on grain yields, yield losses due to N stress, and yield losses due to water stress in 2003.

N RATEMOISTURE

REGIME YIELD PROTEIN YLNS YLWS

kg/ha0 Natural 3090 11.2 611 493

Low Natural 3630 12.8 241 318Medium Natural 3530 13.8 144 516

High Natural 3460 15.3 95 6310 Irrigated 3270 12.4 494 426

Low Irrigated 3720 13.0 334 136Medium Irrigated 3310 14.7 302 578

High Irrigated 3240 14.6 260 690p value ns <.05 ns ns

—N Rate—0 3180 11.8 552 460

low 3680 12.9 288 227Medium 3420 14.3 223 547

High 3350 15.0 177 660p value <.001 <.0001 <.0002 <.0001

lsd (0.05) 208 0.67 144 148

—Moisture Regime—Natural 3428 13.3 273 490Irrigated 3385 13.7 348 457

p value ns <.05 ns ns


Tabl

e 2.

The

influ

ence

of s

oil m

oist

ure

regi

me

and

4 N

rat

es o

n sp

ectr

al in

dice

s an

d re

flect

ance

for

May

20,

200

3.

N R

ate

Moi

stur

eRe

gim

e

SPEC

TRAL

IND

ICES

REFL

ECTA

NCE

ND

VIG

ND

VIG

ND

VIs

ND

WI

NRI

Cgr

een

Cre

dC

NIR

Blue

Gre

enRe

dN

IRM

IRkg

/ha

0N

atur

al0.

280.

380.

99-0

.24

0.99

2.06

1.36

1.49

6.6

9.0

11.6

20.1

32.7

Low

Nat

ural

0.30

0.40

1.04

-0.2

31.

042.

171.

411.

546.

38.

710

.920

.232

.6M

ediu

mN

atur

al0.

320.

411.

07-0

.21

1.07

2.23

1.45

1.58

6.1

8.5

10.5

20.4

31.5

High

Nat

ural

0.28

0.39

1.01

-0.2

31.

012.

091.

381.

516.

79.

111

.620

.732

.90

Irriga

ted0.

280.

380.

99-0

.22

0.99

2.06

1.36

1.49

6.8

9.3

11.8

20.8

32.7

Low

Irriga

ted0.

300.

401.

03-0

.21

1.03

2.14

1.41

1.53

6.6

9.2

11.4

21.1

32.2

Med

ium

Irriga

ted0.

280.

381.

00-0

.23

1.00

2.08

1.37

1.50

6.7

9.1

11.5

20.4

32.4

High

Irriga

ted0.

280.

391.

00-0

.22

1.00

2.08

1.38

1.50

6.7

9.2

11.6

20.6

32.1

p va

lue

nsns

ns<.

1ns

<.00

01ns

nsns

nsns

nsns

—N

Rat

e—0

0.28

0.38

0.99

-0.2

30.

992.

061.

361.

496.

79.

211

.620

.432

.7lo

w0.

300.

401.

03-0

.22

1.03

2.15

1.41

1.54

6.5

8.9

11.2

20.7

32.4

Med

ium

0.30

0.40

1.03

-0.2

21.

032.

151.

411.

546.

48.

811

.020

.431

.9H

igh0.

280.

391.

00-0

.22

1.00

2.09

1.38

1.50

6.7

9.2

11.6

20.7

32.5

p va

lue

<.00

01<.

0001

<.00

01<.

0001

<.00

01<.

0001

<.00

02<.

0001

<.00

01<.

0001

<.00

01<.

0005

<.00

01lsd

(0.0

5)0.

010.

010.

030.

010.

010.

050.

030.

030.

40.

50.

70.

91.

1—

Moi

sture

Reg

ime—

Nat

ural

0.30

0.40

1.03

-0.2

31.

032.

141.

401.

536.

48.

811

.120

.332

.4Irr

igated

0.28

0.39

1.00

-0.2

21.

002.

091.

381.

506.

79.

211

.620

.732

.4p

valu

ens

nsns

nsns

nsns

ns<.

1ns

<.1

nsns

100 Proceedings of the South Dakota Academy of Science, Vol. 85 (2006)Ta

ble

3. T

he in

fluen

ce o

f soi

l moi

stur

e re

gim

e an

d 4

N r

ates

on

spec

tral

indi

ces

and

refle

ctan

ce fo

r Ju

ne 4

, 200

3.

N R

ate

Moi

stur

eRe

gim

e

SPEC

TRAL

IND

ICES

REFL

ECTA

NCE

ND

VIG

ND

VIG

ND

VIs

ND

WI

NRI

Cgr

een

Cre

dC

NIR

Blue

Gre

enRe

dN

IRM

IRkg

/ha

0N

atur

al0.

630.

610.

920.

050.

833.

832.

382.

533.

96.

56.

127

.024

.3Lo

wN

atur

al0.

720.

671.

010.

151.

024.

662.

853.

023.

45.

92.

030

.022

.0M

ediu

mN

atur

al0.

750.

701.

040.

211.

235.

163.

123.

283.

15.

74.

531

.620

.5H

ighN

atur

al0.

710.

671.

000.

161.

004.

632.

843.

013.

45.

95.

129

.921

.80

Irriga

ted0.

620.

600.

900.

060.

803.

712.

322.

474.

16.

86.

427

.324

.4Lo

wIrr

igated

0.71

0.67

1.01

0.17

1.03

4.74

2.90

3.06

3.5

6.1

5.1

31.0

22.1

Med

ium

Irriga

ted0.

690.

660.

980.

130.

964.

452.

722.

883.

56.

15.

429

.322

.6H

ighIrr

igated

0.70

0.67

1.00

0.17

1.00

4.65

2.86

3.02

3.5

6.1

5.3

30.4

21.7

p va

lue

nsns

ns<.

1ns

<.00

01ns

nsns

nsns

nsns

—N

Rat

e—0

0.62

0.61

0.91

0.05

0.82

3.77

2.35

2.50

4.0

6.6

6.3

27.1

24.3

low

0.71

0.67

1.01

0.16

1.02

4.70

2.88

3.04

3.4

6.0

5.1

30.5

22.0

Med

ium

0.72

0.68

1.01

0.17

1.04

4.81

2.92

53.

083.

35.

94.

930

.421

.5H

igh0.

710.

671.

000.

161.

004.

642.

853.

013.

56.

05.

230

.121

.7p

valu

e<.

0001

<.00

01<.

0001

<.00

01<.

0001

<.00

01<.

0002

<.00

01<.

0001

<.00

01<.

0001

<.00

05<.

0001

lsd (0

.05)

0.03

0.02

0.03

0.04

0.09

0.13

0.23

0.23

0.3

0.3

0.5

1.6

1.1

—M

oistu

re R

egim

e—N

atur

al0.

700.

660.

990.

140.

994.

572.

802.

963.

46.

05.

229

.622

.1Irr

igated

0.68

0.65

0.97

0.13

0.95

4.39

2.70

2.86

3.7

6.3

5.6

29.5

22.7

p va

lue

nsns

nsns

nsns

nsns

<.1

ns<.

1ns

ns


Tabl

e 4.

The

influ

ence

of s

oil m

oist

ure

regi

me

and

4 N

rat

es o

n sp

ectr

al in

dice

s an

d re

flect

ance

for

June

29,

200

3.

N R

ate

Moi

stur

eRe

gim

e

SPEC

TRAL

IND

ICES

REFL

ECTA

NCE

ND

VIG

ND

VIG

ND

VIs

ND

WI

NRI

Cgr

een

Cre

dC

NIR

Blue

Gre

enRe

dN

IRM

IRkg

/ha

0N

atur

al0.

880.

810.

940.

520.

729.

135.

165.

331.

73.

92.

438

.312

.2Lo

wN

atur

al0.

910.

850.

980.

570.

8911

.42

6.42

6.62

1.7

3.6

2.0

43.6

11.8

Med

ium

Nat

ural

0.91

0.85

0.98

0.58

0.89

11.4

06.

476.

681.

73.

62.

043

.511

.6H

ighN

atur

al0.

920.

850.

990.

580.

8911

.40

6.47

6.68

1.7

3.6

2.0

43.5

11.6

0Irr

igated

0.89

0.83

0.96

0.55

0.78

10.0

75.

655.

821.

73.

82.

240

.011

.6Lo

wIrr

igated

0.91

0.84

0.97

0.57

0.85

10.7

16.

086.

261.

73.

62.

041

.811

.3M

ediu

mIrr

igated

0.92

0.86

1.00

0.60

0.98

12.5

37.

017.

221.

63.

41.

945

.611

.5H

ighIrr

igated

0.92

0.86

1.00

0.61

1.00

12.8

27.

247.

461.

63.

41.

846

.911

.5p

valu

ens

0.08

0.09

ns0.

030.

06ns

0.09

nsns

nsns

ns—

N R

ate—

00.

890.

820.

950.

530.

759.

605.

415.

581.

73.

92.

339

.211

.9lo

w0.

910.

840.

980.

570.

8711

.06

6.25

6.44

1.7

3.6

2.0

42.7

11.6

Med

ium

0.92

0.85

0.99

0.59

0.93

11.9

76.

746.

951.

73.

52.

044

.511

.6H

igh0.

920.

860.

990.

590.

9712

.37

7.00

7.20

1.6

3.5

1.9

45.9

11.7

p va

lue

<.00

01<.

0001

<.00

01<.

0001

<.00

01<.

0001

<.00

01<.

0001

<.00

5<.

0001

<.00

01<.

0001

<.1

lsd (0

.05)

—M

oistu

re R

egim

e—N

atur

al0.

910.

840.

970.

560.

8610

.97

6.20

6.39

1.7

3.7

2.1

42.5

11.8

Irriga

ted0.

910.

850.

980.

580.

9011

.53

6.49

6.69

1.6

3.6

2.0

43.6

11.5

p va

lue

nsns

nsns

nsns

nsns

nsns

nsns

ns


Correlation Coefficients

On May 20, 2003 reflectance was generally not correlated to N rate, yield, protein, ∆, YLNS, and YLWS (Table 5). The lack of correlation was attributed to small plants and that bare soil confounded the signal (Chang et al. 2004). Different results were observed on June 4th, where reflectance was correlated to N rate, yield, protein, ∆, and YLNS were correlated. YLWS was not correlated to reflectance (Table 6). YLNS had a higher correlation to reflectance then yield, protein, and ∆. Yield was correlated at the 0.01 probability levels for all values except, blue, green, red, and MIR which were correlated at the 0.05 level. Protein showed correlated values at the 0.05 level for GNDVI, GNDVIs, NDWI, Cred, CNIR, and MIR. ∆ showed correlation at the 0.05 level for NDVI, GNDVI, GNDVIs, NDWI, Cred, CNIR, blue, green, red, and MIR. Similar results were observed on 29 June (Table 7). All of the reflectance values correlated to the N rate at the 0.01 probability level except for MIR which was not correlated to N rate. Protein generally had stronger correlations to reflec-tance than yield, ∆, and YLNS except for blue and MIR. Protein was correlated at the 0.01 probability level to all of the reflectance values. Blue was correlated to protein at the 0.05 probability level and MIR was not correlated. 13C discrimina-tion was correlated to NDVI, GNDVI, GNDVIs, NRI, Cgreen, Cred, and NIR at the 0.05 probability level. YLNS was correlated at the 0.01 probability level for all values except for Cgreen, blue, NIR, and MIR which showed correlation at the 0.05 level. Yield and YLNS values had higher correlation to reflectance

Table 5. Correlation coefficients on May 20th, 2003. 3-4 Leaf Stage. Values indicated with * and ** were significant at the .05 and .01 level.

N Rate Yield Protein C13Dis YLNS YLWS

NDVI 0.06 0.06 -0.09 -0.20 -0.18 0.12GNDVI 0.09 0.04 -0.04 -0.22 -0.17 0.12GNDVIS 0.09 0.04 -0.04 -0.22 -0.17 0.12NDWI 0.22 .484* 0.22 -0.08 -0.33 -0.24NRI 0.09 0.03 -0.05 -0.22 -0.16 0.12Cgreen 0.09 0.08 -0.03 -0.22 -0.19 0.10Cred 0.12 0.08 -0.03 -0.23 -0.22 0.13CNIR 0.11 0.05 -0.05 -0.23 -0.20 0.14Blue -0.01 0.11 0.14 0.24 0.08 -0.21Green -0.03 0.14 0.12 0.26 0.07 -0.24Red -0.02 0.09 0.14 0.24 0.10 -0.22NIR 0.07 0.29 0.16 0.18 -0.09 -0.26MIR -0.14 -0.14 -0.03 0.28 0.22 -0.07


Table 6. Correlation coefficients on June 4th, 2003. 5-6 Leaf Stage. Values indicated with * and ** were significant at the .05 and .01 level.

N Rate Yield Protein C13Dis YLNS YLWS

NDVI 0.55** 0.60** 0.42 -0.43* -0.77** 0.07GNDVI 0.58** 0.61** 0.45* -.44* -0.77** 0.06GNDVIS 0.58** 0.61** 0.45* -.44* -0.77** 0.06NDWI 0.59** 0.61** 0.46* -.43* -0.78** 0.06NRI 0.53* 0.61** 0.41 -0.42 -0.76** 0.05Cgreen 0.54** 0.60** 0.42 -0.42 -0.76** 0.06Cred 0.54** 0.60** 0.42* -0.42* -0.77** 0.07CNIR 0.55** 0.61** 0.43* -0.43* -0.77** 0.06Blue -0.51* 0.47* -0.35 0.48* 0.67** -0.14Green -0.53* -0.45* -0.37 0.49* 0.66** -0.14Red -0.52* -0.53* -0.36 0.45* 0.72** -0.11NIR 0.50* 0.68** 0.41 -0.28 -0.75** -0.05MIR -0.60** -0.47* -0.44* .52* 0.71** -0.16

Table 7. Correlation coefficients on June 29th, 2003. Boot Stage. Values indicated with * and ** were significant at the .05 and .01 level.

N Rate Yield Protein C13Dis ylns ylws

NDVI 0.70** 0.359 0.70** -0.42* -0.54** 0.12GNDVI 0.72** 0.316 0.76** -0.46* -0.55** 0.18GNDVIS 0.72** 0.316 0.76** -0.46* -0.55** 0.18NDWI 0.62** 0.404 0.65** -0.30 -0.57** 0.10NRI 0.72** 0.265 0.76** -0.46* -0.54** 0.23Cgreen 0.70** 0.263 .73** -0.44* -0.53* 0.23Cred 0.71** 0.267 0.74** -0.47* -0.56** 0.25CNIR 0.72** 0.268 0.75** -0.48* -0.56** 0.25Blue -0.44* -0.127 -0.50* 0.12 0.51* -0.37Green -0.67** -0.219 -0.78** 0.37 0.58** -0.34Red -0.70** -0.340 -0.71** 0.38 0.56** -0.17NIR .68** 0.309 0.67** -0.47* -0.45* 0.09MIR -0.17 -0.357 -0.29 -0.20 0.52* -0.10


at the V4 to V5 (Feekes 3-5) growth stage than reflectance measured at the late boot stage (Feekes 10.5). This reduction in correlation from V4-5 to late boot was attributed to reflectance at the late boot being influenced by two vegetation types (leaves and head), whereas at the V4-V5 reflectance was only influenced by leaves. Similar findings are observed as corn goes from vegetation to reproductive growth stages. YLWS was not correlated to reflectance at any date. These results are in agreement with the Anova analyis. The lack of correlation between YLWS and reflectance was expected because supplemental irrigation did not increase yields. Results from the reflectance data suggests that reflectance can be used to assess N stress in wheat.

SUMMARY

Nitrogen and water regularly interact and influence grain yields. The ability to quantify YLNS and YLWS by ∆ provides a method to asses yield variability. Yields were increased by N fertilizer and were not increased with supplemental irrigation. Protein content was increased by both N fertilizer and supplemental irrigation. Early in the growing season (May 20) reflectance and the plant parameters were not correlated. As the season progressed to June 4 data showed that reflec-tance could be used as a tool to identify N stress. At late boot, (Feekes 10.5) reflectance data could be used to assess wheat quality. Reflectance data collected at V4-V5 could be used for developing corrective N rates while data collected at late boot could be used for marketing purposes. In many situation premiums are paid for wheat with high protein content. Utilizing protein information, farmers will be better able to manage their resources.

REFERENCES

Araus J.L., G.A. Slafer, I. Romagos. (1999). Durum wheat and barley yields in antiquity estimated from 13C discrimination of archaeological grains: A case study from western Mediterranean basin. Aust. J. Plant Physiol. 26:345-352

Bauer A., R.A. Young, J.L. Ozbum. (1965). Effects of moisture and fertilizer on yields of spring wheat and barley. Agron. J. 57:354-356.

Barnes E.M., T.R. Clarke, S.E. Richard Colaizzi P.D., J. Hoberland, M. Kostoze-wski, P. Waller, C. Choi, R. Riley, T. Thompson. R.J. Lascana, H. Li, M.S. Moran. (2000). Coincidental detection of crop water stress, nitrogen status, and canopy density using ground-based multispectral data. In P.C. Robert et al. ed.) Proc. Int. Conf. On Press. April 5th, Bloomington MN, 16-19 July 2000 [CD-Rom] ASA, CSSA, and SSSA, Madison WI.

Bausch W.C., H.R. Duke. (1996). Remote sensing of plant nitrogen status in corn. Trans. ASAE 29:1869-1875.


Chang J., S.A. Clay, D.E. Clay. (2004). Determining weed free and weed infest-ed areas of a soybean field using NIR reflectance. Weed Sci. 52:642-648.

Clay D.E., S.A. Clay, Z. Liu, C. Reese. (2001b) Spatial variability of C-13 isotopic discrimination in corn (Zea mays). Comm. Soil Sci. Plant Anal. 32:1813-1827

Clay, D.E., S.A. Clay, D.J. Lyon, J.M. Blumenthal. (2005) Can 13C discrimina-tion in corn (Zea mays) grain be used to characterize intra-plant competition for water and nitrogen? Weed Sci. (53:23-29).

Clay D.E., R.E. Engel, D.S. Long, Z. Liu. (2001a) Nitrogen and Water Stress Interact to Influence Carbon-13 Discrimination in Wheat. Soil science So-ciety of America Journal 65, 1823-1828

Clay D.E., Ki-In Kim, J. Chang, S.A. Clay, K. Dalsted. (2006) Characterizing water and nitrogen stress in corn using remote sensing. Agron. J. (98:579-587).

Farquhar G.D., L. Lloyd. (1993). Carbon and oxygen isotopes effects in the exchange of carbon dioxide between terrestrial plants and the atmosphere. P. 47-60 J. 8R. Ehleringer (ed.) Stable isotopes and plant carbon-water rela-tionships. Academic Press, New York

Gitelson A.A., A. Vipa, D.C. Rundquist, V. Ciganda, T.J. Arkebauer. (2005). Remote estimation of canopy chlorophyll content in crops. Geophys. Res. Lett. 32:L08403 doi:10.1029/2005GL022688.

Jackson T.J., D. Chen, M. Cosh, F. Li, M. Anderson, C. Withall, P. Doriaswamy, E. Hunt. (2004). Vegetative water content mapping using Landsat derived normalized water index for corn and soybean. Remote Sens. Environ. 92:475-482

Shanahan J.F., J.S. Schepers, D.D. Francis, G.E. Varvel, W.W. Wilhelm, J.M. Tringe, M.R. Schlemmer, D.J. Major. (2001). Use of remote sensing to estimate corn grain yield. Agron. J. 93:583-589.

Webb R.A. (1972) Use of the boundary line in the analysis of biological data. J. Hort. Sci. 47:309-319.


THE EVALUATION OF OPERANDSAND ITS PROBLEMS IN C++

Dan Day and Steve ShumComputer Science Department


INTRODUCTION

In beginning C++ courses, the precedence of operators is a main focus. It is taught that each operator has certain precedence. It is also said that mathematical operators are math-like in operation while the non-math operators just fit in the mix somewhere. More or less, these courses focus on how the operators work on the values of operands. However, this focus is not the whole story. All implementations have some method of evaluating operands in order to retrieve the values that they store. What is important is that C++ does not provide a strict ordering of when the operands are evaluated, unlike operator precedence. Many assume that the evaluation of operands will coincide with operator precedence, but this is not guaranteed. It is important to know that C++ leaves all ordering of evaluating operands unspecified and a certain subset of those evaluations involving side ef-fects undefined.

EVALUATING OPERANDS

C++ is not strict about how an implementation evaluates operands (and side effects, which will be considered later). It allows for a lot of freedom. Here is the relevant section in the C++ standard: Except where noted, the order of evaluation of operands of individual operators and subexpressions of individual expressions, and the order in which side effects take place, is unspecified. In general, the order of evaluating operands is unspecified. This means that there is no certain order an implementation must follow. For example, Java has a guarantee that all oper-ands are evaluated in a left-to-right order. In C++, an implementation is free to evaluate operands in a left-to-right, right-to-left, or any other imagineable order. Moreover, an implementation does not have to be consistent with evaluating operands. It may order the operands of one expression different than the same, exact expressions that may appear later in the program.What does this all mean? Take this code for example:


int g(){std::cout << “g()” << std::endl;return 0;}int f(){std::cout << “f()” << std::endl;return 0;}int main(){return g() + f();}

It is unknown at compile-time if the implementation will call function g or f first. Consequently, it is unknown whether which text will appear first on the screen. There are two possible ways an implementation could order the evalua-tion of operands, and both are equally available to the implementation.It must be noted that a definite output will occur. The program is well-formed, valid according to the C++ standard, so it must produce some type of output and not crash. This behavior must be contrasted with undefined behavior, which will be discussed shortly.

SEQUENCE POINTS, SIDE EFFECTS,AND UNDEFINED BEHAVIOR

Side Effects and Sequence Points

As shown, all evaluation of operands is done in an unspecified way. However, there is a subset of evaluation that actually can produce undefined behavior.This situation arises when a programmer uses expressions that cause side effects to the operands. C++ defines side effects to be:

“Accessing an object designated by a volatile lvalue, modifying an object, calling a library I/O function, or calling a function that does any of those operations are all side effects, which are changes in the state of the execution environment.”

Evaluation of an expression might produce side effects. In addition, C++ leaves the time of resolution of the side effect (being the actual point in the execution of the program where the side effect is actually applied) up to the implementation. All that C++ requires is that an implementation must resolve all side effects between two sequence points at some point no later than the next sequence point. A sequence point is:


“At certain specified points in the execution sequence called sequence points, all side effects of previous evaluations shall be complete and no side effects of subsequent evaluations shall have taken place.”

Sequence points are mostly found at semicolons, functions, and some opera-tors. This is important because many operators do not have a sequence point; so many side effects can take place before an implementation must resolve the side effects. Sequence points are only found at these places:

“There is a sequence point at the completion of evaluation of each full-expres-sion. When calling a function (whether or not the function is inline), there is a sequence point after the evaluation of all function arguments (if any) which takes place before execution of any expressions or statements in the function body. There is also a sequence point after the copying of a returned value and before the execution of any expressions outside the function. Sev-eral contexts in C++ cause evaluation of a function call, even though no cor-responding function call syntax appears in the translation unit. The sequence points at function-entry and function-exit (as described above) are features of the function calls as evaluated, whatever the syntax of the expression that calls the function might be.

In the evaluation of each of the expressionsa && ba || ba ? b : ca , busing the built-in meaning of the operators in these expressions, there is a sequence point after the evaluation of the first expression.”

The importance of sequence points relates to what a programmer can and cannot do between sequence points. While C++ is very loose towards the imple-mentation, C++ requires programmers to realize that there are certain rules of how many side effects a programmer can apply to a single operand.

THE PROBLEM WITH SEQUENCE POINTS

A problem arising sequence points is that an operand can have multiple side effects between any two sequence points. Consider this piece of code:

int i = 0;i = 5 + i++;

Within the second statement (since we are unconcerned with the declara-tion), there is only one sequence point, the semicolon, and two side effects, the assignment and increment operators. The problem is what is i after the semico-


lon: 5 or 1? Since an implementation can order side effects freely, a programmer would think either is possible. However, the above code actually produces undefined behavior. Undefined behavior simply means that a program is no longer well-formed and may pro-duce any result, including crashing during execution. The fact that the code at-tempts to put multiple side effects on i between two sequence points is the cause of undefined behavior. Why is the above program undefined? It is because C++ only allows so many side effects on any one operand between sequence points. Here is the relevant section:

“Between the previous and next sequence point a scalar object shall have its stored value modified at most once by the evaluation of an expression. Fur-thermore, the prior value shall be accessed only to determine the value to be stored. The requirements of this paragraph shall be met for each allowable ordering of the subexpressions of a full expression; otherwise the behavior is undefined.”

A programmer must be careful about this rule. An operand can only be modified once between sequence points, although its value can be accessed mul-tiple times. It is important to realize the seriousness of this rule. Breaking in this rule can easily create a program that is ill-formed making it not portable and potentially unusable.

EVALUATION, SIDE EFFECTS, AND OPERATOR OVERLOADING

The rules of evaluating operands are mostly inherited from C. Mostly when the topic of sequence points and undefined behavior is brought up; it is talked about in terms of using built-in types, such as ints, chars, and pointers. However, C++ adds operator overloading, so it is worthy to wonder how sequence points related to user-defined objects and operator overloading.

Consider this example:

class X{public:X operator++(int); //Implementation omittedX& operator=(const X& other);};int main(){//...X x;x = x++;//...}


With a built-in type, this would clearly be undefined behavior. However, the type is user-defined. It is necessary to consider the nature of operator overload-ing. The fact that programmers use the operators directly is a matter of conve-nience, but the compiler must translate the code to use the functions defined in the class in order for the program to work. Consider if main was re-written as follows:

int main(){//...X x;x.operator=(x.operator++(0));}

For this article, operator overloading is nothing more than just calling member functions of a class. A compiler would translate the original code to one that uses functions. The presence of functions adds several sequence points, one before the calling of each function and one after each function returns. The existence of many sequence points does not cause undefined behavior like if a built-in type would have been used instead.

SUMMARY

While the order of evaluation of operands is a very technical subject in the C++ language, it is nevertheless an important one. Its subtle rules can lead to unexpected behavior in a program. It is necessary for a programmer to remember that it is incorrect to assume any specific ordering of evaluating operands or to assume any specific resolution of side effects. Assuming anything outside of the very loose rules C++ provides can lead to potential problems later during the development and porting of a software application.

REFERENCES

INTERNATIONAL STANDARD, ISO/IEC14882. Programming languages — C++. 10/15/2003


EFFECTS OF GRAZING ON SMALL MAMMAL ABUNDANCE IN EASTERN SOUTH DAKOTA

Wesley W. Bouska and Jonathan A. JenksDepartment of Wildlife and Fisheries Sciences


ABSTRACT

Variation in small mammal abundance was compared between ungrazed and grazed pasture in Brookings County, South Dakota from 28 September through 27 October 2005. Total relative abundance and absolute abundance of small mammal populations did not differ (p=0.476) between grazed and ungrazed pastures. However, there were significantly more (p<0.05) masked shrews (Sorex cinerus) trapped on ungrazed than grazed pasture. Conversely, there were signifi-cantly more (p<0.05) deer mice (Peromyscus maniculatus) and meadow voles (Mi-crotus pennsylvanicus) trapped on grazed than ungrazed pasture. Because little information is available on effects of grazing on small mammal communities in the Northern Great Plains, these results will aid range and wildlife managers in developing local grazing programs that maintain the abundance and diversity of small mammal populations.

Keywords

Grazing, meadow voles, Microtus ochrogaster, Microtus pennsylvanicus, prairie voles, small mammal abundance, South Dakota.

INTRODUCTION

In the Midwest, cattle production is an economically important practice. However, effects of grazing on small mammal populations in this area have not been sufficiently addressed. Geier and Best (1980) determined that habitat alterations such as those made by cattle, result in loss of cover and changes in wildlife populations. Bock et al. (1984) noted that small mammal abundance was greater in ungrazed than grazed semi-desert grasslands. By contrast, other studies have shown that grazing had minimal effects on small mammal popula-tions and their habitats (Samson et al., 1988; Oldemeyer et al., 1988). In light of these contradicting studies, it is likely that the effect of grazing on small mammal populations is dependent upon site location, habitat type, grazing pressure, and the small mammal community itself. The purpose of this study was to investigate the effects of cattle grazing on small mammal abundance in Brookings County, South Dakota. Because there


is little information on this subject in the Northern Great Plains, study results will aid range and wildlife managers in developing local grazing programs that maintain abundance and diversity of small mammal populations. We predicted that small mammal abundance would be greater in ungrazed pasture than in grazed pasture.

STUDY AREA

The study area for this project was located approximately 0.80 km north of Brookings, South Dakota (elevation 502.01 m above mean sea level, Township 110, Range 050, Section 14). The site was bordered on the north by Six-Mile Creek and was characterized by northern mixed grass prairie. The plant com-munity included but was not limited to: smooth brome grass (Bromus inermis), yellow foxtail (Setaria glauca), Canada thistle (Cirsium aruense), and showy milk-weed (Asclepias syriaca). One half of the study area served as cattle pasture and experienced heavy grazing one month per year, the other half was idle pasture (K. VanderWal, SDSU Range Sciences, personal communication).

METHODS

Small mammal abundance was measured on grazed and ungrazed pasture with Museum Special snap traps (Woodstream Corporation, Lititz, Pennsylva-nia). Traps were placed in transects 20-m away from any edge habitat, with one transect each on grazed and ungrazed pasture. Transects were placed randomly and each contained 25 snap traps with 5-m spacing between traps. Trapping occurred weekly for 5 weeks with two trap nights per trap per week; trapping began on 28 September 2005 and was completed on 27 October 2005. Traps were baited with peanut butter and checked every 24 hours after placement (Wiener and Smith, 1972). All specimens were handled in accordance with the 1998 Animal Care and Use Guidelines developed by the American Society of Mammalogists (ASM, 1998). After sampling, captures per trap night were calculated for each species by transect. Relative abundance was calculated as the total number of individuals caught per 1000 trap nights: Relative Abundance = (Total captures/Total trap nights) x 1000 (Lancia et al., 2005). Absolute abundance was calculated as the total number of animals captured on each treatment divided by the number of sample days (10). Systat (SPSS 2000) was used to compare relative and total abundance using a pooled variance t-test. A Pearson Chi-squared analysis also was conducted to compare the frequency distributions of small mammals cap-tured on grazed and ungrazed habitats.


RESULTS

Over 250 trap nights, 58 small mammals were trapped on the ungrazed area; captures included masked shrews (Sorex cinereus) (n=19), short-tailed shrews (Blarina brevicauda) (n=12), prairie voles (Microtus ochrogaster) (n=13), and meadow voles (Microtus pennsylvanicus) (n=13). On the grazed area, 250 trap nights produced 47 small mammals including meadow voles (n=20), prairie voles (n=15), short-tailed shrews (n=6), masked shrews (n=4), and deer mice (Peromyscus maniculatus) (n=2). Relative abundance for grazed and ungrazed pastures was 188 and 232, respectively. A pooled variance t-test with a 95% confidence interval indicated there was no difference (t = -0.728, df = 18, p-value = 0.476) between small mammal absolute abundance on grazed versus ungrazed transects (Table 1). A Pearson Chi-squared test with a 95% C.I. was used to compare species frequencies between grazed and ungrazed pasture. Masked shrew captures represented 33.3% of total captures on ungrazed pasture. On grazed pasture, masked shrews accounted for only 8.6% of total captures which was significantly less (p< 0.05). Meadow voles made up 42.6% of total captures on the grazed side but only 22.8% of total captures on the ungrazed pasture; these percentages were significantly different (p< 0.05). Finally, deer mice constituted 4.3% of captures on the grazed treatment. No deer mice were captured on the ungrazed transect (Table 2). Capture frequencies of short-tailed shrews and prairie voles did not differ (p>0.05) across the treatments. The grazed pasture had slightly higher species richness (5 species captured) than the ungrazed pasture (4 species captured).

Table 1: Abundance of small mammals on grazed and ungrazed pasture, Brookings County, South Dakota, 28 September through 27 October 2005.

TOTAL RELATIVE ABUNDANCE OF SMALL MAMMALS

Day Grazed Ungrazed

1 40 402 0 1203 360 2004 320 2405 160 2406 160 4807 440 2808 200 36609 160 16010 40 200


DISCUSSION

The type of rotational grazing system used on the grazed section of the study site did not negatively affect small mammal abundance. Severson and Urness (1994) determined that wildlife benefit from grazing systems when compared to seasonal or year-long grazing. Grazing systems can also be tailored to benefit certain species (Severson and Urness, 1994). Light to moderate grazing has been shown to increase plant diversity (Smith et al., 1996). Grasses and seeds are preferred food items of the meadow vole (Burt and Grossenheider, 1980). This potential increase in grasses and seeds produced by grazing could have been responsible for the higher frequency of meadow voles found on the grazed pasture. Masked shrews tend to select moist habitats (Burt and Grossenheider, 1980). The higher frequencies of masked shrews on the ungrazed habitat could possibly be explained by the sites close proximity to Six-Mile Creek when compared to the grazed transect. Verme (1958) also identified moisture as a factor affecting distribution of the masked shrew. In his studies, he noted that shrew abundance could vary between fundamentally similar environments as a result of subtle dif-ferences in microhabitat. The higher frequency of deer mice (n=2) captured on the grazed pasture might not be biologically significant. This low capture rate (1 per 125 trap nights) indicates that deer mice were uncommon at this study site. However, other studies have shown grazing to significantly increase relative abundance of deer mice (Matlack et al., 2001; Meaney et al., 2002). Deer mice select grazed areas because they have less plant litter and foraging for seeds can be accom-plished more efficiently (Matlack et al., 2001). In our study, grazing only occurred 1 month out of the year, allowing 11 months of recovery time for plant and animal communities. We believe that ad-ditional research regarding grazing and its impact on small mammal communi-ties would help define thresholds of grazing intensity that affect small mammal communities.

Table 2: Small mammal species frequencies and row percents on grazed and ungrazed pasture, Brookings County, South Dakota, 28 September through 27 October 2005.

SPECIESFREQUENCIES

GRAZEDFREQUENCIES

UNGRAZEDROW PERCENTS

GRAZEDROW PERCENTS

UNGRAZED

masked shrews 4 19 8.511 33.333

short-tailed shrews 6 12 12.76 21.053

deer mice 2 0 4.255 0

prairie voles 15 13 31.915 22.807

meadow voles 20 13 42.553 22.807

Total 47 57 100 100


MANAGEMENT IMPLICATIONS

Results of this study indicated that high use, low frequency grazing does not affect small mammal abundance. Range or wildlife managers could implement this grazing system in areas where small mammal community health or threat-ened species of small mammals are of concern.

ACKNOWLEDGEMENTS

We thank the Department of Wildlife and Fisheries Sciences at South Dakota State University for their support. Nate Gosch provided helpful com-ments on an earlier draft of our paper.

LITERATURE CITED

American Society of Mammalogists. 1998. <http://www.mammalsociety.org>. Accessed 2005 Nov 29.Bock, C. E., J. H. Bock, W. R. Kenny, and V. M. Hawthorne. 1984. Responses

of birds, rodents, and vegetation to livestock exclosure in a semidesert grass-land site. Journal of Range Management 37:239-242.

Burt, W. H., and R. P. Grossenheider. 1980. A field guide to the Mammals. Third edition. Houghton Mifflin Company, New York, New York, USA.

Geier, A. R., and L. B. Best. 1980. Habitat selection by small mammals of riparian communities: evaluating effects of habitat alterations. Journal of Wildlife Management 44:16-24.

Lancia, R. A., W. L. Kendall, K. H. Pollock, and J. D. Nichols. 2005. Estimating the number of animals in wildlife populations, pg. 106-153 in C.E. Braun, editor, Techniques for Wildlife Investigations and Management. The Wild-life Society, Bethesda, Maryland.

Matlack, R. S., D. W. Kaufman, and G. A. Kaufman. 2001. Influence of graz-ing by bison and cattle on deer mice in burned tallgrass prairie. American Midland Naturalist 146:361-368.

Meaney, C. A., A. K. Ruggles, N. W. Clippinger, and B. C. Lubow. 2002. The impact of recreational trails and grazing on small mammals in the Colorado piedmont. The Prairie Naturalist 34:115-136.

Oldemeyer, J. L., and L. R. Allen-Johnson. 1988. Cattle grazing and small mam-mals on the Sheldon National Wildlife Refuge, Nevada. Proceedings of a symposium on the management of amphibians, reptiles, and small mammals in North America. Pages 391-398. U.S. Forest Service General Technical Report. RM-166.

Samson, F. B., F. L. Knopf, and L. B. Hass. 1988. Small mammal response to the introduction of cattle into a cottonwood floodplain. Proceedings of a symposium on the management of amphibians, reptiles, and small mammals in North America. Pages 289-299. U.S. Forest Service General Technical Report. RM-166.


Severson, K. E., and P. J. Urness. 1994. Livestock grazing: a tool to improve wildlife habitat. Pages 232-249 in M. Vavra, W. A. Laycock, and R. D. Pieper, editors. Ecological Implications of Livestock Herbivory in the West. Society for Range Management, Denver, Colorado, USA.

Smith, G., J. L. Holechek, and M. Cardenas. 1996. Wildlife numbers on excel-lent and good condition Chihuahuan desert rangelands: an observation. Journal of Range Management 49:489-493.

SPSS Science and Marketing Department. 2000. Systat version 10. SPSS incorporated,Chicago, Illinois.

Verme, L. J. 1958. Localized variation in masked shrew abundance. Journal of Mammalogy 39:149-150.

Wiener, J. G., and M. H. Smith. 1972. Relative efficiencies of four small mam-mal traps. Journal of Mammalogy 53:868-873.


DETECTION OF BISON/CATTLEHYBRIDIZATION IN CUSTER STATE PARK

BREEDING BULLS USING MICROSATELLITEAND MITOCHONDRIAL DNA MARKERS:

TOOLS FOR CONSERVATION MANAGEMENT

Cynthia M. Anderson, Traci L. Berger, Forrest Cain and Shane K. Sarver Center for the Conservation of Biological Resources

Black Hills State UniversitySpearfish, SD 57799

ABSTRACT

Previous studies have shown the presence of cattle genes in the Custer State Park (CSP) bison herd. In this study the degree of hybridity in breeding bulls, was assayed in extant breeding bison bulls as well as samples of archived blood from Custer State Park bulls dating from 1994-2001. Hybridity was determined using both mitochondrial and microsatellite markers. The mitochondrial assay utilized a highly conserved region as a positive control and a cattle specific re-gion in multiplexed PCR reactions. Presence of the cattle specific PCR product and the control PCR product was diagnostic for the presence of maternal cattle DNA. Microsatellite markers were diagnostic for cattle and bison and were also used to assess hybridity. Using seven previously developed microsatellite markers, 90 bison bulls were genotyped on an automated genetic analyzer. The extent of hybridity detected with the mitochondrial test was 3.3% and hybridity detected using microsatellite markers was 10%. The extended goal of the pres-ent project is to utilize DNA-based assessments of hybridity in the selection of breeding bulls in the CSP bison herd.

Keywords

Bison, Custer State Park, hybrid, introgression

INTRODUCTION

Historically, the population of American Bison (Bison bison) on the Great Plains of North America had been estimated in excess of 50 million. During the expansion westward, bison were slaughtered for their meat, and valuable hides. By the late 1800’s, bison were nearly extinct, numbering less than 1000 individuals. During the late 1800s, a small number of private ranchers attempted to rescue the bison from extinction by harboring them on their ranches. These individuals are responsible for saving the bison from extinction and for creating small foundation herds. It is known that many of these ranchers hybridized bison


to domestic cattle (Bos taurus) experimentally. These small, private herds were then used to stock federal and state bison populations. Pete Dupree, a South Dakota rancher, captured five bison calves in the Dakota’s to found his own bison herd in 1881. James “Scotty” Philip purchased the entire Dupree herd in 1901 which had grown to about 83 bison. In 1914, 36 bison from the estate of Scotty Philip were sold to the state of South Dakota and immediately brought onto Custer State Forest lands. These 36 bison became the founders of what is now the Custer State Park herd (Dary, 1974). It is well documented that there was a history of hybridization between cattle and bison in the Dupree herd (Coder, 1975), but the status of the 36 bison used to found the CSP herd is not known. There is no record of experimental hybridization following the foundation of the Custer State Park herd, however, recent studies have indicated the existence of cattle genes in the current herd (Polziehn et al, 1995; Ward et al, 1999; Halbert et al, 2005). These cattle genes are thought to trace back to the original 36 animals used to found the CSP herd. Hybridity can be assessed in bison using microsatellite and mitochondrial DNA (mtDNA) markers. Mitochondrial DNA (mtDNA) is haploid, inherited maternally, lacks recombination and provides a relatively simple approach to DNA testing. Mitochondrial DNA will reveal hybidity in maternal pedigrees, but the absence of cattle mtDNA is not proof of bison purity because of this maternal mode of transmission. Nuclear microsatellite markers have become the preferred molecular marker in many genetic studies because of their high muta-tion rate, abundance in the genome and the simple codominant phenotypes. In addition, microsatellite markers can be used to determine the extent of nuclear introgression that is not revealed with mitochondrial analysis. The present study was undertaken at the request of park biologists to in-vestigate the genetic status of the breeding bulls in CSP to evaluate the possible inclusion of DNA testing in the future selection of breeding bulls. Seven mic-rosatellite markers diagnostic for cattle and bison alleles were used to screen each of the 90 breeding bulls in addition to a mitochondrial assay for the presence of cattle mtDNA.

METHODS

DNA extraction

Archived whole blood samples for 90 breeding bulls were obtained from Custer State Park. These individuals represented the animals used for breed-ing over a period of eight years from 1994-2001. Total genomic DNA from whole blood was extracted using a Qiagen DNeasy Tissue Extraction Kit as per manufacturer’s instructions (Qiagen Inc, Valencia, CA).

Mitochondrial Assay

A mitochondrial assay was used to detect the presence of cattle mitochon-drial DNA in bison. PCR was performed for each individual blood sample with 2 mitochondrial markers. One primer pair (Bov16s2878/Bov16s2284, table 1)


amplified a segment of the mitochondrial genome of any animal in the Bovidae subfamily (Derr et al, 1992). This marker was used as a control since it amplified segments of both the domestic cattle, and the bison mitochondrial genome. The second primer pair (c-spec16264/c-spec15907, table 1) specifically amplifies a portion of the control region of the mitochondrial genome of cattle (Bos taurus) and does not amplify bison DNA (Ward, 2000). Therefore, the presence of this second PCR product in a bison sample indicates the presence of cattle mitochon-drial DNA. Each amplification was carried out in 20 ul reactions containing 0.5 U Taq Polymerase, 200 uM each dNTPs, 1.5 mM MgCl2, Tris buffer (Promega, Madison, WI), and 0.5 uM of each primer (Integrated DNA Technologies, Coralville, IA). Thermal profiles for these two primer sets are specified in table

Table 1. Primers and amplification protocols used for amplification of mitochondrial sequences and for nuclear microsatellite loci. * Expected size range for C-spec and Bov 16S are as presented in Ward, 2000; the size range for the microsatellite loci are as presented in Halbert et al, 2005.

LOCUS PRIMER SEQUENCEAMPLIFICATION

PROTOCOL

SIZERANGE

CATTLE*(bp)

SIZERANGEBISON*

(bp)

Mitochondrial markers

C-spec F: 5’-AGCTAACATAACACGCCCATAC-3’ d 357 no product

R: 5’-CCTGAAGAAAGAACCAGATGC-3’

Bov 16SrDNA F: 5’-CCCGCCTGTTTATCAAAAACAT-3’ d 594 594

R: 5’-CCCTCCGGTTTGAACTCAGATC-3’

Nuclear microsatellite markers

BM1314 F: 5’-HEX/TTCCTCCTCTTCTCTCCAAAC-3’ b 143-167 137

R: 5’-ATCTCAAACGCCAGTGTGG-3’

BM4513 F: 5’-HEX/GCGCAAGTTTCCTCATGC-3’ a 139-166 132-134

R: 5’-TCAGCAATTCAGTACATCACCC-3’

BMS2270 F: 5’-HEX/CTGCGTTAACACCCCACC-3’ c 80-98 66-70

R: 5’-GCAGGAAGGCTGATGCAC-3’

BMS4040 F: 5’-HEX/GTCCATAGGGTCACACAGAGTC-3’ a 85-99 75

R: 5’-CCAAATCTTACCATAGCAAAGG-3’

CSSM042 F: 5’-6-FAM/GGGAAGGTCCTAACTATGGTTGAG-3’ c 173-217 167-171

R: 5’-ACCCTCACTTCTAACTGCATTGGA-3’

CSSM36 F: 5’-6-FAM/GGATAACTCAACCACACGTCTCTG-3’ a 162-185 158

R: 5’-AAGAAGTACTGGTTGCCAATCGTG-3’

TGLA227 F: 5’HEX/CGAATTCCAAATCTGTTAATTTGCT-3’ b 79-106 73

R: 5’-ACAGACAGAAACTCAATGAAAGCA-3’ a) 94°C 2 min; 35 cycles of 94°C 15 sec, 54.4°C 15 sec +1sec/cycle, 72°C 30 sec; 72°C 2 min. b) 94°C 2 min: 35 cycles of 94°C 15 sec, 54.4°C 15 sec, 72°C 30 sec; 72°C 2 min. c) 94°C 2 min; 35 cycles of 94°C 15 sec, 60.1°C 15 sec, 72°C 30 sec; 72°C 2 min. d) 94°C 2 min; 35 cycles of 94°C 30 sec, 55°C 30 sec, 74°C 1 min.


1. Amplification of cattle DNA was performed with each group of reactions as a positive control. The presence or absence of PCR products from the above reactions was determined using 1.5 % agarose gel electrophoresis. The gels were then visualized and photographed. Figure 1 depicts the expected banding pat-tern for each of the marker sets when used to detect cattle, bison or cattle/bison hybrids.

Microsatellite marker amplification

Seven microsatellite markers were used to detect hybridity in the breeding bull samples from CSP. The forward primers for each of the seven loci were 5’ end-labeled with a fluorophore (HEX or 6-FAM) for genotype analysis. Amplifica-tion was performed in 20 ul reactions using an Eppendorf Gradient thermocycler (Eppendorf, Westbury, NY). Marker names, primer sequences and thermal profiles are outlined in Table 1. Each 20 ul reaction contained 0.5 U Taq Poly-merase, 200 uM each dNTPs, 1.5 mM MgCl2, Tris buffer (Promega), and 0.5 uM of each primer (Integrated DNA Technologies).

Figure 1. Expected results of the mitochondrial assay. This figure depicts the band patterns with bovine primers Bov16s2878/Bov16s2284 and cattle specific primers c-spec16264/c-spec15907 for bison, cattle and bison/cattle hybrid. The Bov primers amplify a fragment approximately 700 bp in length of DNA from any animal of the bovidae subfamily. The c-spec primers amplify a frag-ment approximately 400 bp in length from cattle DNA only, and are therefore diagnostic for the presence of cow mitochondrial DNA.


Genotyping

Fragments amplified for each individual bison bull were separated, sized, and genotyped on an ABI 3100 Avant Automated Genetic Analyzer (Applied Biosystems, Foster City, SD). ROX 350 (Applied Biosystems) was used as an internal size standard. Data collection was done using Genescan software version 3.7 (Applied Biosystems). Data analysis for allele determination was done using Genotyper v3.7 (Applied Biosystems). Genotyping was performed by Western South Dakota DNA Core Facility (WestCore), Black Hills State University, Spearfish, SD.

RESULTS

Mitochondrial Assay

The mitochondrial assay was performed on 90 bison bulls. Each of the 90 bulls assayed showed the presence of a band with the Bov16s2878/Bov16s2284 primers, and were therefore positive for the control DNA. Three of the 90 bulls assayed showed the presence of a band amplified with the c-spec16264/c-spec15907 primers. The c-spec primers were specific for cattle mitochondrial DNA and, therefore, indicate that 3.3% of the bison tested using this method are hybrids.

Microsatellite marker amplification

A total of seven microsatellite markers were used to genotype 90 bison and cattle controls. These microsatellite markers are able to amplify both bison and cattle DNA. Hybridity is detected by the presence or absence of alleles that are specific to bison or cattle. Table 2 depicts the genotypes of 11 bison individuals found to be hybrids. In this set of bison bulls, only three of the seven microsatel-lite markers used detected hybridity in nine animals, a hybrid frequency of 10%. At locus BM4513, two bulls had a cattle allele. At locus BM1314, three bulls had a cattle allele. At locus BMS2270, four bulls had a cattle allele. There were no bison that had more than one cattle allele, and only one bison with a cattle allele at locus BM4513 also tested positive for cattle mtDNA. The total amount of hybridity, mtDNA and microsatellite data combined, detected among the 90 bulls tested was 12.2% (Table 3).

DISCUSSION

In the past decade, with the advance of molecular genetics techniques, the impact of hybridization between wild and domestic or nonindigenous flora and fauna on the wild species populations has raised new concerns for wildlife conservation. It has been recognized that hybridization events can threaten a rare species existence (Rhymer and Simberloff, 1996). While hybridization in


naturally occurring populations of species can provide a source for novel genetic variation that is sometimes beneficial to species evolution (Lewontin and Birch, 1966; Dowling and Secor, 1997), anthropogenic hybridization poses a threat. The most extreme example is the complete admixture of the two parental species that leaves no pure stock to conserve (Allendorf and Leary, 1988; Allendorf et al., 2001; Rhymer et al., 1994). In the case of bison, we know that a bottleneck occurred in the North American population as the species was decimated and population numbers plummeted from more than 50 million animals to 1000 animals. The subsequent hybridization of some of the remaining animals with domestic cattle and the ensuing introgression has left its mark in some of the remaining wild and captive herds. However, none of the remaining populations

Table 2. Genotypes of the eleven hybrid bison. * = the Cattle microsatellite alleles that were amplified; NP = no PCR product was amplified; af = amplification failed.

SAMPLENUMBER BM1314 BM4513 BMS2270 C-SPEC

BR950156 (132/132) (127/143*) (059/063) NPBR950131 (132/154*) (127/129) (061/061) NPBR960163 af (127/129) (061/079*) NPBR970301 (132/132) (127/129) (059/061) 400bpBR970308 (132/132) (127/127) (061/081*) NPBR970327 (132/132) (127/129) (059/061) 400bpBR980201 (132/132) (127/143*) (063/063) 400bpBI990055 (132/154*) (127/129) (059/063) NPBI990052 (132/154*) (127/129) (063/063) NPBI990076 af (127/127) (059/081*) NPBI210057 (132/132) (129/129) (059/081*) NP

Table 3: Comparison of the frequency of hybrids detected in bison from this study with previous studies.

LOCUSNUMBER OF

INDIVIDUALS

HYBRIDFREQUENCYDETECTED

PUBLISHEDFREQUENCY

mtDNA and nuclear loci 11 12.22% namtDNA 3 3.33% 20.6%1

Nuclear loci 9 10% 30.8%2

BM1314 3 3.4% 4.05%2

BM4513 2 2.22% 0.00%2

BMS2270 4 4.44% 2.56%2

1Ward et al., 1999; 2Halbert et al., 2005


that contain remnant hybrid animals, including the Custer State Park popula-tion, have undergone introgression to this extent. For example, Halbert et al., 2005 tested 100 microsatellite markers covering 29 of the 30 bison autosomes and the X chromosome for their ability to differentiate cattle and bison genomic regions. They found only 14 diagnostic regions, of which only seven regions over five autosomes showed evidence of cattle introgression. The Halbert study included 14 bison populations, five of which were found to have some small degree of cattle introgression. While overall CSP bison harbored cattle alleles on four different autosomes, only five individuals in the Halbert study were found to harbor two introgressed genomic regions, and only seven were found to harbor a single introgressed region. A Y chromosome marker BYM-1 has also been developed and when tested revealed that there has been no Y chromosome introgression in any of the herds where mtDNA haplotypes were found (Ward et al., 2001). We argue that implementation of a management plan for natural herds that includes testing individuals for hybridity would be a valuable tool to work toward ridding herds containing hybrid animals of remnant cattle DNA. This approach would be especially useful for herds such as CSP that are known to harbor a low frequency of hybrid animals and which are an important source for supplementing and founding new herds. The results presented here are consistent with previous findings that reported evidence of bison-cattle hybridization that occurred over 100 years ago (Polziehn et al, 1995; Ward et al, 1999; Halbert et al, 2005). The hybridization event(s) most likely occurred in bison owned by Peter Dupree in the late 1800’s. Bison from this herd were subsequently sold to James “Scotty” Philip and are known to be the founders of the CSP herd. The hybridity seen today represents the remnant introgressed cattle DNA. An estimated 30-35 generations have passed since the CSP herd was established. While it is not known how many indi-viduals of the initial herd were hybrids, previous studies and historical accounts provide ample evidence that at least some of the original 36 individuals were second generation hybrids (Garretson, 1938; Coder 1975; Dary 1989; Polziehn, 1995; Ward, 1999; Halbert, 2005). With the passage of 30 or more generations, and many of the animals backcrossing to wild-type bison, the amount of genetic introgression would decrease and become more fragmented throughout the ge-nome. These results show that using seven microsatellite loci in combination with the mtDNA detected about ½ the amount of hybrid animals as previous studies using a few more loci. To increase the power of detecting hybrid indi-viduals, we could include the use of other diagnostic microsatellite markers. Certainly, the choice of genetic markers used in a management program that looks for hybridity in individuals is very important. The mtDNA marker is only able to detect the maternally inherited mitochondrial DNA. While this test can be useful, it is important to reiterate that a positive mtDNA test confirms hybridity, but a negative mtDNA test does not mean the animal tested is not a hybrid. In order to rule out hybridity a number of diagnostic nuclear mark-ers should be used. A management strategy that utilizes DNA testing to detect hybridity in a bison herd could benefit from the utilization of mtDNA markers, nuclear microsatellite markers, and other diagnostic nuclear markers that are cur-rently being developed (R. Schnabel, University of Missouri-Columbia, personal


communication) together in the most efficient way possible. Perhaps the most cost effective strategy would be to run the mtDNA test first to identify quickly and inexpensively hybrids harboring cattle mtDNA since those individuals would very likely harbor cattle genomic DNA. Those animals testing negative for the mtDNA can then be tested for hybridity using nuclear markers. Data from these markers could be used in various aspects of management from deter-mining kinship and paternity, to potentially identifying the presence of disease resistant loci. While genotype information on all individuals in the herd is valuable in-formation, the implementation of this is problematic. In a herd such as that residing in Custer State Park or in Yellowstone National Park external tags or markings are not consistent with a non-domestic animal and detracts form the aesthetics. Thus the challenge is not in obtaining the genetic information for each individual but rather in the adoption of an efficient method of identify-ing individuals in the field. Once this is accomplished there is the potential for gathering other useful information such as herd kinship, sire contribution to the herd, herd structure (maternal groupings), inbreeding, and selective culling. The implementation of a management strategy that utilizes genetic technology will greatly benefit the conservation of bison in a natural state, and the preservation of the existing genetic diversity of the species.

ACKNOWLEDGEMENTS

We would like to thank Custer State Park for providing samples and his-torical information. Special thanks to Gary Brundidge, Ron Walker, and Chad Kremer. Funding for this research was provided in part by grants from the Na-tional Science Foundation (NSF-MRI #0320651), and the U.S. Environmental Protection Agency #CR-83152201. Partial funding was also provided by NIH Grant Number 2 P20 RR016479 from the INBRE Program of the National Center for Research Resources. Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NIH.

LITERATURE CITED

Allendorf, F.W., R.F. Leary, P. Spruell, and J.K. Wenburg. 2001. The problems with hybrids: setting conservation guidelines. Trends Ecol. Evol. 16: 613-6622

Allendorf, F.W., Leary, R.F. 1988. Conservation and distribution of genetic vari-ation in a polytypic species, the cutthroat trout. Conserv. Biol. 2:170-184

Coder, G.D. 1975. The national movement to preserve the American buffalo in the United States and Canada between 1880 and 1920. PhD Thesis, Ohio State University.

Dary, David A. 1974. The Buffalo Book; The Full Saga Of The American Animal p. 231-232


Dowling, T.E., Secor, T.L. 1997. The role of hybridization and introgression in the diversification of animals. Ann. Rev. Ecol. Syst. 28:593-619

Garretson, M.S. 1938. The American Bison: The Story of its Extermination as a Wild Species and its Restoration Under Federal Protection. New York Zoological Society, New York.

Lewontin R.C., and Birch, L.C. 1966. Hybridization as a source of variation for adaptation to new environments. Evolution 20:315-336

Polziehn, R.O., Strobeck, C.M., Sheraton, J., Beech, R. 1995. Bovine mtDNA discovered in North American Bison Populations. Conserv. Biol. 9:1638-1643

Rhymer, J.M., D. Simberloff. 1996. Extinction by hybridization and introgres-sion. Ann. Rev. Ecol. Syst. 27: 83-109.

Rhymer, J.M., M.J. Williams, and M.J. Braun. 1994. Mitochondrial analysis of gene flow between New Zealand mallards (Anas platyrhynchos) and grey ducks (A. superciliosus). Auk 111: 970-978.

Ward, T.J., Skow, L.C., Gallagher, D.S. Schnabel, R.D., Nall, C.A., Kolenda, C.E., Davis, S.K., Taylor, J.F., Derr, J.N. 2001. Differential introgression of uniparentally inherited markers in bison populations with hybrid ancestries. Animal Genetics 32:89-91

Ward, T.J., Bielawski, J.P., Davis, S.K., Templeton, J.W., Derr, J.N. 1999. Iden-tification of domestic cattle hybrids in wild cattle and bison species: a general approach using mtDNA markers and the parametric bootstrap. Animal Conservation 2:51-57

Wilson, P.J. et al. 2000. DNA profiles of the eastern Canadian wolf and the red wolf provide evidence for a common evolutionary history independent of the gray wolf. Can. J. Zool. 78: 2156-2166.


DIETARY YEAST CULTURESUPPLEMENTATION DURING INITIAL

REARING OF THREE SALMONID SPECIES

Michael E. Barnes and Brian FletcherSouth Dakota Department of Game, Fish and Parks

Spearfish, SD 57783

Dan J. DurbenBlack Hills State University

Spearfish, SD 57799

Stuart G ReevesDiamond V. Mills

Cedar Rapids, IA 52404

ABSTRACT

Three experiments were undertaken to evaluate the effects of supplementa-tion with yeast-based DVAqua (Diamond V Mills, Cedar Rapids, Iowa) during the initial feeding of three salmonid species. Inclusion levels of 0.125% and 0.25% DVAqua into commercial starter feeds were compared to control diets containing 0.25% grain blank during the initial feeding of domesticated Erwin strain rainbow trout O. mykiss, feral Soda Lake brown trout Salmo trutta, and feral Lake Oahe fall Chinook salmon Oncorhynchus tshawytscha. A pattern of increased growth and decreased feed conversion with increasing levels of dietary DVAqua were observed in each of the experiments, although small samples likely precluded statistical significance. However, brown trout length and weight did increase significantly with increasing DVAqua concentrations. Mortality rates also decreased, albeit not significantly, during the rainbow trout and Chinook salmon experiments. Mean percent mortality was 40% less in the Chinook salm-on tanks receiving DVAqua-supplementation compared to the control tanks, but this difference only approached statistical significance. DVAqua supplementa-tion during initial feeding likely produced benefits in salmonid growth and survival.

Keywords

Yeast, initial feeding, DVAqua, rainbow trout, brown trout, Chinook salm-on, Oncorhynchus mykiss, Oncorhynchus tshawytscha, Salmo trutta


INTRODUCTION

Yeast, and yeast-containing products, have been used in aquaculture diets for a variety of reasons. Dietary yeast may improve fish growth (Noh et al. 1994; Lara-Flores et al. 2002; Li and Gatlin 2003, 2004; Ghosh et al. 2005; Barnes et al. In Press), replace fish meal as a protein source (Dabrowski et al. 1980; Rumsey et al. 1990; Rumsey et al. 1991; Sanderson and Jolly 1994; Oliva-Teles and Gon-calves 2001; Cheng et al. 2004), change muscle color (Johnson et al. 1977, 1980; Whyte and Sherry 2001; Storebakken et al. 2004), and act as an immunostimu-lant (Siwicki et al. 1994; Nakano et al. 1995,1999; Ortuño et al. 2002; Li and Gatlin 2004, 2005). Yeast has also been used during initial feeding, particularly with species typically not cultured intensively, such as grass carp Ctenopharyngo-don idella (Appelbaum and Uland 1979); vendace Coregonus albula (Appelbaum 1979), European sea bass Dicentrarchus labrax (Tovar et al. 2002; Tovar-Ramírez et al. 2004), and sharptooth catfish Clarias gariepnus (Hecht 1981). There have been a few studies to indicate that yeast inclusion in cultured-fish diets may not be beneficial however (Rumsey et al. 1991; Whyte and Sherry 2001; Li et al. 2005; Lim et al. 2005). Aside from Barnes et al. (2006), little experimentation has focused directly on yeast-supplementation effects during initial feeding on salmonid growth and survival. Research involving yeasts during salmonid culture has focused on juveniles (Johnson et al. 1980; Rumsey et al. 1991; Nakano et al. 1999) or fish greater than 100 g (Siwicki et al. 1994; Nakano et al. 1995; Cheng et al. 2004; Storebakken et al. 2004). To the best of our knowledge, no one has evaluated the effects of yeast-supplementation during the initial feeding of domesticated rain-bow trout Oncorhynchus mykiss, feral brown trout Salmo trutta, or fall Chinook salmon O. tshawytscha swim-up fry. Thus, the objective of this study was to de-termine the effect of a dried, fully permented yeast culture supplement (DVAqua from Diamond V Mills, Cedar Rapids, Iowa) on the growth and survival of these salmonid species and strains.

METHODS

The trials were conducted at McNenny State Fish Hatchery, Spearfish, South Dakota, USA. Well water at a constant temperature of 11ºC (total hardness as CaCO3, 360 mg L-1; alkalinity as CaCO3, 210 mg L-1; pH, 7.6; total dissolved solids, 390 mg L-1) was used throughout rearing. Egg incubation and sac-fry holding procedures were the same for each species used in this study, with newly-hatched fry kept in vertical-flow incubators (Flex-a-lite Consolidated, Tacoma, Washington) prior to initial feeding.

Rainbow Trout

Immediately after yolk sac absorption on November 2, 2004, approximately 2,000 (236 g) domesticated, Erwin-strain rainbow trout swim-up fry from the same incubator were placed into each of six 100-L cylindrical tanks (12,000 fish


total). Flows in each tank were set at 20 L/min. Just before moving the fish from the incubator to the tanks, 30 fish were weighed to the nearest mg and measured to the nearest mm and total tank weights were recorded. Feeding com-menced the day after placement of the fish in the tanks and daily feed rations were dispensed hourly from 08:00 to 16:00 using automatic feeders (Sweeney Enterprises, Inc., Boerne, Texas). Because of the constant water temperature, feeding levels for the tanks were determined by the hatchery constant (HC) method, with a planned feed conver-sion of 1.1 (Buterbaugh and Willoughby 1967). Because the fish were just learn-ing to eat, and to prevent the buildup of wasted feed in the tanks, for the first three days of feeding, HC was 3.96. It was increased to 5.94 for the next seven days, and maintained at 6.6 for the duration of the experiment. After reaching the HC of 6.6, feed levels were changed every week. Feed amounts were weighed to the nearest g. The six tanks were assigned to one of three diets. Two tanks received a com-mercially-prepared Silvercup® starter trout feed (Nelson and Sons, Inc., Murray, UT) with 0.25% grain blank mixed just prior to feeding. Two tanks received the same trout feed mixed with 0.125% 0.5 mm DVAqua and the last two tanks received starter mixed with 0.25% 0.5 mm DVAqua yeast product. To produce DVAqua, a molasses-based medium containing Saccahromyces cerevisiae was fer-mented anaerobically, mixed with a grain carrier, dried, and ground to the final particle size. The grain blank used in the control diet consisted of the same cereal grain mixture used in the production of DVAqua. At the end of four weeks, 40 fish per tank were again weighed and measured. Total tank weights to the near-est g were also recorded. Tanks were cleaned daily with mortalities removed and counted after cleaning.

Brown Trout

This study was conducted similarly to the aforementioned rainbow trout ex-periment, with feral brown trout as the test organisms. These fish were produced from spawn obtained from free-swimming brown trout in Soda Lake, WY. Im-mediately after yolk sac absorption on December 7, 2004, approximately 1,000 (113 g) brown trout swim-up fry from the same incubator were placed into each of the six 100-L cylindrical tanks (6,000 fish total). Experimental methodology was similar to the rainbow trout experiment with a few exceptions. Feeding rates were lower than those used with rainbow trout because brown trout historically have exhibited slower growth during hatchery rearing. For the first seven days of feeding, HC was 2.97. It was increased to 3.96 for the next seven days, and maintained at 4.95 for the remaining 8 days of the experiment. This trial could not last more than three weeks because the experimental tanks were needed for Chinook salmon testing (the subsequent experiment). The six tanks were again assigned to one of three diets. Two tanks received a commercially-prepared starter diet traditionally used by the South Dakota De-partment of Game, Fish and Parks to start feral brown trout on feed (Biodiet # 1, Bio-Oregon, Inc., Warrenton, OR) with 0.25% grain blank mixed just prior to feeding. Two tanks received the Biodiet mixed with 0.125% DVAqua and the


last two tanks received Biodiet mixed with 0.25% DVAqua. At the end of three weeks, 40 fish per tank were again weighed and measured and data collected as described previously in the rainbow trout experiment.

Chinook Salmon

This experiment used progeny of feral Fall Chinook salmon from Lake Oahe, SD, and was conducted similar to the other initial feeding experiments mentioned previously. Immediately after yolk sac absorption, approximately 500 (142 g) Chinook salmon swim-up fry from the same incubator were placed into each of the six tanks (3,000 fish total) on December 29, 2004. Methodology differed from the other two experiments in the following ways, primarily because of differences in fry size and projected growth rates. For the first six days of feed-ing, HC was 3.96. It was increased to 5.28 for the next seven days, 6.6 for the next seven days, and 7.92 for the remaining seven days of the trial. Biodiet #2 (Bio-Oregon, Inc., Warrenton, OR) was used as the base starter feed for mixing the grain blank and two DVAqua concentrations to make the three diets during the first week of the experiment. To replicate normal hatchery operations, for the remainder of the experiment, the base diet was a mix containing 50% #2 Biodiet and 50% #1 salmon crumbles (Silvercup salmon feed, Nelson and Sons, Inc., Murray, UT). In addition, after the first week of rearing, the particle size of grain blank and DVAqua mixed into the experimental diets increased from 0.5 to 1.0 mm.

Statistical Analysis

Data were analyzed by Analysis of Variance (ANOVA). Because of the small sample sizes used in these experiments, pairwise mean comparisons were performed using Fisher’s Protected Least Significance Difference, with signifi-cance predetermined at P < 0.05 (Ott 1984). As per Fisher’s Protected Least Significance Difference, pairwise comparisons were not conducted unless the initial ANOVA indicated significant differences between the treatments (Ott 1984). All mortality percentage data were arcsine transformed prior to analysis to stabilize the variances (Ott 1984).

RESULTS

Rainbow Trout

Mean ending tank weights ranged from 780 g in the control tanks to 855 g in the tanks receiving 0.25% DVAqua (Table 1). However, substantial variation in the two control tanks (SE = 106) precluded any statistically-significant dif-ferences among the treatments. Mean feed conversion was also not significantly different among the treatments and ranged from 1.23 in the control tanks, to 1.17 in the 0.125% DVAqua tanks, to 1.09 in the 0.25% DVAqua tanks. The


control tanks experienced a mean mortality percentage of 4.9%, compared to 4.3% and 3.4% in the 0.125% and 0.25% DVAqua diets respectively. No significant differences in length, weight, or condition factor were de-tected among the fish in any of the treatments (Table 2). Although not statisti-cally different however, a pattern of increasing fish length with increased dietary DVAqua was observed.

Brown Trout

Despite a nearly 10% improvement in mean total tank ending weights, gain, and feed conversion with the addition of DVAqua into initial brown trout diets, the means were not significantly different (Table 3). Percent mortality was similar among the treatments.

Table 1. Data (mean ± SE) for weight gain, feed conversion, and mortality rates for tanks of Erwin strain rainbow trout fed diets containing no yeast (control), 0.125% yeast product, or 0.25% yeast product for 28 rearing days from initial feeding (swim-up).

DIAMOND V YEAST SUPPLEMENT

NONE 0.125% 0.25% P

Tanks 2 2 2Start weight (g) 173 173 173End weight (g) 780 ± 106 808 ± 25 855 ± 5 0.792Gain (g) 623 ± 106 635 ± 25 682 ± 5 0.792Food fed (g) 742 742 742Conversion 1.23 ± 0.21 1.17 ± 0.05 1.09 ± 0.01 0.746% mortality 4.9 ± 0.7 4.3 ± 0.7 3.4 ± 0.2 0.316

Table 2. Data (mean ± SE) weights, lengths, and condition factors (K)a of Erwin strain rainbow trout fed diets containing no yeast (control), 0.125% yeast product, or 0.25% yeast product for 28 rearing days from initial feeding (swim-up).


NONE 0.125% 0.25% P

Start N 30 30 30End N 40 40 40Start weight (g) 0.09 ± 0.01 0.09 ± 0.01 0.09 ± 0.01End weight (g) 0.52 ± 0.02 0.54 ± 0.02 0.53 ± 0.02 0.741Start length (mm) 23.3 ± 0.2 23.3 ± 0.2 23.3 ± 0.2 End length (mm) 38.9 ± 0.5 39.2 ± 0.4 39.5 ± 0.4 0.608Start K 0.68 ± 0.01 0.68 ± 0.01 0.68 ± 0.01 End K 0.87 ± 0.01 0.88 ± 0.01 0.85 ± 0.01 0.148

a Condition factor (K) = 105 x (weight)/(length3)


Brown trout length significantly increased with increased dietary DVAqua, and individual fish weight increases approached statistical significance (P=0.074) (Table 4). Mean condition factor was nearly identical among the treatments.

Chinook Salmon

Chinook salmon rearing performance characteristics followed the same pat-tern observed in the domesticated rainbow trout and feral brown trout experi-ments. Tank ending weights and gain increased with increasing DVAqua con-centrations, although these increases were not significantly different (Table 5). Mean conversion was 2.44 in the control tanks, compared to 2.05 in the tanks

Table 3. Data (mean ± SE) for weight gain, feed conversion, and mortality rates for tanks of Soda Lake strain brown trout fed diets containing no yeast (control), 0.125% yeast product, or 0.25% yeast product for 21 rearing days from initial feeding (swim-up).


NONE 0.125% 0.25% P

Tanks 2 2 2Start weight (g) 113 113 113End weight (g) 219 ± 3 236 ± 8 236 ± 5 0.196Gain (g) 106 ± 3 123 ± 8 123 ± 5 0.196Food fed (g) 116 116 116Conversion 1.10 ± 0.03 0.95 ± 0.06 0.94 ± 0.04 0.160% mortality 1.6 ± 0.1 1.3 ± 0.3 1.5 ± 0.2 0.644

Table 4. Data (mean ± SE) weights, lengths, and condition factors (K)a of Soda Lake strain brown trout fed diets containing no yeast (control), 0.125% yeast product, or 0.25% yeast product for 21 rearing days from initial feeding (swim-up). Means with different letters are significantly dif-ferent (P < 0.05).


NONE 0.125% 0.25% P

Start N 30 30 30End N 40 40 40Start weight (g) 0.11± 0.01 0.11 ± 0.01 0.11 ± 0.01End weight (g) 0.24 ± 0.01 0.25 ± 0.01 0.26 ± 0.01 0.074Start length (mm) 25.0± 0.2 25.0 ± 0.2 25.0 ± 0.2 End length (mm) 31.6 ± 0.3 z 32.1 ± 0.3 zy 32.5 ± 0.2 y 0.028Start K 0.72 ± 0.01 0.72 ± 0.01 0.72 ± 0.01 End K 0.75 ± 0.01 0.74 ± 0.01 0.75 ± 0.01 0.984



receiving 0.125% DVAqua and 1.99 in the 0.25% DVAqua fed tanks. Decreases in mortality with the addition of DVAqua to the feed approached statistical sig-nificance (P=0.109). Mean individual fish weight, length, and condition factor all increased, albeit not significantly, with increased DVAqua concentrations in the starter feed (Table 6).

DISCUSSION

Small sample sizes and the corresponding low statistical power (Curtis et al. 1991) obviously hindered the ability to make conclusive, statistically-significant,

Table 5. Data (mean ± SE) for weight gain, feed conversion, and mortality rates for tanks of Oahe strain Chinook salmon fed diets containing no yeast (control), 0.125% yeast product, or 0.25% yeast product for 27 rearing days from initial feeding (swim-up).


NONE 0.125% 0.25% P

Tanks 2 2 2Start weight (g) 142 142 142End weight (g) 307 ± 15 337 ± 5 343 ± 5 0.140Gain (g) 165 ± 15 195 ± 5 201 ± 5 0.140Food fed (g) 400 400 400Conversion 2.44± 0.22 2.05 ± 0.07 1.99 ± 0.05 0.174% mortality 5.0 ± 0.8 3.0 ± 0.2 3.3 ± 0.1 0.109

Table 6. Data (mean ± SE) weights, lengths, and condition factors (K)a of Oahe strain Chinook salmon fed diets containing no yeast (control), 0.125% yeast product, or 0.25% yeast product for 27 rearing days from initial feeding (swim-up).


NONE 0.125% 0.25% P

Start N 20 20 20End N 40 40 40Start weight (g) 0.28± 0.01 0.28 ± 0.01 0.28 ± 0.01End weight (g) 0.68 ± 0.02 0.69 ± 0.02 0.73 ± 0.02 0.113Start length (mm) 34.5± 0.2 34.5 ± 0.2 34.5 ± 0.2 End length (mm) 45.7 ± 0.4 45.7 ± 0.4 46.3 ± 0.4 0.413Start K 0.70 ± 0.01 0.70 ± 0.01 0.70 ± 0.01 End K 0.71 ± 0.01 0.71 ± 0.01 0.74 ± 0.02 0.165



statements concerning the effects of DVAqua supplementation during these ex-periments. However the pattern of increased Saccahromyces cerevisiae-containing DVAqua yeast in the diet corresponding to improved weight gain, decreased feed conversion, and increased growth was consistent among all three of the experi-ments. Barnes et al. (2006) also observed this pattern during the initial feeding of feral rainbow trout receiving DVAqua supplementation, and further replica-tion led to statistical significance. We believe it likely that further replication of our experiments would produce significant differences between the diets tested. In the two other studies which have looked at yeast-supplementation during initial-feeding, the results have been inconsistent. Tovar-Ramírez et al. (2004) observed similar growth patterns in European sea bass fed diets containing a different yeast species, Debaryomyces hansenii. However, supplementation with S. cerevisiae did not improve the growth of initial-feeding European sea bass (Tovar et al. 2002). In other first-feeding studies where yeast was the sole dietary component (i.e. not just a supplement), grass carp fed only alkan yeast (Candida lypolitica) were significantly larger after 10 d than larvae fed either Artemia nau-plii or a commercial fish food (Appelbaum and Uland 1979). An initial diet of only torula yeast (Candida utilis) produced the best growth in sharptooth catfish larvae at 10 d post-initial-feeding compared to several other diets, including S. cerevisiae (Hecht 1981). In addition, S. cerevisiae-fed larvae exhibited greater mortality. Alkan yeast was also an acceptable first food for vendace (Appelbaum 1979). S. cerevisiae is a protein source by conventional definition (Cheng et al. 2004) and in non-salmonids, fish growth has improved with dietary supplementation of S. cerevisiae and other yeast species (Noh et al. 1994; Oliva-Teles and Gon-calves 2001; Lara-Flores et al. 2002; Li and Gatlin 2003, 2004). In addition to direct nutritional benefits, yeast may also be improving fish nutrition indirectly by adhering to fish intestinal mucus and producing polyamines (Vázquez-Juárez et al. 1993; Tovar et al. 2002; Tovar-Ramírez et al. 2004). The improvement in survival with the inclusion of dietary DVAqua, particu-larly as observed in the Chinook salmon experiment, is similar to that reported by Barnes et al. (2006). They reported an approximate 50% decrease in mortal-ity during the first four weeks when DVAqua was added to the starter diet of feral rainbow trout. The non-statistically-significant 40% decrease we observed dur-ing Chinook salmon initial feeding appears to follow that pattern. The decrease in mortality we observed with the domesticated rainbow trout is much smaller than that reported by Barnes et al. (2006) with a feral rainbow trout strain. DVAqua may improve survival by acting as a feed stimulant or attractant. Most of the mortality observed in our experiments consisted of starved fish which appeared to have not eaten anything since yolk sac absorption. Yeasts, such as the S. cerevisiae in DVAqua contain high levels of both essential and non-essential amino acids (Shcherbina et al. 1987; Appelbaum 1979; Cheng et al. 2004), which have been identified as the chemicals frequently responsible for increasing fish food palatability (Adron and Mackie 1978; Johnsen and Adams 1986; Harada 1989; Heinsbroek and Kreuger 1992; Papatryphon and Soares 2000). DVAqua, as a fully fermented yeast culture, contains a range of unidenti-fied metabolites that could serve a similar function. If DVAqua is indeed palat-


ability-enhancing, its relatively low cost would be atypical compared to other feed stimulant ingredients (Barrows and Hardy 2001). It is also possible that the decreased mortality associated with dietary DVAqua was due to yeast immunomodulatory effects. Yeast in general has been shown to improve immunological function in fish (Siwicki et al. 1994; Nakano et al. 1995, 1999; Ortuño et al. 2002; Li and Gatlin 2004, 2005) and DVAqua in particular improved the disease resistance of Pacific white shrimp (Burgents et al. 2004). On the basis of these studies, we would recommend the use of DVAqua at either 0.125% or 0.25% in starter diets during the initial feeding of salmonid species. Such supplementation likely produces benefits well in excess of the cost of the product, leading to a very favorable cost-to-benefit ratio.

ACKNOWLEDGEMENTS

We thank Rachel Sanders, Will Sayler, Rick Cordes, and Eric Krebs for their culture assistance, the many Black Hills State University students who participated in this study, and the reference librarians at the South Dakota State Library. This study was supported by Diamond V Mills, Cedar Rapids, IA.

REFERENCES

Adron, J. W., and A. M. Mackie. 1978. Studies on the chemical nature of feed-ing stimulants for rainbow trout, Salmo gairdneri Richardson. Journal of Fish Biology 12:303-320.

Appelbaum, S., and B. Uland. 1979. Intensive rearing of grass carp larvae Ctenopharyngodon idella (Valenciennes 1844) under controlled conditions. Aquaculture 17:175-179.

Barnes, M. E., D. J. Durben, S. G. Reeves, and R. Sanders. 2006. Dietary yeast culture supplementation improves initial rearing of McConaughy strain rainbow trout. Aquaculture Nutrition: 12:388-394.

Barrows, F. T., and R. W. Hardy. 2001. Nutrition and feeding. Pages 483-558 in G. A. Wedemeyer, editor. Fish hatchery management, second edition. American Fisheries Society, Bethesda, Maryland.

Burgents, J. E., K. G. Burnett, and L. E. Burnett. 2004. Disease resistance of Pacific white shrimp, Litopenaeus vannamaei, following the dietary adminis-tration of a yeast culture food supplement. Aquaculture 231:1-8.

Buterbaugh, G. L., and H. Willoughby. 1967. A feeding guide for brook, brown and rainbow trout. Progressive Fish-Culturist 29:210-215.

Cheng, Z. J., R. W. Hardy, and J. J. Huige. 2004. Apparent digestibility coeffi-cients of nutrients in brewer’s and rendered animal by-products for rainbow trout (Oncorhynchus mykiss (Walbum)). Aquaculture Research 35:1-9.

Curtis, C. R., M. D. Salman, and S. Shott. 1991. Power and sample size. Jour-nal of the American Veterinary Medical Association 197:838-840.


Dabrowksi, K., S. Hassard, J. Quinn, T. J. Pitcher, and A. M. Flinn. 1980. Ef-fect of Geotrichum candidum protein substitution in pelleted fish feed on the growth of rainbow trout (Salmo gairdneri Rich.) and on utilization of the diet. Aquaculture 21:213-232.

Ghosh, K., S. K.Sen, and A. K. Ray. 2005. Feed utilization efficiency and growth performance in rohu, Labeo rohita (Hamilton 1822), fingerlings fed yeast extract powder supplemented diets. Acta Ichthyologica Et Piscatoria 35:111-117.

Harada, K. 1989. Feeding attraction activities of L-dipeptides for abalone, oriental weatherfish, and yellowtail. Bulletin of the Japanese Society of Sci-entific Fisheries 55:1629-1634.

Hecht, T., 1981. Rearing of sharptooth catfish larvae (Clarias gariepnus Buchell, 1822: Clarida) under controlled conditions. Aquaculture 24, 301-308.

Heinsbroek, L. T. N., and J. G. Kreuger. 1992. Feeding and growth of glass eels, Anguilla anguilla L.: the effect of feeding stimulants on feed intake, energy metabolism and growth. Aquaculture and Fisheries Management 23:327-336.

Johnsen, P. B., and M. A. Adams. 1986. Chemical feeding stimulants for the herbivorous fish, Tilapia zillii. Comparative Biochemistry and Physiology 83A:109-112.

Johnson, E. A., D. E. Conklin, and M. J. Lewis. 1977. The yeast Phaffia rho-dozyma as a dietary pigment source for salmonids and crustaceans. Journal of the Fisheries Research Board of Canada 34:2417-2421.

Johnson, E. A., T. G. Villa, and M. J. Lewis. 1980. Phaffia rhodozyma as an astaxanthin source in salmonid diets. Aquaculture 20:123-134.

Lara-Flores, M., M. A.Olvera-Novoa, B. E.Guzmán-Médez, and W. López-Madrid. 2002. Use of the bacteria Streptococcus faecium and Lactobacillus acidophilus, and the yeast Saccharomyces cerevisiae as growth promoters in Nile tilapia (Oreochromis niloticus). Aquaculture 216:193-201.

Li, P., and D. M. Gatlin III. 2003. Evaluation of brewers yeast (Saccharamyces cerevisiae) as a feed supplement for hybrid striped bass (Morone chrysops x M. saxatilis). Aquaculture 219:681-692.

Li, P., and D. M. Gatlin III. 2004. Dietary brewer’s yeast and the prebiotic gro-biotic AE influence growth performance, immune responses and resistance of hybrid striped bass (Morone chrysops x M. saxatilis) to Streptococcus iniae infection. Aquaculture 231:445-456.

Li, P., and D. M. Gatlin III. 2005. Evaluation of the prebiotic GroBiotic® -A and brewers yeast as dietary supplements for sub-adult hybrid striped bass (Morone chrysops x M. saxatilis) challenged in situ with Mycobacerium mari-num. Aquaculture 248:197-205.

Li, P., G. S. Burr, J. Goff, K. W. Whiteman, K. B. Davis, R. R. Vega, W. H. Neill, and D. M. Gatlin III. 2005. A preliminary study on the effects of dietary supplementation of brewers yeast and nucleotides, singularly or in combination, on juvenile red drum (Sciaenops ocellatus). Aquaculture Re-search 36:1120-1127.

Lim, E. H., T. J. Lam, and J. L. Ding. 2005. Single-cell protein diet of a novel recombinant vitellogenin yeast enhances growth and survival of


first-feeding tilapia (Oreochromis mossambicus) larvae. Journal of Nutrition 135:513-518.

Nakano, T., M. Tosa, and M. Takeuchi. 1995. Improvement of biochemical features in fish health by red yeast and synthetic astaxanthin. Journal of Agriculture and Food Chemistry 43:1570-1573.

Nakano, T., Y. Miura, M. Wazawa, M. Sato, and M. Takeuchi. 1999. Red yeast Phaffia rhodozyma reduces susceptibility of liver homogenate to lipid peroxi-dation in rainbow trout. Fisheries Science 65:961-962.

Noh, S. H., I. K. Han, T. H. Won, Y. J. Choi. 1994. Effect of antibiotics, en-zyme, yeast culture and probiotics on the growth performance of Israeli carp. Korean Animal Science 36:480-486.

Oliva-Teles, A., and P. Goncalves. 2001. Partial replacement of fishmeal by brewers yeast Saccharomyces cerevisae in diets for sea bass Dicentrarchus labrax juveniles. Aquaculture 202:269-278.

Ortuño, J., A. Cuesta, A. Rodríguez, M. A.Eseban, J. Mseguer. 2002. Oral administration of yeast, Saccharomyces cerevisiae, enhances the cellular innate immune response of gilthead seabream (Sparus aurata L.). Veterinary Im-munology and Immunopathology 85:41-50.

Ott, L. (1984) An introduction to statistical methods and data analysis. PWS Publishers, Boston, MA.

Papatryphon, E., and J. H. Soares, Jr. 2000. Identification of feeding stimulants for striped bass, Morone saxatilis. Aquaculture 185:339-352.

Rumsey, G. L., S. G. Hughes, and J. L. Kinsella. 1990. Use of dietary yeast Sac-charomyces cerevisiae nitrogen by lake trout. Aquaculture 21:205-209.

Rumsey, G. L., J. E. Kinsella, K. J. Shetty, and S. G. Hughes. 1991. Effect of high dietary concentration of brewer’s dried yeast on growth performance and liver uricase in rainbow trout (Oncorhynchus mykiss). Animal Feed Sci-ence and Technology 33:177-183.

Sanderson, G. W., and S. O. Jolly. 1994. The value of Phaffia yeast as a feed ingredient for salmonid fish. Aquaculture 124:193-200.

Shcherbina, M. A., L. N. Trofimova, I. A. Sal’kova, and A. V. Grin’. 1987. Availability of amino acids in yeast raised on hydrocarbons for carp, Cyprinus carpio. Journal of Ichthyology 27:23-28.

Siwicki, A. K., D. P. Anderson, and G. L. Rumsey. 1994. Dietary intake of im-munostimulants by rainbow trout affects non-specific immunity and protec-tion against furunculosis. Veterinary Immunology and Immunopathology 41:125-139.

Storebakken, T., M. Sørensen, B. Bjerekeng, and S. Hiu. 2004. Utilization of astaxanthin from red yeast, Xanthophyllomyces dendrorhous, in rainbow trout, Oncorhychus mykiss: effects of enzymatic cell wall disruption and feed extru-sion temperature. Aquaculture 236:391-403.

Tovar, D., J. Zambonino, C. Cahu, F. J. Gatesoupe, R Vázquez-Juárez, and R. Lésel. 2002. Effect of live yeast incorporation in compound diet on diges-tive enzyme activity in sea bass (Dicentrarchus labrax) larvae. Aquaculture 204:113-123.

Tovar-Ramírez, D., J. Zambonino Infante, C. Cahu, F. J. Gatesoupe, and R. Vázquez-Juárez. 2004. Influence of dietary live yeast on European sea bass


(Dicentrarchus labrax) larval development. Aquaculture 234:415-427.Vázquez-Juárez, R., F. Ascencio, T. Andlid, L. Gustafsson, and T. Wadström.

1993. The expression of potential colonization factors of yeasts isolated from fish during different growth conditions. Canadian Journal of Micro-biology 39:1135-1141.

Whyte, J. N. C., and K. L. Sherry. 2001. Pigmentation and composition of flesh of Atlantic salmon fed diets supplemented with yeast Phaffia rhodozy-ma. North American Journal of Aquaculture 63:52-57.


COMPUTER AIDED STATISTICAL ANALYSIS OF SATELLITE SENSOR DATA CROSS-CALIBRATION

Stuart Ness and Daniel SwetsAugustana College

Sioux Falls, SD

ABSTRACT

The space borne Moderate Resolution Imaging Spectroradiometer (MO-DIS) instrument was designed as the leading edge of global observation technol-ogy. The MODIS instrument and later follow-on satellite sensors (e.g., NPP, VIIRS) represent the best technology for further observations of global change, land use/land cover change, and for global mapping. However, the value of MODIS data for these applications has been limited because of its short history. The satellite sensor that serves as the precursor to MODIS is the Advanced Very High Resolution Radiometer (AVHRR). The USGS National Center for Earth Resource Observation and Science (EROS) has an archive of AVHRR data cov-ering the conterminous United States dating from 1989 to the present. Using the year 2003 where we have data from both sensors, we investigated a method to transform the heritage AVHRR data to a form useful for comparison with the MODIS data. This will allow for comparative studies, such as climate and environmental change, which require the long history of AVHRR data with the current and future data supplied by MODIS. We have found that a simple linear regression does not appear to provide an accurate transform between the sensors. In addition, it has been observed that divisions by land type, position in the season, and geographical area need to be addressed for accurate comparisons. Within the data, we have seen some obvious problems with heteroskedasticity, and suspect that the data could be cross-sec-tional in nature to account for the reported variance.

INTRODUCTION

The use of remote sensing has continued to increase over the past decade as increasingly advanced technology becomes available to give more accurate measurements and faster evaluation times of current seasonal data. One of the most widely used sets of data from remote sensing sensors has been the Normal-ized Difference Vegetation Index (NDVI). NDVI is routinely calculated using the visible and near infrared light spectrum acquired by satellite sensor data. Until recently, the data has been collected using the NOAA Advanced Very High Resolution Radiometer (AVHRR) series satellites. As these satellites have begun to age, a new set of sensors have been created to give better measurements and higher resolutions. The sensor that will be used to replace AVHRR is the Visible/Infrared Imager/Radiometer Suite (VIIRS). While this sensor has not


been put into production yet, the VIIRS sensor will be very similar to the NASA Moderate Resolution Imaging Spectroradiometer (MODIS) sensor. Many studies have used the long history of NDVI data that has been col-lected from the AVHRR sensor [11], [10], [9], [17], and [19]. A number of these studies have used the almost twenty years worth of sensor data to monitor changes in both vegetation and a variety of land surface properties. Being able to understand how current AVHRR derived NDVI and current and future sensor derived NDVI data relate is crucial to being able to allow future long term trend analysis studies to continue.

PROJECT BACKGROUND

Background

A small number of studies have previously compared observed and simulated MODIS and AVHRR data ([1], [4], [5], [8], [13], [18]), which have provided varied results, mainly in support of high correlation values between cross sensor evaluations. Gitelson, et. al, shows that MODIS NDVI values are slightly greater than those from AVHRR in simulated data from the red and near infrared (NIR) bands [3]. In addition, Gallo et. al ([4], [5]) both found that MODIS NDVI data has shown good agreement with other sensor NDVI data when properly corrected for water vapor, ozone, Rayleigh scattering, and other atmospheric conditions in a manner discussed by [12] and [16]. There is evidence that there is a poor correlation between the NOAA-14 AVHRR satellite and the MODIS Terra sensor [2]. However, while the NOAA-14 and NOAA-15 satellite showed poor correlation, the NOAA-16 satellite has a slightly adjusted range of bands that are used to determine NDVI, which should allow for a better correlation [18]. Many of the studies have used simulated data rather than actual data in order to avoid the atmospheric and other sensor based problems that exist within satel-lite image data. The majority of these studies have found that there is a strong linear relationship between the two sensors. In particular, Steven et. al has noted that the reflectance for spectral band effects can be corrected to approximately a value of plus or minus 0.02 [13]. To deal with the effects of satellite images that simulated data does not factor in, it appears that topography, solar angles, and viewing angles have a contribution to the differences between satellite data [13]. One problem with the data is the problem of geo-registration (not being able to accurately map the data to the same locations) [18]. Further, AVHRR and MODIS NDVI values can vary with land cover type, simply by observation of the data sets, and noted that correctional algorithms should take the land cover differences into account [4].

Initial Hypothesis

The initial regressions were created under the hypothesis that the differences between the AVHRR and MODIS sensors were a result of the general differences


between sensors, including spectral band differences and time of day, as well as the contamination based on clouds. It was also believed that there may be a gen-eral link between sensor differences among different types of land cover classes. Under some brief exploration of data sets, our hypothesis shifted to include some type of correction for approximate time of season.

Comparison Data Setup

Our study compares four satellite sensors. Two satellites, N16 and N17, use the AVHRR sensor; the other two satellites, AQUA and TERRA, use the MO-DIS sensor. We use a piece-wise linear regression technique to compare the data from these sensors. We classify the data by land cover type and the 16-day composite obser-vation number. The land cover type was determined from the adjusted 1992 NLCD land cover map [19]. The land cover classes we use are: deciduous forest, evergreen forest, mixed forest, grassland and herbaceous, shrub land, row crops, small grains, pasture and hay, residential and commercial, water and ice, and the generic other. This study extends Gallo’s work, which examined the first nine classes, as water and ice have many notable problems with NDVI values being artificially low [5]. In addition, the land cover classified as the “other” category defines too broad of a region to be useful in making an accurate evaluation. The compilations of the data sets were done in two ways. The first was to use a pixel to pixel comparison using images of the US. In total the comparison used 13,251,843 points per composite, which was divided into the eleven land type groupings. While this comparison is ideal in the comparison of data, a number of issues, such as accuracy problems with the land type map, geo-registration problems, as well as artificially low data point values due to image contamina-tion, led to the use of the second data set aggregation. The second compilation of the data set used 20km by 20km sample sites located around the United States where the contents of the sample is 80% or greater of one land type. Each sample was averaged, using only the pixels that according to the modified NLCD land type map were of the dominant land type. The averaging also removed the effects of cloud- and water-masked pixels. One issue within the data set is that the N16 satellite began to fail midway through the 2003 data year (at image observation number 16), resulting in an incomplete year of data for comparison between the four different satellites. In the regression analysis, all depictions of N16 are used through the last good date of N16 to gain full value of the regression values. Unfortunately, due to the failure, there is only a limited amount of data between N16 and the Aqua satel-lites. The 2003 data year was chosen as it provides the greatest amount of data between N16 and Aqua allowing for the maximum number of data comparisons in one season.

Running Regressions

In order to use the statistical package SAS, each land cover type was divided into a separate file that was ordered by seasonal composite image number. These


groupings from each image were then sequentially linked together to allow for data evaluation. The particular regressions that were run were simple ordinary least squares first order linear regressions. Under our initial hypothesis, the differences be-tween the two data sets were dependent on the differences between the sensors, which should affect all data equally and should be easily corrected based on some simple linear regression. In addition, some work completed by Kevin Gallo at the National Center for EROS suggested that there may be explainable differ-ences between land cover types [5]. Following this, we divided the data into nine divisions, using the modified NLCD map and omitting the water and bare rock/sand/clay land covers. As we worked further into the project, it became ap-parent that the inability to accurately geo-register the satellite images pixels to the correct latitude and longitude position was creating problems with regressions analysis. To solve this, another program was used to extract sample areas with large coverage of the desired land cover class and average a twenty by twenty pixel area. We then ran regressions in SAS and the open source program Gretl accord-ing to the seasonal observation number on all of the data set combinations, as well as doing an inclusive land type regression dismissing the observation num-ber on all sample sites. The observation numbers discussed in the article refer to the time offset from the beginning of the year based on 23 composite observa-tions in one year. No corrections were done in the techniques to solve problems with heteroskedasticity and no correlation problems appeared in the data sets.

DATA RESULTS

The results from the regressions have shown to be very troublesome in some areas of the data. As a general rule, land types one and three (deciduous and mixed forests) were the most troublesome of the nine evaluated land cover classes. In addition, there appears to be a seasonal trend that is portrayed in the adjusted R2 as well as the coefficient and the intercept. It should also be noted that as the NDVI values increase and reach the peak of the growing season, usually in the middle of the season around observations 10-12, there is a greater chance for large NDVI deviations between satellites to occur. In evaluation of the R2 on all combinations, running a regression on all 16-day composite samples of the same land type resulted in R2 values between .7 and .95 in addition to matching the hypothesized sign and approximate value of the coefficient. Based on the initial hypothesis, a linear regression was chosen to be the first model structure. It was estimated that the intercept should occur between -0.2 and 0.2, with a positive slope. The data would be piece-wise as well, allowing for slope changes between 16-day composites. However, in the plotting of the data without separation for observation number, it became apparent that using either a squared regression or a higher polynomial function may offer more explanation between the combinations of sensors. The plots presented in the next section of the paper also offer some other insights into potential differential factors in the data as well.


Common Features

The following basic plot charts were created using pairs of sensors (with MODIS sensor on the vertical axis, and the AVHRR sensor on the horizontal access). The plots were done according to the dataset and the land cover type number. A general trend is evident in most of these plots. In examining all data points together, there is a definite general data comparison line; however, there is also a large amount of deviation from the regression line as shown in figure 1.

In simply separating the data into forest and non-forest, the data shows characteristics based on land cover. In general, there is a “floor” on all data derived from the non-forest land areas, and a “ceiling” on all forest land areas. The differences can be seen in figure 2 and figure 3. In addition to identifying which land cover in general has the “floor” and “ceiling” effects, the data also has a considerably smaller amount of variation, due to specifying the forest and non-forest samples. This ability to improve our results by splitting the data suggests that there could be further gain from analyzing the different land cover types separately. Examining the data plots, it also is interesting to see that even though the regression lines plotted over the data points are different, if we could remove the problem areas of the data, it may be seen that the other factors such as seasonality may be involved in determining the makeup of the general com-parison of the data calibration. To begin breaking apart these data sets, these simple relations were first broken into the nine land cover classes to be evaluated separately.

Figure 1. All land cover and observations. Notice the general trend line evident in the data. However, there is also a great deal of scatter. While a regression offers a relatively high R2, the amount of variation limits the ability of the regression to explain the sensor differences by com-bining all observations together.


Figure 2. All forest land cover, all observations, with the ceiling effect visible, but the floor effect missing, suggesting that the lower saturation level on the MODIS based sensor exists in the non-forest land cover. We also note that the non-scattered and non ceiling saturated data represents a very similar data plot to the non-forest land covers. This suggests that there may be other com-mon factors to these two data sets, such as seasonality which should be accounted for.

Figure 3. All Non-forest land covers for all observations. Notice the floor effect is present, but the ceiling effect is missing from this data. We notice especially that while a floor effect exists heavily on the MODIS sensor at the 0 NDVI level, it also exists on the AVHRR sensor around the 0.2 NDVI level. We also note the similar characteristics as noted on Figure 3-2.


Growing and Non-Growing Season

Deciduous Forest

In the Deciduous Forest land cover class, as shown in figure 4, a distinct discontinuity in the data values appeared. In the lower ranges of AVHRR (under 0.4 NDVI), there is a large scatter of the MODIS points from 0 NDVI to 0.7 NDVI. However, there also appears to be a strong relationship of the sensors between 0.35 NDVI to 0.6 NDVI in the AVHRR sensor and 0.37 NDVI to 0.63 NDVI in the MODIS sensor. However, at 0.6 NDVI in AVHRR and 0.7 NDVI in MODIS there appears to be a discontinuity. It also appears that there is a “ceiling effect” occurring at 0.9 NDVI on the MODIS sensor. Looking at figure 4, it becomes clear that there are two separate regions, a lower region with a typical slope and an upper region with a less steep slope.

This occurrence could suggest that there may be an additional dimension that needs to be accounted for, such as geographic region. However, under further analysis, we see that the data is broken into at least two of sections, as can be seen in figure 5 and figure 6. These figures show that simply breaking the image into parts of the growing season result in greater grouping of the data. However, in breaking the data into these two groups, the models have a relatively low explanation factor, with an R2 of approximately 0.78 for the off-growing season and 0.34 for the growing season. This is a large drop considering the full deciduous forest model R2 is approximately 0.88.

Figure 4. Deciduous Forest N17 vs. Terra – All Observations. Two separate regions appear in the data plot of the deciduous forest land cover. The lower portion has a great deal of scatter on the AVHRR sensor from 0.1 to 0.4 NDVI, while the MODIS sensor has a ceiling effect around a 0.9 NDVI. The R2 for the regression is 0.88, which may be artificially boosted due to the large number of samples used for the regression.


In figure 5 there is also ambiguity between the two sensors in the N16 0.1 to 0.3 NDVI ranges. In examining the individual plots from the data, these values are caused by Observations 2, 3, and 21 of the given year. This could be possibly explained with respect to snow and ice reflection, as well as cloud shad-ows or other sky clarity and atmospheric issues that have not been detected and accounted for in the cloud or snow mask.

Figure 5. Deciduous Forest Observations 1-9. These observations make up the non-growing season. It is easily noticeable that the regression follows a regression line that has a reasonable slope and intercept. Interestingly below 0.4 NDVI there is a great deal of variation between the two sensors. However, above 0.4 NDVI, there appears to be a well behaved data correlation.

Figure 6. Deciduous Forest Observation 10-16. The growing season for the deciduous forest ap-pears to be a much higher regression combination, possibly due to a sensor saturation with the MODIS NDVI calculation. There also appears to be a set of points that are considerable outliers that may be a result of mis-classified data points of low-NDVI growing seasons.


The explanation of this reduction could also be in part due to the number of observations used in each regression model. Since there are close to 2000 points in the full regression, the R2 is artificially boosted, giving an inflated value which will decrease when reducing the number of observations in a set.

Evergreen Forest

The evergreen forest did not appear to have a distinct break, as seen in the deciduous forest. However, it did become apparent that there is a greater varia-tion between the data points as well as possibly a quadratic relationship. It is interesting to note that the N17 sensor combined with either MODIS sensor had a smaller root mean squared error (RMSE) as well as having a higher R2. The most noticeable difference between the N17 and Terra comparison is the scattering of data points above the main concentration, as seen in figure 7.

In breaking this image apart, we see two distinct time regions. Observations 1-9 and 19-23, which has a larger standard error (0.015), lower R2 (0.657) and show the non-uniform parts of the graph (figure 8). Observations 10-18 however have a nice distribution and offer a high R2 (0.888) and an even lower standard error (0.009) (figure 9).

Figure 7. Notice the higher polynomial curve shape of the concentrated data points. In addition, a peculiar trend showing more variation on the high Terra, low N17 side than on the opposite side.


Figure 8. Evergreen Forest Observations 10-18 The growing season for this forest appears to be well behaved and have little residual data points with large variations. There are a few samples with low NDVI values on the AVHRR sensor, however, the split to the growing season offers a great deal of explaination between the two sensors.

Figure 9. Evergreen Forest Observations 1-9 and 19-23. The non-growing season in comparison to the growing season has a greater variation. It is obvious that there is something that occurs in the early and late part of the season which needs to be accounted for. This could be a result of residual clouds not detected by the CLAVR cloud mask, or other atmospheric contamination or snow and ice contamination.


Mixed Forest

The mixed forest is difficult due to a relative lack of data points. In the N17 vs. MODIS, there were approximately 550 data points, and 367 points in the N16 vs. MODIS comparison. The N16 vs. MODIS comparisons were fairly straightforward. The N17 vs. MODIS comparisons (Figure 10), however, resulted in a similar effect as the evergreen forest observations 10-18 (figure 9). The overall data plot shown in Figure 10 appears to be a cross between the char-acteristics of both the deciduous forest, with a discontinuous regression, and the increased scatter found within the evergreen forest. With the smaller amount of data, the problems could simply be a result of a lack of data observations to successfully determine the variations from the actual data points.

Grasslands / Herbaceous

The Grasslands land cover type offers another problem. In all of the samples, it appears that at low levels of NDVI, the MODIS sensor measures NDVI at approximately 0, but the AVHRR sensor shows a scattered range between 0 and 0.2. This could be due to snow and clouds (as the difference is occurring at very low levels of NDVI), typically found during winter months. The low levels can be seen by combining observations 1 through 6 and 19 through 23. As depicted in Figure 11, the floor effect occurs in the off-season. While there is still a considerable amount of good data points in this particular non-growing season, there is a significant problem with these off-growing season observations. The resulting growing season then lacks any floor effect and has minimized scatter (figure 12).

Figure 10. Mixed Forest all observations. This land cover has features which appear to be a cross between the deciduous and evergreen forest land cover classes. Due to the relative lack of data points, it is difficult to derive any solid deductions on this data set.


Evaluating the standard ordinary least squares method, it was observed that the floor effect, occurring within the non-growing portion of the season, has an R2 of .687 while the growing season has an R2 of .912. This offers a promising method for correcting off-season and on-season data with a great improvement for the growing season as the R2 of the entire season for the fourth land cover is .852.

Figure 11. Grasslands/Herbaceous Non-Growing Season Observations 1-6 and 19-23. The off-growing season of the grasslands contains a large floor on MODIS values that range from 0 to 0.4 NDVI (based on AVHRR), which may be a result of snow, clouds, or other atmospheric contamination. The upper regions of the data are well-behaved, and follow a regression line relatively well.

Figure 12. Grasslands/Herbaceous Growing Season Observations 7-18. The growing season of the grassland offers a promising comparison between the two sensors, as the data plot follows the regression line relatively closely, having an R2 of 0.92.


Shrub land

Similar to the other non-forest land cover classes, the data has relatively little data scatter. Without any time-of-year separation to the data, the regression has a relatively high R2 at 0.88 (figure 13). However, the shrub land does not clearly separate into a season and off-season. While the same floor is visible within the data to a smaller degree, the primary contributors are the first four observa-tions.

One possible way to better explain these data would be to employ a piece-wise regression technique. Basically, instead of only splitting the data by grow-ing- and non-growing season, this technique groups the data by composite obser-vation number. These smaller data sets are then evaluated in a linear regression technique with the hopes of avoiding any problems that exist due to changes in conditions as a result of different time different sun angle, and other factors that occur with different times of the year. These individual regressions can then be examined to see if any pattern exists of if there exist condition changes between growing and non-growing times of the year. In examining the piece-wise regressions, we find no clear observable pattern of the data. By examining the intercept of the regression line, we can observe trends for both within- and outside the growing season. As shown in table 1, the later observations within the growing season (periods 11-16) tend to have a similar intercept. Evaluation of these six periods in the later growing season (fig-ure 14) reveals a more uniform data set compared to the whole growing season (figure 15). An R2 of 0.93 shows that this smaller period correlates better..

Figure 13. Shrub land: Total Season Observations 1 – 16. The shrub land is difficult to divide into a season and non-season data set. There is relatively little variation between the two sensors, and the small floor that appears on the MODIS sensor is barely recognizable.


It is difficult to create an explanation for the change in behavior of this land cover. However, because of the small number of observations (just over 60 per time period), it is possible that there may not be completely accurate representa-tions due to water vapor or other factors that may occur at more frequently at certain times of the year, which are adjusted by the MODIS sensor but not by the AVHRR sensor.

Table 1. Shrub land: Slope-Intercept Table sorted by Intercept. This table is an example of the intercepts and slopes obtained using a piece-wise regression. While the R2 on all of these re-gressions was low, it was largely due to a lack of data points. By examining the intercepts and slopes, we can pick out observations that are approximately from the same time, which we would expect may have a common seasonal shape.

OBSERVATION INTERCEPT SLOPE

12 -0.51 1.0413 -0.0661 1.0816 -0.062 1.0714 -0.0429 1.0311 -0.0392 1.0715 -0.0371 0.9873 -0.0207 1.121 0.002 1.078 0.0038 1.017 0.00875 0.962 0.0105 0.9964 0.0116 1.029 0.0121 0.9426 0.0163 0.9365 0.0219 0.974

10 0.0243 0.908


Figure 14. Shrub land: Late Growing Season 11-16. Using a piece-wise regression technique, we found a set of observations that appear to have a relatively close fit together. Since we were using all late growing-season observations, it may be possible that the correlation between data points may also have a factor dealing with green-up or green-down parts of the growing-sea-son.

Figure 15. Shrub land: Growing Season 9– 16. Compariablly to the late growing season, this shows a “dual” set of data points. However, this may also be impacted by different seasonality dates, depending on where the data point is located.


The off season (figure 16) also should be noted as having a similar R2 (0.91) to the complete growing season. However, as shown in table 1, there is not a clear pattern showing the lead-up to the season. Instead what we see is a range of intercepts from 0.002 to 0.0243. This variation along with the variation of slopes suggests that there are some other factors involved other than time of season.

Row Crops

Row Crops typically attain higher levels of NDVI during the growing sea-son, but otherwise behave similarly to the other non-forest land cover types. While these values do not exhibit the same drastic ceiling behavior found in the forest land covers, it is possible that this effect could be observed during years of extremely high growth. Figure 17 illustrates two areas of concern. The first is the floor values that oc-cur along with the high amount of scatter in the lower NDVI range. The second concern is that the data appears to be related under a higher order polynomial. While these areas lower the regression explanation, the R2 maintains itself above 0.91.

Figure 16. Shrub land: Non-Growing 1-8. The non-growing season is held to a minimal variation, however, it does have a small floor on the MODIS NDVI values, as well as have a slight spread between the 0.1 NDVI AVHRR and 0.35 NDVI AVHRR. This suggests that there may be other fac-tors that lead into the explaination other than growing/non-growing season evaluation.


Breaking the season into the growing and non-growing sections, there is clearly a correlation again to the floor effect and the non-growing part of the season. In addition, most of the scatter occurs in the non-growing part of the season, as shown in figure 18. Due to the scatter of the comparison, as well as the floor effect of the MODIS sensor, the R2 measures a 0.68 for this off-season data.

Figure 17. Row Crops: Entire Season. The row crop land cover has a unique in that it achieves high NDVI levels, as well as low NDVI levels. While it is not extremely obvious, there does appear to be a slight saturation level, as well as a quadratic shape to the data. Also noticeable is the floor, as well as the low NDVI level scatter.

Figure 18. Row Crops: Non-Growing Season: Observations 1 - 8 and 20 – 23. By using the non-growing season, we are able to extract the scatter as well as the floor of the MODIS samples. There are a few higher NDVI values, which may be a result of misclassifications or early grow-ing seasons. Floor values have been shown to be removed by the addition of a snow mask (not shown), and improved the regression R2 above the 0.68 from the non-snow masked regression.


The growing season for the row crops also has an interesting feature. The data appears to have a non-linear correlation. The data would suggest that a higher order polynomial function may be needed to model the data points. In addition, the data also appears to have a slight ceiling effect, which could also be contributing to the need for a polynomial function. Figure 19 shows the data with a simple linear regression. While it is possible that a higher polynomial function should be used, the regression does have a rather high rate of explana-tion (R2 = 0.92). In addition, the possibility of using piece-wise regression, as discussed in a later section, eliminates the need for the higher order function.

Small Grains

The last three land cover types provide interesting attributes. Similar to row crops, a slight higher order polynomial function appears in the data comparison. Figure 20 shows a linear regression over the entire season for the small grains, producing an R2 of 0.93, which is one of the highest regression values we have been observed in this study. However, compared to the other land cover classes, we have relatively few observations, only 775 data points for the entire season.

Figure 19. Row Crops: Growing Season: Observations 9- 19. The row crop growing season appears to have a slight ceiling saturation from the MODIS data, as well as a quadratic shape to the data. This may be in part due to differences within seasonal characteristics; however, the linear regression has a relatively high R2 of 0.92.


Because of the relatively good linear fit to the data, it was difficult to divide the data into any clear separation based on observations. However, in order to eliminate the floor effect data, observations 6 through 19 were separated and compared with an R2 of 0.91 (figure 21) and the remaining observations (figure 22) had an R2 of 0.91. One possible explanation for the relatively similar data comparison through-out the entire season could be the length of the season for small grains. Because small grains have an earlier growing season and have varied lengths of seasons, depending on area, it may be difficult to accurately separate the growing/non-growing season.

Figure 20. Small Grains: Entire Season. The small grains regressions are limited severely by the number of data samples available. In addition, it is difficult to determine a seasonal difference clearly by observation of the data. This is in part due to the wide variety of crops that are planted, which have seasons that overlap, and start very early in the year. As a result, the seasonal characteristics depend largely on what type of small grain is being observed.


Figure 21. Small Grains: Growing Season: Observations 6 – 19. Creating a growing season portion of the data lowers the R2 value from 0.93 to 0.91, due to the reduction in the number of data points. There is a small amount of variation in the upper NDVI values; however, in general it forms a relatively good linear regression line.

Figure 22. Small Grains: Non-Growing Season: Observations 1-5 and 20-23. The non-growing season is much like the growing season, except for an additional floor that occurs around 0 NDVI. The 0.91 R2 is the same as the growing season, with a similar range of NDVI values. This in large part is due to the near continuous growing season offered by different types of small grains. As a result, the extremely low values may only last for a short time, suggesting that there is a much smaller non-growing season that what has been depicted.


Pasture Hay

Compared to the row crop and small grain land cover, pasture hay has a greatly reduced correlation between the sensors. Using the first 16 observations, due to N16 sensor failure, figure 23 shows a set of data with a wider spread along with a small set of floor pixels with a limited amount of pixels that fall below 0.2 NDVI. The R2 for all pasture hay pixels over the sixteen good observations is 0.81.

By dividing the data into the respective parts of the season, we see a drop in the R2 values to 0.71 for the non-growing season (figure 24) and 0.75 for the growing season (figure 25). While the regression has a lower R2 value, noting visually that there is not a great deal of difference between the three regressions suggests that the entire season R2 value was artificially inflated due to the number of data points used. Also, in figure 24, the NDVI values reach almost the same levels as the grow-ing season. This may be in part due to the geographic location of the data points. Examining the location of the sample sites, the pasture hay land cover class pri-marily runs along the Mississippi river, in a north-to-south fashion. This land cover type has a considerable amount of the sample locations derived from the southern part of the U.S., where growth levels do not typically fall to extremely low levels. This could result in high levels of NDVI during the non-growing season.

Figure 23. Pasture Hay: Entire Season: Observations 1 – 16. The pasture/hay has greatly reduced explaination by linear regression compared to the previous land cover classes. There is a much larger scatter range, as well a more extreme set of floor values that exist. In addition, there ap-pears to be some type of non-linear regression involved with the heaviest part of the data plot.


Figure 24. Pasture Hay: Non-Growing Season: Observations 1-8 . The non-growing season has a relatively high NDVI value. This may be in large part due to the southern geographic location of the data points. We also expect that because of the geographic location differences between the sample locations, the growing seasons may need to be divided in a much more careful manner specific to the geographic location.

Figure 25. Pasture Hay: Growing Season: Observations 8 – 16. The growing season has a much less scattered appearance than its non-growing counterpart. This is largely due to all locations having a high NDVI value. While the R2 of 0.75 suggests that there are other factors involved in the regression, in examining the difference between the non-growing season, we see that a closer look at starting and ending dates of the season are important.


Urban

The urban land, consisting of commercial, industrial and residential areas, has a limited number of data points available for comparison. Due in part to the lack of data points, it was difficult to find any explanation between the sets of data. A general explanation line between the sensor data appears to form within the data; however, there is a great deal of scatter that could possibly be explained by the differences in latitude and longitude of the data areas. The entire season regression (figure 26) only had a 0.67 R2; however, it is obvious that other factors are influencing the data. One possible factor could be the differences between geographic locations, as this was not considered in the study. This may be an important factor and a possible area for further study of how different locations will be affected by the differences of light reflection.

Especially in the urban areas, there is very limited ability of separating out the scatter among different observations. However, there does appear to be a correlation between a few lower floor values and the off-growing season. Figure 27 shows the observations during the growing season. Overall, this collection of data points had a R2 of 0.47, which is significantly lower than the entire season. However, as was stated earlier, while there is less explanation, the higher R2 value could simply be due to more observations. Figure 28, shows the non-growing season part of the year, with a R2 value of 0.54. Surprisingly, in this case, the off-growing season has a better regression

Figure 26. Urban: Entire Season. The Urban land cover suffers from three problems. The first problem is a lack of data sample sites, resulting in semi-visible regressions. The second problem is that the urban sample areas are scattered around the U.S. and have different lengths of growing seasons. The third complication is that the urban land cover is known to be problematic within a growing year due to the wide range variety of areas classified as urban.


than the growing season. This could potentially be due to more data points; however, it may also be a result of atmospheric conditions, such as smog or water vapor, which if one sensor does not correct for, could show greater reduc-tions, since smog and water vapor tend to be more prevalent during the summer months, causing reduced correlation.

Figure 27. Urban: Growing Season: Observations 9-19. Aside from eliminating the lower NDVI values, very little is gained from the separation into a typical growing/non-growing season set of data.

Figure 28. Urban: Non-Growing Season Observations 1 - 8 and 20 – 23. We notice a floor on both sensors; however, due to the lack of data points, it is difficult to show how great of an impact the floor values have. In addition, there is extensive scattering, suggesting that some other factors must be in play.


Specific Satellite Features

Between each of the different satellite sensors, there are a few unique fea-tures. In general, the N17 satellite tended to have a greater spread of the data points over the N16 satellite. This tended to make the N16 regressions a bit cleaner and gave better explanation between the sensors. However, due to the limited data from the N16 satellite, N17 data was primarily used for demonstra-tion purposes due to the complete 2003 data set. Another consideration when evaluating the data sets is to examine the cross calibration of the different satellites. While this was not a focus of this paper, it was interesting to note that while in general the sensors have limited variation, the variation between satellites increase in a similar pattern as the calibration be-tween the AVHRR and MODIS comparisons. This increase was not of a similar magnitude, yet, the comparisons did note a difference in variation in growing and non-growing parts of the season. For an example of the evaluation, a comparison of the Aqua and Terra satel-lites is considered for the Evergreen Forest Land Cover Type. In figure 32, the non-growing season is displayed, with a R2 of 0.79. However, this can be com-pared with figure 33, which shows a much tighter data set, that has a R2 value of 0.97. While this does not match up to the variation change of the AVHRR to MODIS regression, it does show that the analysis of these sensors is based on more than a sensor mis-calibration, and there are other factors that could be influencing the data.

Figure 29. Aqua vs. Terra: Observations 1-8 and 19-23. Despite these two satellites being of the same sensor type, there is a considerable amount of scatter between the data points. Differences between these two sensors may include cloud problems, atmospheric contaminations, as well as the sun angle due to the time of day that the over-pass of the satellite occurred. While these were not considered in the scope of this project, we note it to suggest that there is very little hope of achieving a perfect calibration between two sensors.


Piece-wise Regression

The consideration of piece-wise regression is employed to consider further explanation of the data sets, allowing smaller errors and data issues to be con-sidered. This area has not been completely explored, however, some preliminary work has been done to show the possibility that it would have a significant im-pact on the data set. For analysis of the 2003 data set, by considering the individual observation numbers, an assumption was made that the data’s seasonal values per land type are similar values. If dealing with multi-year data sets, a consideration into off-setting seasonal conditions should be employed. In addition, the second primary focus of moving to a piece-wise regression is to eliminate the need for higher order polynomial functions in the calculation of the data sets. This work is preliminary as only a single year of data has been used. If this area is to be considered further, a multiyear data set would be needed to give more data points, as well as deal with seasonal offsets.

DISCUSSION

Within each land cover type, there appears to be a connection between the growing part of the season and a higher R2 for linear regressions. There are most notably two exceptions to this generalization, the urban and the pasture hay land cover types. While it is difficult to conclude any solid observations as to

Figure 30. Aqua vs. Terra: Observations 9-18. As we noted, in the growing season with the other sensor regressions, the growing season has better explanation. It appears that this is similar to the same sensor type. This suggests that there is definatley something greater than sensor calibra-tion that is off, but includes other conditions that need to be accounted for. However, this does show that the part of the season has an important impact on the regression calibration.


what makes these two land covers different, one speculative answer could relate to the geographically diverse areas that the sample sites were taken from. Both land covers had a wide range of sites taken primarily in the north to south fash-ion. This would suggest that the relation could involve another factor, namely latitude. However, without more data, there is nothing certain as to what causes this. Generally, the N17 satellite sensor had a greater amount of scatter when compared with either MODIS sensor than the N16 satellite sensor. Also by looking at the comparison between Aqua and Terra, it is clear that even between the same sensors, there are factors that impede explanation, especially in the non-growing season. The result from this is that it becomes apparent that some type of relationship exists between a growing and a non-growing time of the year. Furthermore, it appears that in certain groupings of land, different factors are influential at different times of the year. For instance, the deciduous forest when looked at in a two-part regression allows for smaller groups of data that can more easily be reproduced. It has however, become apparent through the examination of the data that the MODIS sensors tend to saturate at a 0.92 NDVI level. Since the MODIS sensors typi-cally find a higher value of NDVI than AVHRR, the MODIS sensor is not able to evaluate high levels of NDVI values. The result is that in areas of extreme growth, such as forests, a ceiling is created for the MODIS NDVI values, causing a two-part regression being mandated to be used. This is primarily the case in the forest areas over the non-forest areas due to the higher rates of growth. In essence, the non-growing season/growing season relationship is far less a matter of time of the season, but rather a saturation problem of the sensor. There is still strong evidence however, to support the need for a split regres-sion in the examination of the evergreen forest land cover. Such a split in this land cover does allow for better explanation of the growing season. This does not solve the off-growing season problem; however, it would allow some seasonality metrics to be used with some modifications. The correlation of the evergreen forest non-growing season’s high rate of scatter may have something to do with the problem of snow and unmasked clouds found in the non-forest land cover types. It has generally been assumed that the most desirable NDVI values are the maximum values. It has also been the assumption that things such as snow, wa-ter, ice, smoke, haze, and smog would decrease the values of NDVI. The result-ing rationalization of the increased scatter in the evergreen forest results around the concept that during the non-growing season, some areas would still produce a high-level of NDVI due to the nature of the continual greenness. However, if some areas were covered by snow or ice, and not masked out by the sensor, it is possible that the reduction of one sensor may be greater than the other, causing a scatter, which would be relatively unpredictable. This same concept then would apply to the non-forest land cover types which have a unique problem with the floor saturation level created on by the MODIS sensor. The isolation of the floor effect to the off-season would suggest that the problem could be a result of snow and ice cover. As the floor effect primarily occurred in land cover types that are relatively non-obstructed by snow


masking growth, the MODIS sensor lowered the NDVI values greater than the AVHRR sensor. The majority of the focus within studies has been dealing with artificially low NDVI values and attempting to correct them. However, as alluded to earlier in the discussion, it has been shown in the forest land cover areas in section 3.2 that there are major issues with the comparison of extremely high NDVI levels with the MODIS sensor. This issue was noted by [7], noticing that the initial twelve months of MODIS vegetation data some intensively measured test sites appeared to have a saturation level of 0.90 NDVI. The problem clearly stated observes that, at extremely high values data is information lost, which exists extensively in the deciduous forest land cover but in the other forest types as well as row crops covers, is the result of the problem of the MODIS sensor saturating at 0.92 NDVI. The problem of saturation raises an interesting and critical question. By adjusting our data to match a sensor that cannot fully describe the differences be-tween two levels of growth due to saturation, should we be creating a calibration that loses data? The NDVI metric was created to work with AVHRR data and has been adapted for a number of sensors. With the MODIS sensor, however, we have an additional problem of having a generally larger NDVI value at all data points. This mixed with the saturation level of 0.92, will result in decreases of accuracy in high NDVI growth areas during the peak time of the year. As the primary use of this data focuses on the higher growth time period, this loss of explanation may not give an accurate representation of the characteristics of a season. Fixing this problem may need to be considered at the fundamental level by attempting to modify either the calculation of NDVI, or finding a way to remove the saturation level from the MODIS sensor. While the MODIS sensor is not viewed to be the next platform for NDVI, the goal being the VIIRS sensor, the MODIS platform will allow for comparison of data that will be necessary in order to allow for continuity between AVHRR’s long history of data, and VIIRS continuing coverage of remote sensing issues. Despite the problem of saturation however, there is still a considerable amount of data that can be gathered by employing some technique of adjust-ment between legacy data and current data at non-maximum values of NDVI.

CONCLUSION

Through the evaluation of the data we have seen factors that improve the calibration of the AVHRR data to the MODIS data. Most of the factors that have been found deal with MODIS NDVI values being reported as too low. Factors that have significant influence in the regressions include the need to mask clouds and buffering around clouds to deal with cloud interference as well as atmospheric moisture that is surrounding clouds and clouds not detected by the CLAVR algorithm. In addition, although no direct data shows the benefit of using a snow mask, the affects of snow appear to be evident in the non-growing season parts of the year data.


The primary focus of this study has been examining the effect of time peri-ods on the calibration of the sensors. As can be noticed in the data, the middle part of the year, which has been referred to as the growing part of the season has a relatively narrow collection of data running on the regression line. However it should be noted that the edges of the season have characteristics of the season, as well as the off-season. The reason for this is possibly related to a latitude based relation, as the different latitudes will effect the light reflections differently, which cannot be accounted for completely in simply a growing/non-growing season correction. To further prove this would require a closer look at evaluating data based on metric derived values in a more dynamic sense. The possibility of using a piece-wise regression may prove to be very useful in relating the data. However, without more data, the results may not be able to be applied to a longer calibration based algorithm. The main benefit of this study has shown that the effect of season based regression is a likely possibility for correctional based algorithms. While more research into the breakdown of the season as well as adding additional factors such as geographic location may improve the results, there is clearly a need to utilize some type of adjustment based on the time period of the season.

ACKNOWLEDGMENTS

This study was partially supported by the South Dakota Space Grant Con-sortium, with the assistance of National Center for ERO, and Augustana Col-lege.

RESOURCESCihlar, J. (1996). Identification of Contaminated Pixels in AVHRR Composite

Images for Studies of Land Biosphere, Remote Sensing of Environment 56, 149-163.

Fensholt, R. (2004). Earth observation of vegetation status in the Sahelian and Sudanian West Africa: comparison of Terra MODIS and NOAA AVHRR satellite data, International Journal of Remote Sensing, 25, 1641-1659.

Gitelson, A. and Kaufman, Y. (1998). MODIS NDVI optimization to fit the AVHRR data series – spectral considerations, Remote Sensing of Environ-ment, 66, 343-350.

Gallo, K., Ji, L., Reed, B., Dwyer, J., and Eidenshink, J. (2004). Comparison of MODIS and AVHRR 16-day normalized difference vegetation index composite data. Geophysical Research Letters, 31, L07502, doi:10.1029/2003GL019385.

Gallo, K., Ji, L., Reed, B., Eidenshink, J., and Dwyer, J. (2005). Multi-platform comparisons of MODIS and AVHRR normalized difference vegetation in-dex data. Remote Sensing of Environment, 99, 221-231.

Gao, X., Huete, A.R., and Didan, K. (2003). Multisensor comparisons and vali-dation of MODIS vegetation indices at the semiarid Jornada Experimental Range, IEEE Trans. Geoscience and Remote Sensing, 41, 2368-2381.


Hou, Yu-Tai, Campana, Kenneth A., Mitchell, Kenneth E., Yang, Shi-Keng, Stowe, Larry L. (1993) Comparison of an Experimental NOAA AVHRR Cloud Dataset with Other Observed and Forecast Cloud Datasets, Journal of Atmospheric and Oceanic Technology, 10, 833-849.

Huete, A., Didan, K., Miura, T., Rodriquez, E.P., Gao, X., and Ferreira, L.G. (2002) Overview of the radiometric and biophysical performance of the MODIS vegetation indices, Remote Sensing of Environment, 83, 195-213.

Lu, D., Mausel, P., Brondízio, E., and Moran, E. (2003) Change detection tech-niques, International Journal of Remote Sensing, 25, 2365-2407.

Myneni, R. B., Tucker, C. J., Asrar, G., and Keeling, C. D. (1998) Interannual variations in satellite-sensed vegetation index data from 1981 to 1991. Jour-nal of Geophysical Research, 103, 6145-6160.

Potter, C., Tan, P. Steinbach, M. Klooster, S. Kumar, V., Myneni, R. and Geno-vese V. (2003) Major disturbance events in terrestrial ecosystems detected using global satellite data sets. Global Change Biology, 9, 1005-1021.

Saleous, N. E., Vermote, E., and Roger, J. (1994) Operational atmospheric cor-rection of AVHRR visible and near infrared data. International Geoscience and Remote Sensing Symposium 1994; Pasadena, Calif., 220–222.

Steven, M., Malthus, T., Baret, F., Xu, H., and Chopping, M. (2003) Intercali-bration of vegetation indices from different sensor systems. Remote Sensing of Environment, 88, 412-422.

Stowe, L. L., Davis, P. A., and McClain, E. P. (1999) Scientific basis and initial evaluation of the CLAVR-1 global clear/cloud classification algorithm for the Advanced Very High Resolution Radiometer. Journal of Atmospheric and Oceanic Technology, 16, 656-681.

Swets, D. L., B. C. Reed, J. D. Rowland, S. E. Marko, “A Weighted Least-Squares Approach to Temporal NDVI Smoothing,” in Proceedings, 1999 AMPRS Annual Conference, Portland, OR, 1999.

Teillet, P. M. (1991). Radiometric and atmospheric correction procedures for AVHRR preprocessing in the solar reflective channels., Proceedings 5th Colloquium, Physical Measurements and Signatures in Remote Sensing, Courchevel, France, January 14-18, No ESA SP-319, 1991, 101-104.

Tucker, C. J., Slayback, D. A., Pinzon, J. E., Los, S. O. Myneni, R. B., and Tay-lor, M. G. (2001) Higher northern latitude normalized difference vegetation index and growing season trends from 1982 to 1999. International Journal of Biometeorology, 45,184-190.

Venturini, V., Bisht, G. Islam, S., and Jiang, L. (2004). Comparison of evapo-rative fractions estimated from AVHRR and MODIS sensors over South Florida. Remote Sensing of Environment, 93, 77-86.

Vogelmann, J.E., Howard, S.M ., Yang, L., Larson, C.R., Wylie, B.K., & Van Driel, J.N., (2001). Completion of the 1990’s National Land Cover Data SEt for the conterminous United States. Photogrammetric Engineering and Remote Sensing, 67, 650-662.

Zhou, L., Tucker, C. J., Kaufmann, R. K., Slayback, D., Shabanov, N. V., and Myneni, R. B. (2001) Variations in northern vegetation activity inferred from satellite data of vegetation index during 1981 to 1999. Journal of Geophysical Research, 106, 20,069-20,083.


RARE AND DECLINING FISHES OF SOUTH DAKOTA: A RIVER DRAINAGE SCALE PERSPECTIVE

Christopher W. Hoagstrom, Cari-Ann Hayer,Jason G. Kral and Steven S. Wall

Department of Wildlife & Fisheries Sciences

Charles R. Berry, Jr.U.S. Geological Survey

South Dakota Cooperative Fish and Wildlife Research Unit

South Dakota State UniversityBrookings, South Dakota 57007

ABSTRACT

We summarized the status of fishes that have declined from one or more of the 14 major river drainages in South Dakota and of fishes that are restricted to only one river drainage in the state, even if they have not declined. These species are of conservation concern because declines indicate sensitivity to environmen-tal change and restricted distributions indicate relatively high extinction risk. We documented 35 species that had declined from one or more river drainages and six species that have not declined, but are restricted to only one river drainage. The species were not necessarily of equal conservation concern because some had declined more than others, and some maintained greater present-day (post-1990) distributions than others. Thus, we determined relative conservation concern by combining the numeric rank of each species by the number of drainages from which it was missing with the number of drainages presently occupied. We also used a literature review to summarize impacts that affect each species elsewhere. This review suggested that impacts of erosion (siltation, pollution) and channel modification (channelization, riparian degradation, etc.) are the most substan-tial, but barriers to dispersal, water withdrawals, and wetland drainage are also important. This analysis is limited because it only considers declines at the river-drainage scale, but it nonetheless provides the first comprehensive summary of the status of South Dakota fishes.

Keywords

Fish conservation, South Dakota, human impacts, conservation assessment, river drainage scale

INTRODUCTION

There is increasing concern for freshwater fish conservation in North Amer-


ica (Pister 1999, Abell 2002, McKinney 2002). Many taxa are extinct (Miller et al. 1989), extinction rates are increasing (Wilcove et al. 1992), and may increase further (Ricciardi and Rasmussen 1999). Eight species are extinct from South Dakota (Hoagstrom 2006) and nine are state threatened or endangered, one of which is also federally endangered (SDGFP 2006). Another species, Topeka shiner Notropis topeka, is federally endangered, though not state endangered. There are often no quantifiable characteristics of endangered or threatened species. That is, the status of species afforded legal protection can vary greatly and some protected species may be more secure than unprotected ones. For ex-ample, the northern redbelly dace Phoxinus eos is widespread throughout South Dakota and a new population has been recently discovered (Morey and Berry 2004). This species is presumably listed as state threatened because it inhabits small, isolated habitats though it has not declined in South Dakota. In contrast, the lake chub Couesius plumbeus was once widespread within streams of the Black Hills and was also present in the Crow Creek and Little Missouri River drainages (Bailey and Allum 1962, Isaak et al. 2003). Based on recent surveys, only one substantial population now represents the species (Isaak et al. 2003), but despite these declines, the lake chub is unprotected. The primary reasons for such dis-crepancies are: (1) fish species status is difficult to assess because it relies on the availability of information that varies greatly by species and location; and (2) formal protection of fish species is instituted via a political process that is subject to many factors, only one of which is scientific data. Our purpose in this paper is to provide a perspective on species status by applying a standard assessment that treats all native South Dakota fishes equally. The intent of this approach is to provide the first statewide synthesis of native fish species declines, which we use to summarize patterns of decline, identify conser-vation priorities, recommend conservation strategies, and recommend research and management practices. Our hope is that this effort will improve awareness of fish conservation issues and highlight opportunities for conservation and res-toration. We do not necessarily want to increase the number of species that are given legal protection. Rather, we hope that increasing knowledge of fish species status will reduce the need for legal action and focus justifiable legal action where it is most needed.

METHODS

Hoagstrom (2006) reviewed fish collection records from South Dakota and constructed a list of fish species by major river drainage. Given the sparsity of data on fish distribution and abundance throughout South Dakota, we consid-ered the river drainage scale to be the smallest spatial scale suitable for a statewide analysis. Hoagstrom (2006) recognized 14 major river drainages within the state and divided the Missouri River Valley (the mainstem Missouri River with minor direct tributaries) into two sections with Fort Randall Dam as the boundary (Fig-ure 1). We used his list to identify native fish species that have declined or are rare at the river drainage scale. Declines and rarity at the river drainage scale are


a concern because they represent substantial range losses (Patton et al. 1998) and restriction to a low number of river drainages increases extinction risk (Moyle and Williams 1990). For this assessment, we used post-1990 surveys as representative of modern status because most of the major river drainages have been surveyed since that time (Kral and Berry 2005, Hoagstrom 2006). We defined rare species as native fishes that have not necessarily declined, but are restricted to one river drainage within South Dakota because restriction to one river drainage increases the risk of extinction (Moyle and Williams 1990). Our definition of declining species was native fishes that were missing from post-1990 collections in well-sampled river drainages. Well-sampled drainages were those in which post-1990 surveys were clearly more extensive than historical (pre-1990) surveys. We reasoned that if a fish species was documented by sparse historical surveys, but undetected by relatively extensive recent surveys, then there was legitimate reason to consider a species as truly ‘missing’. However, it is always possible that future surveys will discover undocumented populations of species we report as missing (Hayer et al. 2006). All river drainages of South Dakota were more extensively sampled recently than historically, except for the Bois de Sioux River, Crow Creek, and the Little Missouri River (Hoagstrom 2006). Species missing only from recent collections in one or more of these three poorly sampled drainages were not considered de-clining. However, we noted declines from the three poorly sampled river drain-

Figure 1. Map of South Dakota and adjacent areas that shows the 14 river drainages and two sections of the Missouri River Valley. River drainages are the (1) Little Missouri River, (2) Grand River, (3) Moreau River, (4) Cheyenne River (drainage 4a represents the Black Hills, a unique physiographic regions lying entirely within the Cheyenne River drainage), (5) upper Missouri River valley, (6) Bad River, (7) White River, (8) Niobrara River, (9) Ponca Creek, (10) Crow Creek, (11) lower Missouri River valley, (12) James River, (13) Vermillion River, (14) Big Sioux River, (15) upper Minnesota River, (16) Bois de Sioux River.


ages for species that were missing from one or more well sampled drainages. We present a review of the distribution for each declining and rare fish spe-cies and a general review of the biology of each species in relation to conservation in South Dakota and range-wide. Based on this literature review, we summarize impacts that affect declining and rare fish species and rank the impacts based on the number of species affected. We also summarize patterns of species declines and of species persistence by river drainage. These summaries are used to recom-mend priorities for native fish conservation, restoration, and management.


We documented 35 fish species that had declined within South Dakota at the river drainage scale and six species that were rare (Table 1). Eight of the 35 declining fish species were absent from all recent surveys and are presumably ex-tinct from South Dakota. Declines of an additional seven species have restricted them to one river drainage. Below, we provide a species by species review with an accompanying review of their conservation ecology. Following the species accounts, we summarize general patterns among species.

Table 1. Status of rare and declining fish species of South Dakota by river drainage (Hoagstrom 2006). Italics = species absent from post-1990 collections, n = native species. Fish names follow Nelson et al. (2004).

FAMILIES,

SPECIES, AND

SUBSPECIES

CENTRAL LOWLANDS GREAT PLAINS

Bois

de S

ioux

Riv

er

Upp

er M

inne

sota

Riv

er

Big

Siou

x Ri

ver

Verm

illio

n Ri

ver

Jam

es R

iver

Miss

ouri

Valle

y (lo

wer

)

Nio

brar

a Ri

ver

Ponc

a C

reek

Whi

te R

iver

Cro

w C

reek

Bad

Rive

r

Che

yenn

e Ri

ver

Mor

eau

Rive

r

Gra

nd R

iver

Miss

ouri

Valle

y (u

pper

)

Littl

e M

issou

ri Ri

ver

PETROMYZONTIDAE lampreysIchthyomyzon unicuspis n n n silver lampreyACIPENSERIDAE sturgeonsAscipenser fulvescens n lake sturgeonPOLYODONTIDAE paddlefishesPolyodon spathula n n n n n paddlefishLEPISOSTEIDAE garsLepisosteus osseus n n n n n n longnose garAMIIDAE bowfinsAmia calva n bowfin


FAMILIES,

SPECIES, AND

SUBSPECIES


Bois

de S

ioux

Riv

er

Upp

er M

inne

sota

Riv

er

Big

Siou

x Ri

ver

Verm

illio

n Ri

ver

Jam

es R

iver

Miss

ouri

Valle

y (lo

wer

)

Nio

brar

a Ri

ver

Ponc

a C

reek

Whi

te R

iver

Cro

w C

reek

Bad

Rive

r

Che

yenn

e Ri

ver

Mor

eau

Rive

r

Gra

nd R

iver

Miss

ouri

Valle

y (u

pper

)

Littl

e M

issou

ri Ri

ver

HIODONTIDAE mooneyesHiodon tergisus n mooneyeANGUILLIDAE freshwater eelsAnguilla rostrata n n n n n n American eelCYPRINIDAE carps and minnowsCouesius plumbeus n n n lake chubHybognathus argyritis n n n n n n n n n n n n western silvery minnowH. hankinsoni n n n n n n n n n n n n n n n brassy minnowH. placitus n n n n n n n n n plains minnowMacrhybopsis gelida n n n n n n sturgeon chubM. meeki n n sicklefin chubM. hyostoma n shoal chubNocomis biguttatus n n n hornyhead chubNotemigonus crysoleucas n n n n n n n n n n n n n golden shinerNotropis blennius n n river shinerN. heterodon n blackchin shinerN. heterolepis n n n n n n n n blacknose shinerN. percobromus n n carmine shinerN. shumardi n n n silverband shinerPhenacobius mirabilis n n n n suckermouth minnowPhoxinus erythrogaster n southern redbelly dacePlatygobio gracilis n n n n n n n n n n n flathead chub

Table 1 continued.


FAMILIES,

SPECIES, AND

SUBSPECIES


Bois

de S

ioux

Riv

er

Upp

er M

inne

sota

Riv

er

Big

Siou

x Ri

ver

Verm

illio

n Ri

ver

Jam

es R

iver

Miss

ouri

Valle

y (lo

wer

)

Nio

brar

a Ri

ver

Ponc

a C

reek

Whi

te R

iver

Cro

w C

reek

Bad

Rive

r

Che

yenn

e Ri

ver

Mor

eau

Rive

r

Gra

nd R

iver

Miss

ouri

Valle

y (u

pper

)

Littl

e M

issou

ri Ri

ver

Rhinichthys cataractae cataractae n n n n n n n n n n longnose daceR. obtusus n n n n n n n n western blacknose daceCATOSTOMIDAE suckersCarpiodes velifer n n highfin carpsuckerCatostomus catostomus n longnose suckerCatostomus platyrhynchus n mountain suckerHypentelium nigricans n n northern hog suckerIctiobus niger n n n black buffaloICTALURIDAE NorthAmerican catfishesIctalurus furcatus n n n n blue catfishNoturus flavus n n n n n n n n n n n n stonecatN. gyrinus n n n n n tadpole madtomPylodictis olivaris n n n n n n flathead catfishPERCOPSIDAE trout-perchesPercopsis omiscomaycus n n trout-perchGADIDAE codsLota lota maculosa n n n n burbotFUNDULIDAE topminnowsFundulus kansae n northern plains killifishF. sciadicus n n n n n n n plains topminnowPERCIDAE perchesEtheostoma exile n n n n n n n n n n n n n n Iowa darterPercina phoxocephala n slenderhead darter

Table 1 continued.


SPECIES ACCOUNTS

Silver Lamprey (Ichthyomyzon unicuspis).— South Dakota status: missing.

In South Dakota, the silver lamprey is missing from the Ver-million River, lower Missouri River valley, and Crow Creek drainages (Figure 2), which it formerly occupied (Bailey and Allum 1962). The silver lamprey is typi-cal of large rivers and lakes (Becker 1983). It is para-sitic and is most commonly associated with sturgeons and catfishes, though it may para-sitize a variety of fishes (Becker 1983). Silver lamprey requires clear-water streams without excessive silt where ammocoetes (a developmental stage) can burrow (Trautman 1981). The ammocoete stage lasts 4 to 7 years (Scott and Crossman 1973). Dams impact silver lampreys by blocking migra-tions of adults, which drift downstream after transforming from ammocoetes and later migrate upstream to spawn (Trautman 1981). Thus, the siltation of nursery streams and dam construction presumably impacted the silvery lampreys of South Dakota.

Lake sturgeon (Ascipenser fulvescens).— South Dakota status: restricted na-tive range.

In South Dakota, the lake sturgeon is present in the lower Missouri River valley below Gavins Point Dam (Kral per-sonal communication), which is the extent of its known range in the state (Figure 3). Lake sturgeon captures are rare in South Dakota (Shearer personal communication) and there is no record of the spe-cies spawning in the state.

Figure 2. Silver lamprey present and historical distribu-tion by drainage in South Dakota.

Figure 3. Lake sturgeon present and historical distribu-tion by drainage in South Dakota.


Paddlefish (Polyodon spathula).— South Dakota status: declining.

In South Dakota, the pad-dlefish is missing from the Big Sioux River drainage (Figure 4), which it formerly occupied (Bailey and Allum 1962). The species is still present in the Vermillion River (Kral per-sonal communication), James River (Berry et al. 1993), lower Missouri River valley (Wickstrom 1997, Berry and Young 2004), and upper Mis-souri River valley (Lott et al. 1994) drainages. It is man-aged as a game fish species in South Dakota. Paddlefish are typical of large rivers and lakes (Becker 1983). The species undergoes spawning migrations (e.g., Lein and DeVries 1998, Paukert and Fisher 2001) and thus may be negatively impacted by dams (Trautman 1981). The decline of paddlefish from South Dakota is presumably related to dams and habitat degradation, such as siltation that impacts spawning habitat.

Longnose gar (Lepisosteus osseus).— South Dakota status: declining.

In South Dakota, the long-nose gar is missing from the Upper Minnesota River, Big Sioux River, and Crow Creek drainages (Figure 5), which it historically occupied (Bailey and Allum 1962). The species is still present in the Vermillion River (Kral personal commu-nication), James River (Shear-er and Berry 2003), and lower Missouri River valley (Berry and Young 2004, Shuman et al. 2005) drainages. The longnose gar is most successful in clear-water habitats (Trautman 1981). Spawning habitat includes either silt-free rocky stream bottoms with moderate current or calm waters with vegetation (Cross 1967) and the species may undertake migrations to reach such habitat (Netsch and Witt 1962). Thus, the decline of longnose gar from South Dakota is likely related to dams that impede migrations (if they occur) and the

Figure 4. Paddlefish present and historical distribution by drainage in South Dakota.

Figure 5. Longnose gar present and historical distribu-tion by drainage in South Dakota.


degradation of spawning habitat due to siltation or loss of instream vegetation. Increasing turbidity due to erosion may also impact the species.

Bowfin (Amia calva).— Status: missing.

In South Dakota, the bow-fin was presumably present in the Upper Minnesota River drainage (Figure 8) and may have been widespread in the eastern portion of the state (Bai-ley and Allum 1962). The western extent of the historical range of bowfin is uncertain (Bailey and Allum 1962, Cross 1967). If the species was historically present in South Dakota, it may have been eliminated by degrada-tion of riverine wetlands such as backwaters and oxbows via drought and human impacts (Bailey and Allum 1962, Cross 1967).

Mooneye (Hiodon tergisus).— South Dakota status: missing.

In South Dakota, the mooneye is missing from the Big Sioux River drainage (Figure 7), where it formerly was present (Gilbert 1978). The mooneye occupies large rivers and lakes with clear waters (Trautman 1981). The species undergoes spring mi-grations for spawning (Traut-man 1981, Becker 1983). Human impacts that reduced water clarity and inhibited mi-gration, both of which have occurred in the Big Sioux Riv-er drainage (Sinning 1968), presumably eliminated the mooneye from South Dakota waters.

Figure 8. Bowfin present and historical distribution by drainage in South Dakota.

Figure 7. Mooneye present and historical distribution by drainage in South Dakota.


American eel (Anguilla rostrata).— South Dakota status: declining.

In South Dakota, the American eel is missing from the Upper Minnesota River, Big Sioux River, Vermillion River, and upper Missouri River valley drainages (Figure 8), where it formerly was present (Bailey and Allum 1962, Lee 1978). The species is still pres-ent in the James River and lower Missouri River valley drainages (Berry et al. 1993). The American eel is catad-romous, meaning it migrates downstream (to the sea) to spawn. Dams impede these migrations and have eliminated eels from many waters. In South Dakota, presence of dams has likely eliminated eels from the upper Missouri River valley and they presumably contributed to the decline of eels from the Upper Minnesota River, Big Sioux River, and Vermillion River drainages as well.

Lake chub (Couesius plumbeus).— South Dakota status: declining.

In South Dakota, the lake chub is missing from the Crow Creek and the Little Mis-souri River drainages (Figure 9), where it formerly occurred (Bailey and Allum 1962). The species is still present in the Cheyenne River drainage in the Black Hills, but has de-clined in distribution (Isaak et al. 2003). The major remnant population inhabits Deerfield Reservoir (Isaak et al. 2003). The lake chub may in-habit lakes or small streams (McPhail and Lindsey 1970, Brown 1971) but is typically found in cool or cold waters (Becker 1983). In South Dakota, the lake chub was known from clear and cold streams (Evermann and Cox 1896). The decline of the lake chub from the state may have resulted from stream warming due to habitat degradation such as siltation and the destruction of riparian areas. Water withdrawals may have also played a role. For example, the flow of many Black Hills streams is

Figure 8. American eel present and historical distribu-tion by drainage in South Dakota.

Figure 9. Lake chub present and historical distribution by drainage in South Dakota.


largely diverted (Burr et al. 1999), which certainly increases water temperatures in depleted reaches. It is also possible that introduced trout impacted lake chub via predation and competition.

Western silvery minnow (Hybognathus argyritis).— South Dakota status: declining.

In South Dakota, the west-ern silvery minnow is missing from the Big Sioux River, Ver-million River, Crow Creek, and Bad River drainages (Figure 10), where it formerly was present (Bailey and Allum 1962). The species is still present in the lower Missouri River valley (Berry and Young 2004), Niobrara River (Cun-ningham et al. 1995), White River (Cunningham et al. 1995, Fryda 2001, Harland 2003), Cheyenne River (Cun-ningham et al. 1995, Hampton and Berry 1997, Doorenbos 1998, Duehr 2004, Hoagstrom 2006), Moreau River (Loomis et al. 1999, Duehr 2004), Grand River (Erickson personal communication), upper Missouri River valley (Har-land 2003), and Little Missouri River drainages (Erickson personal communi-cation). The western silvery minnow is typical of relatively large streams of the Mis-souri River drainage (Baxter and Stone 1995) but is also known from some smaller tributaries (Duehr 2004). Although commonly associated with the similar plains minnow Hybognathus placitus, the distributions of these species in South Dakota are not identical (Bailey and Allum 1962, Hoagstrom 2006). As a large-river species, the decline of the western silvery minnow is probably associated with the upstream and downstream impacts of dams and reservoirs on the mainstem Missouri River and most of the major tributary rivers (Hesse et al. 1993). Changes associated with dams that appear to impact the western silvery minnow include sediment starvation and substrate scour, as well as the introduction of nonnative piscivorous fishes (Quist et al. 2004). The loss of shal-low, low-velocity habitats due to river channel degradation or modification also impacts this species (Pflieger and Grace 1987, Welker and Scarnecchia 2004). Dissection is necessary to distinguish the western silvery minnow from the plains minnow (Bailey and Allum 1962). This dissection is absolutely necessary to accurately document the distributions of both declining species.

Figure 10. Western silvery minnow present and historical distribution by drainage in South Dakota.


Brassy minnow (Hybognathus hankinsoni).— South Dakota status: declin-ing.

In South Dakota, the brassy minnow is missing from the Bois de Sioux River, Crow Creek, and Little Missouri River drainages (Figure 11), which it historically inhabit-ed (Bailey and Allum 1962). The species is still present in the Upper Minnesota River (Dieterman and Berry 1994, USGS 2002, 2003), Big Sioux River (Dieterman and Berry 1998, Blausey 2001, Milews-ki 2001, USGS 2002, 2004, Hayer et al. 2006), Vermillion River (Braaten and Berry 1997, Blausey 2001, USGS 2001, 2002, 2003, 2004), James River (Blausey 2001, USGS 2002, 2003, Shearer and Berry 2003), lower Missouri River valley (USGS 2002, Berry and Young 2004, Wickstrom 2004), Niobrara River (Cunningham et al. 1995, USGS 2003, Harland and Berry 2004), Ponca Creek (USGS 2003), White River (Cunningham et al. 1995, USGS 2002, 2004, Harland 2003), Cheyenne River (Erickson personal com-munication), Moreau River (Loomis et al. 1999), Grand River (Erickson per-sonal communication), and upper Missouri River valley (Johnson et al. 1995, Harland 2003) drainages. The brassy minnow is characteristic of small sluggish streams (Becker 1983). The species is highly tolerant and mobile, but is susceptible to mortality in dry-ing pools (Scheurer et al. 2003). It is sensitive to human impacts that reduce aquatic vegetation, increase water temperature, and increase turbidity (Cross and Moss 1987). Deep pools that persist throughout dry periods, perennial stream sections, and the absence of barriers that block dispersal are necessary to main-tain populations (Scheurer et al. 2003).

Figure 11. Brassy minnow present and historical distri-bution by drainage in South Dakota.


Plains minnow (Hybognathus placitus).— South Dakota status: declining.

In South Dakota, the plains minnow is missing from the lower Missouri River valley and Niobrara River drainages (Figure 12), which it historical-ly inhabited (Bailey and Allum 1962). The species is still pres-ent in the White River (Cun-ningham et al. 1995, Fryda 2001, USGS 2001, Harland 2003), Bad River (Milewski 2001, Harland 2003), Chey-enne River (Cunningham et al. 1995, Hampton and Berry 1997, Doorenbos 1998, Duehr 2004, Hoagstrom 2006), Moreau River (Loomis et al. 1999, Duehr 2004), Grand River (USGS 2001), upper Missouri River valley (USGS 2001, Harland 2003), and Little Missouri River (Erickson personal communication) drainages. The decline of the plains minnow is presumably caused by factors similar to those that have caused the decline of the western silvery minnow (see above). However, negative impacts of dams on plains minnow are better documented. The plains minnow appears to be negatively impacted by excessive dewatering by surface water diversion or groundwater pumping (Cross and Moss 1987, Bonner and Wilde 2000), sediment starvation and reduced water temperatures below dams (Anderson et al. 1983), and population fragmentation by dams (Winston et al. 1991, Pittenger and Schiffmiller 1997). Plains minnow may be particularly susceptible to population fragmentation because their semibuoyant-nonadhesive eggs and early protolarvae are susceptible to downstream drift and may be trans-ported into reservoirs or over diversion dams (Fausch and Bestgen 1997, Platania and Altenbach 1998). Over time, downstream losses of eggs and larvae could potentially deplete upstream populations because dams and reservoirs preclude recolonization from downstream. Finally, altered flow regimes below dams may impact the plains minnow, particularly because spawning occurs in conjunc-tion with high flow events (Cross and Moss 1987, Lehtinen and Layzer 1988). Although the early life history of the western silvery minnow is unknown, it is likely similar to that of the plains minnow and thus both species presumably are impacted similarly by dams and reservoirs.

Figure 12. Plains minnow present and historical distribu-tion by drainage in South Dakota.


Sturgeon chub (Macrhybopsis gelida).— South Dakota status: declining.

In South Dakota, the stur-geon chub is missing from the Grand River, upper Missouri River valley, and Little Mis-souri River drainages (Figure 13), which it formerly occupied (Bailey and Allum 1962). The species is still present in the lower Missouri River valley below Gavins Point Dam (Kral personal communication), the mainstem White River (Cun-ningham et al. 1995, Fryda 2001, USGS 2002, 2003), and the mainstem Cheyenne River (Cunningham et al. 1995, Hampton and Berry 1997, Hoagstrom 2006). Factors associated with the decline of the sturgeon chub are similar to those associated with other Great Plains fish-es, such as the western silvery minnow and plains minnow (see above). The sturgeon chub is most common in turbulent, swift-water habitats where sub-strate is relatively coarse (Bailey and Allum 1962, Cross 1967). Thus, sturgeon chub may be less adversely affected by chan-nel degradation and channel-ization than western silvery minnow and plains minnow (Pflieger and Grace 1987). However, it may be more ad-versely affected by dewatering than those species because it depends on swift-water habi-tats. For example, the combined effects of drought and habitat fragmentation by a reservoir could have led to the extinction of sturgeon chub from the Little Missouri River (Kelsch 1994). The early life-history of the sturgeon chub may be similar to that of the western silvery minnow and plains minnow (see above). In 2003, the senior author used a Moore egg collector (Altenbach et al. 2000) to collect drifting eggs from the Cheyenne River near the Plum Creek confluence (Haakon County), one of which was a sturgeon chub egg (Figure 14).

Figure 13. Sturgeon chub present and historical distribu-tion by drainage in South Dakota.

Figure 14. Captive reared sturgeon chub from egg col-lected using a Moore pelagic egg collector.


Shoal chub (Macrhybopsis hyostoma).— South Dakota status: restricted na-tive range.

The shoal chub is present in the lower Missouri Riv-er below Gavins Point Dam (Kral personal communica-tion), which is the extent of its known distribution in South Dakota (Figure 15). The shoal chub was formerly grouped with several closely related spe-cies under the name speckled chub Macrhybopsis aestivalis, but is now recognized as a dis-tinct species (Eisenhour 1999, 2004, Nelson et al. 2004).

Sicklefin chub (Macrhybopsis meeki).— South Dakota status: declining.

In South Dakota, the sick-lefin chub is missing from the upper Missouri River valley drainage (Figure 16), where it was previously present (Bailey and Allum 1962). The species is still present in the lower Missouri River valley drain-age below Gavins Point Dam (Berry and Young 2004, Kral personal communication). Given that the sicklefin chub is restricted to only the largest rivers of the Missouri River drainage, it is likely that major modifications to these rivers, primarily dams and reservoirs, have caused the decline of this species. Remaining populations of sicklefin chub are larg-est where rivers are least modified (Welker and Scarnecchia 2004). Like the sturgeon chub, the species is not as impacted by river channelization as western silvery minnow and plains minnow, presumably because benthic swift-water habitat is still available (Pflieger and Grace 1987).

Figure 15. Shoal chub present and historical distribution by drainage in South Dakota.

Figure 16. Sicklefin chub present and historical distribu-tion by drainage in South Dakota.


Hornyhead chub (Nocomis biguttatus).— South Dakota status: declining.

In South Dakota, the hornyhead chub is missing from the Bois de Sioux River and Big Sioux River drainages (Figure 17), where it was his-torically present (Bailey and Al-lum 1962). It is still present in the Upper Minnesota River drainage (Dieterman and Ber-ry 1994, Shearer unpublished data). The hornyhead chub com-monly inhabits small to me-dium-sized streams with clear water and gravel substrate (Cross 1967, Scott and Crossman 1973, Becker 1983). The species is impacted by siltation (Trautman 1981). In the Upper Minnesota River drainage of South Dakota, the hornyhead chub is associated with relatively cool waters and moder-ately sized streams (Dieterman and Berry 1994). Thus, siltation due to upland erosion, degradation to riparian vegetation, and depletion of water flows most likely explain the decline of hornyhead chubs within the state, similar to findings in Kansas (Cross 1967, Cross and Moss 1987).

Golden shiner (Notemigonus crysoleucas).— South Dakota status: declining.

In South Dakota, the gold-en shiner is missing from the Big Sioux River and James River drainages (Figure 18), where it was historically present (Bailey and Allum 1962). The species is still present in the Vermil-lion River (Braaten 1993), Upper Minnesota River, lower Missouri River valley (Berry and Young 2004), Nio-brara River (Cunningham et al. 1995, Harland and Berry 2004), White River (Cun-ningham et al. 1995, USGS 2003), Bad River (Milewski 2001), Cheyenne River (Duehr 2004), Moreau River (Loomis et al. 1999, Duehr 2004), Grand River (USGS 2001, 2003), upper Missouri River valley (Johnson et al. 1995, Lott et al. 2004), and Little Missouri River (Berry and Young 2004, Wickstrom 2004) drainages. Low-gradient streams and wetlands with clear water and abundant vegeta-tion characterize the habitat of the golden shiner (Scott and Crossman 1973,

Figure 17. Hornyhead chub present and historical distri-bution by drainage in South Dakota.

Figure 18. Golden shiner present and historical distribu-tion by drainage in South Dakota.


Trautman 1981, Becker 1983). As a result, this species is impacted by increased turbidity and siltation, wetland drainage, and channelization (Trautman 1981). Thus, it is unsurprising the golden shiner has declined from the river drainages of eastern South Dakota where wetland drainage, channelization, and siltation caused by erosion are widespread. The golden shiner is a commonly used bait-fish and introduced populations may become established in ponds, lakes, and impoundments (Eddy and Underhill 1974, Trautman 1981, Baxter and Stone 1995). As a result, this species may be more widespread than our records indi-cate because isolated populations may be present in lakes and impoundments.

River shiner (Notropis blennius).— South Dakota status: declining.

In South Dakota, the river shiner is missing from the Ver-million River drainage (Figure 19), where it was once present (Bailey and Allum 1962). The species is still present in the lower Missouri River valley (Berry and Young 2004, Shu-man et al. 2005). Little is known about the river shiner except that it typi-cally inhabits large rivers and lakes (Becker 1983). In the lower Missouri River, river shiner abundance increased after the river was channelized and impounded upstream, possibly because de-creased turbidity favored it (Pflieger and Grace 1987). Thus, the decline of river shiner from the Vermillion River may have resulted from siltation and pollution that increased turbidity.

Blackchin shiner (Notropis heterodon).— South Dakota status: missing.

In South Dakota, the black-chin shiner may have been pres-ent in the upper Minnesota River drainage (Bailey and Al-lum 1962, Figure 20). The blackchin shiner is characteristic of glacial lakes and is very sensitive to hu-man impacts (Trautman 1981, Becker 1983). It typically oc-cupies clear, quiet waters with abundant vegetation (Scott and Crossman 1973, Traut-

Figure 19. River shiner present and historical distribu-tion by drainage in South Dakota.

Figure 20. Blackchin shiner present and historical distri-bution by drainage in South Dakota.


man 1981). The rapid rate of its disappearance from throughout its historic range (Harlan and Speaker 1956, Eddy and Underhill 1974, Trautman 1981, Becker 1983) supports the supposition that this species was once present in South Dakota, but rapidly declined following human settlement.

Blacknose shiner (Notropis heterolepis).— South Dakota status: declining.

In South Dakota, the blac-knose shiner is missing from the Bois de Sioux River, James River, lower Missouri River valley, White River, and up-per Missouri River valley drainages (Figure 21), which it formerly occupied (Bailey and Allum 1962, Fryda 2001). The species is still present in the Upper Minnesota Riv-er (USGS 2004), Big Sioux River (USGS 2004), and Nio-brara River (Cunningham et al. 1995) drainages. The blacknose shiner typically inhabits clear waters with abundant vegeta-tion and clean substrates in glacial lakes and low gradient streams (Trautman 1981, Becker 1983). As with the blackchin shiner, the blacknose shiner has rap-idly declined throughout its range due to wetland loss, increased water turbidity, and siltation caused by erosion and pollution (Hubbs 1951, Harlan and Speaker 1956, Cross 1967, Smith 1979, Trautman 1981, Becker 1983, Cross and Moss 1987). These impacts most likely explain the dramatic decline of blacknose shiner from South Dakota (Bailey and Allum 1962).

Carmine shiner (Notropis percobromus).— South Dakota status: declining.

In South Dakota, the car-mine shiner is missing from the Big Sioux River drainage (Figure 22), where it formerly occurred (Dieterman and Berry 1998). The species is still present in the Upper Minne-sota River drainage (Dieter-man and Berry 1994, USGS 2004). The carmine shiner was formerly grouped with several other species under the name rosyface shiner (Notropis

Figure 21. Blacknose shiner present and historical distri-bution by drainage in South Dakota.

Figure 22. Carmine shiner present and historical distri-bution by drainage in South Dakota.


rubellus) but is now recognized as a distinct species (Wood et al. 2001, Nelson et al. 2004). The carmine shiner typically occupies large streams and small rivers with clear water and clean substrates (Harlan and Speaker 1956, Smith 1979). The species is normally found in habitats with moderate flow and high gradients (Cross 1967, Smith 1979, Becker 1983). Overall, the species is sensitive to turbidity, though some populations may be more tolerant than others (Harlan and Speaker 1956, Becker 1983). The decline of carmine shiner from the Big Sioux River drainage was presumably associated with stream modifications and siltation.

Silverband shiner (Notropis shumardi).— South Dakota status: missing.

In South Dakota, the sil-verband shiner is missing from the Vermillion River, lower Missouri River valley, and upper Missouri River valley drainages (Figure 23), which it formerly occupied (Bailey and Allum 1962). The silverband shiner is only known from large riv-ers and is tolerant of turbid-ity (Gilbert and Bailey 1962, Cross 1967, Smith 1979). It was a typical inhabitant of the historical Missouri River prior to human modifications for navigation and flood control but has declined since (Cross and Moss 1987). The decline of this species from the Missouri River and its major tributaries was apparently associated with the construction of dams and reservoirs, alteration of flow regimes, and degradation of riverine habitats.

Suckermouth minnow (Phenacobius mirabilis).— South Dakota status: declining.

In South Dakota, the suck-ermouth minnow is missing from the Big Sioux River and Crow Creek drainages (Figure 24), which it formerly occupied (Bailey and Allum 1962, Diet-erman and Berry 1998). The species is still present in the lower Missouri River valley (Lott et al. 1994) and upper Missouri River valley (Lott et al. 2004) drainages.

Figure 23. Silverband shiner present and historical distri-bution by drainage in South Dakota.

Figure 24. Suckermouth minnow present and historical distribution by drainage in South Dakota.


The suckermouth minnow primarily inhabits riffles in prairie streams with turbid waters, rich deposits of organic material, and clean gravel substrates (Cross 1967, Trautman 1981, Becker 1983). In the eastern portions of its range, the species has expanded its distribution in association with agricultural expan-sion (Trautman 1981, Becker 1983). The suckermouth minnow has declined from the western edge of its range due to declining base flows (Cross and Moss 1987). The reasons for the decline of this species from South Dakota are most likely associated with siltation of riffles or declining streamflow.

Southern redbelly dace (Phoxinus erythrogaster).—Status: restricted native range.

In South Dakota, the southern redbelly dace is only known from the Big Sioux River drainage (Springman and Banks 2005, Figure 25). It was first collected from the state in 2003 (Shearer, per-sonal communication), but was historically present nearby in the Big Sioux River drain-age of Minnesota (Underhill 1957, Springman and Banks 2005).

Flathead chub (Platygobio gracilis).—Status: declining.

In South Dakota, the flat-head chub is missing from the Big Sioux River and Vermil-lion River drainages (Figure 26), where it was once pres-ent (Bailey and Allum 1962). The species is still present in the lower Missouri River val-ley (USGS 2002, Berry and Young 2004), Niobrara River (Harland and Berry 2004), White River (Cunningham et al. 1995, Fryda 2001, USGS 2001, 2002, 2003, 2004, Harland 2003), Bad River (Milewski 2001, Harland 2003), Cheyenne River (Cunningham et al. 1995, Hampton and Berry 1997, Doorenbos 1998, USGS 2001, 2002, 2004, Duehr 2004), Moreau River (Loomis et al. 1999, USGS 2002, 2003, Duehr 2004),

Figure 25. Southern redbelly dace present and historical distribution by drainage in South Dakota.

Figure 26. Flathead chub present and historical distribu-tion by drainage in South Dakota.


Grand River (USGS 2001, 2002, 2004), upper Missouri River valley (Johnson et al. 1995, Harland 2003), and Little Missouri River drainages. The flathead chub occupies a wide range of stream sizes (Brown 1971) but is most typical of turbid rivers (McPhail and Lindsey 1970, Baxter and Stone 1995). It was a typical inhabitant of the Missouri River basin prior to human modifications for navigation and flood control, but has recently declined (Cross and Moss 1987, Pflieger and Grace 1987, Hesse et al. 1993). Range-wide de-clines are associated with changes in flow regimes and substrate caused by dams and reservoirs (Quist et al. 2004, Welker and Scarnecchia 2004). In South Da-kota, the flathead chub is still widely distributed among western rivers, but has declined from eastern rivers, perhaps due to river channel modifications (e.g., channelization). It declined from the lower Missouri River following channeliza-tion, presumably because habitat suitability declined (Pflieger and Grace 1987).

Longnose dace (Rhinichthys cataractae).— South Dakota status: declining.

In South Dakota, the long-nose dace is missing from the Bois de Sioux River, Bad Riv-er, and upper Missouri River valley drainages (Figure 27), where it was historically pres-ent (Bailey and Allum 1962). It is still present in the Nio-brara River (Cunningham et al. 1995, Harland and Berry 2004), Ponca Creek (USGS 2001), White River (Cun-ningham et al. 1995, Fryda 2001, USGS 2002, 2003, 2004, Harland 2003), Chey-enne River (Cunningham et al. 1995, Hampton and Berry 1997, Doorenbos 1998, USGS 2001, 2002, 2003, 2004, Duehr 2004), Moreau River (Loomis et al. 1999, USGS 2002, 2004, Duehr 2004), Grand River (USGS 2001, 2002, 2003, 2004), and Little Missouri River (Erickson personal communication) drainages. The longnose dace is normally found in flowing water and is most abundant in riffles (McPhail and Lindsey 1970, Becker 1983, Baxter and Stone 1995). The species is tolerant of turbidity and fluctuating environmental conditions (McPhail and Lindsey 1970, Becker 1983). Presumably, reservoirs have im-pacted longnose dace in the upper Missouri River valley drainage and siltation and channel degradation have impacted it in the Bois de Sioux and Bad River drainages.

Figure 27. Longnose dace present and historical distri-bution by drainage in South Dakota.


Western blacknose dace (Rhinichthys obtusus).— South Dakota status: declin-ing.

In South Dakota, the west-ern blacknose dace is missing from the lower Missouri Riv-er valley, White River, and Crow Creek drainages (Fig-ure 28), where it was histori-cally present (Bailey and Al-lum 1962, Fryda 2001). It is still present in the Upper Minnesota River (Dieterman and Berry 1994, USGS 2002, 2003, 2004), Big Sioux River (Dieterman and Berry 1998, Blausey 2001, Milewski 2001, USGS 2002, 2004, Hayer et al. 2006), Vermillion River (Blausey 2001), James River (Blausey 2001, Shearer and Berry 2003, mislabeled as longnose dace), and Niobrara River (Cunning-ham et al. 1995, Harland and Berry 2004) drainages. The western blacknose dace was once grouped with the eastern blacknose dace (Rhinichthys atratulus), but is now considered a distinct species (Nelson et al. 2004). The western blacknose dace typically inhabits permanent streams with mod-erate to high gradients, cool and clear waters, and clean substrates (Trautman 1981, Becker 1983). It is sensitive to human impacts such as siltation and defor-estation (Trautman 1981). Declines of the western blacknose dace from South Dakota are presumably associated with increased turbidity and siltation due to erosion, the degradation of riparian vegetation, and reductions in base flows.

Highfin carpsucker (Carpiodes velifer).— South Dakota status: declining.

In South Dakota, the high-fin carpsucker is missing from the Big Sioux River drainage (Figure 29), where it was for-merly present (Lee and Platania 1978). The species is still present in the lower Missouri River valley (Berry and Young 2004). The highfin carpsucker inhabits streams and rivers (Harlan and Speaker 1956, Cross 1967, Trautman 1981, Becker 1983). It has declined throughout its range (Becker

Figure 28. Western blacknose dace present and histori-cal distribution by drainage in South Dakota.

Figure 29. Highfin carpsucker present and historical distribution by drainage in South Dakota.


1983) and especially from the western edge of its historical distribution (Cross 1967). Its decline is associated with siltation and water pollution (Smith 1979) and it is sometimes migratory (Trautman 1981). Thus, dams and siltation in the Big Sioux River drainage may explain the decline of the species in South Dakota. A major issue in documenting the range of the highfin carpsucker is that small juveniles (< 7.6 cm) cannot be visually distinguished from the river carp-sucker (Carpiodes carpio) or quillback carpsucker (Carpiodes cyprinus; Trautman 1981). Thus, adult carpsucker captures should be inspected closely to search for the presence of highfin carpsuckers. Alternatively, molecular analyses can be conducted on juveniles where the presence of more than one carpsucker species is suspected.

Longnose sucker (Catostomus catostomus).—South Dakota status: restricted native range.

In South Dakota, the longnose sucker is present in the Cheyenne River drainage (Figure 30), which is the ex-tent of its known range within the state (Bailey and Allum 1962). It is mainly restricted to tributary streams that issue from the Black Hills (Isaak et al. 2003).

Mountain sucker (Catostomus platyrhynchus).—South Dakota status: re-stricted native range.

In South Dakota, the mountain sucker is present in the Cheyenne River drainage (Figure 31), which is the extent of its known range within the state (Bailey and Allum 1962). It is restricted to mountain-ous streams of the Black Hills (Isaak et al. 2003).

Figure 30. Longnose sucker present and historical distri-bution by drainage in South Dakota.

Figure 31. Mountain sucker present and historical distri-bution by drainage in South Dakota.


Northern hog sucker (Hypentelium nigricans).— South Dakota status: miss-ing.

In South Dakota, the northern hog sucker is miss-ing from the upper Minnesota River and Big Sioux River drainages (Figure 32), where it was formerly present (Bailey and Allum 1962, Dieterman and Berry 1998). The northern hog suck-er is an inhabitant of swift-water habitats in clear-water streams with clean substrate (Cross 1967, Trautman 1981). The species is impacted by siltation, pollution, channel modification, and depleted base flows (Cross 1967, Eddy and Underhill 1974, Smith 1979, Trautman 1981). Because the northern hog sucker migrates be-tween summer habitat in small streams to winter habitat in larger streams and rivers (Harlan and Speaker 1956, Smith 1979, Trautman 1981), it may also have been impacted by dams, which are barriers to dispersal.

Black buffalo (Ictiobus niger).— South Dakota status: missing.

In South Dakota, the black buffalo is missing from the Big Sioux River, James River, and lower Missouri River valley drainages (Figure 33), where it once was present (Moen 1970, Shute 1978). The black buffalo com-monly inhabits large rivers and reservoirs (Cross 1967, Becker 1983). The decline of this species from South Dakota is most likely associated with modification of the Big Sioux, James, and Missouri rivers for navigation, flood control, and water withdrawal. The black buffalo is commonly confused with the smallmouth buffalo (Ic-tiobus bubalus) and bigmouth buffalo (Ictiobus cyprinellus), so care must be taken in their identification (Smith 1979, Trautman 1981).

Figure 32. Northern hog sucker present and historical distribution by drainage in South Dakota.

Figure 33. Black buffalo present and historical distribu-tion by drainage in South Dakota.


Blue catfish (Ictalurus furcatus).— South Dakota status: declining.

In South Dakota, the blue catfish is missing from the James River and lower Missouri River valley drainages (Figure 34), where it was formerly col-lected (Bailey and Allum 1962, Berry et al. 1993). It is still present in the Big Sioux River (Dieterman and Berry 1998) and upper Missouri River valley (Johnson et al. 1995) drainages. The blue catfish inhab-its large rivers and reservoirs (Harlan and Speaker 1956, Cross 1967, Smith 1979). It may be impacted by river modifications for navi-gation (Trautman 1981). The species has been stocked into the Missouri River on the Nebraska border and into Lewis and Clark reservoir (Hesse et al. 1989). The overall rarity of the blue catfish in South Dakota despite stocking suggests that modern habitat conditions are poor for this species.

Stonecat (Noturus flavus).— South Dakota status: declining.

In South Dakota, the ston-ecat is missing from the Up-per Minnesota River drain-age (Figure 35), where it was formerly present (Bailey and Allum 1962). The species is still present in the Big Sioux River (Dieterman and Berry 1998, Blausey 2001, Milewski 2001, Hayer et al. 2006), Ver-million River (Braaten and Berry 1997, Blausey 2001), James River (Blausey 2001, USGS 2002, Shearer and Berry 2003), lower Missouri River valley (Berry and Young 2004, Shuman et al. 2004), Niobrara River (Cunningham et al. 1995), White River (Cunningham et al. 1995, Fryda 2001, USGS 2001, 2002, 2003, 2004, Harland 2003), Cheyenne River (Cunningham et al. 1995, Hampton and Berry 1997, Doorenbos 1998, USGS 2001, 2003, Duehr 2004), Moreau River (Loomis et al. 1999, USGS 2002, 2004), Grand River (USGS 2001, 2002, 2004), upper Missouri River (Johnson et al. 1995), and Little Missouri River (Erickson personal communication) drainages.

Figure 34. Blue catfish present and historical distribu-tion by drainage in South Dakota.

Figure 35. Stonecat present and historical distribution by drainage in South Dakota.


The stonecat typically occupies swift-water habitats in larger streams and rivers (Scott and Crossman 1973). The species is susceptible to pollution and siltation caused by erosion (Cross 1967, Eddy and Underhill 1974, Trautman 1981). It does not survive in impoundments but may occur downstream from dams (Trautman 1981). The loss of the stonecat from the Upper Minnesota River drainage in South Dakota presumably resulted from increased siltation or pollution.

Tadpole madtom (Noturus gyrinus).— South Dakota status: declining.

In South Dakota, the tad-pole madtom is missing from the lower Missouri River valley (Figure 36), which it formerly occupied (Bailey and Allum 1962). The species is still present in the Upper Min-nesota River (Dieterman and Berry 1994), Big Sioux River (Dieterman and Berry 1998, Blausey 2001, Milewski 2001, USGS 2002, 2004, Hayer et al. 2006), Vermillion River (Braaten 1993, Blausey 2001, USGS 2002, 2004), and James River (Blausey 2001, Shearer and Berry 2003) drainages. The tadpole madtom inhabits calm waters in lakes and streams and is somewhat tolerant of increasing turbidity (Cross 1967, Trautman 1981, Becker 1983). Wetland drainage, siltation, and stream channelization impact the habi-tats of this species (Trautman 1981). Improving water quality in the Big Sioux River drainage apparently benefited this species (Dieterman and Berry 1998). The absence of the tadpole madtom from the lower Missouri River valley pre-sumably resulted from erosion and siltation, which degrades tributary streams, and channel scour and flood control in the mainstem Missouri River, which eliminates floodplain habitats and habitat connectivity.

Figure 36. Tadpole madtom present and historical distri-bution by drainage in South Dakota.


Flathead catfish (Pylodictis olivaris).— South Dakota status: declining.

In South Dakota, the flat-head catfish is missing from the White River drainage (Figure 37), where it was formerly pres-ent (Bailey and Allum 1962). The species is still present in the Big Sioux River (Dieter-man and Berry 1998, Kir-by 2001), Vermillion River (Braaten 1993), James River (Shearer and Berry 2003), lower Missouri River valley (Wickstrom 1997, Berry and Young 2004, Shuman et al. 2005), and upper Missouri River valley (Lott et al. 1994) drainages. The flathead catfish inhabits large rivers and their reservoirs (Cross 1967, Eddy and Underhill 1974). The species is impacted by pollution (Trautman 1981). Reasons for their absence from the White River are unknown. However, it may only occupy the river during high flow years, as suggested by Cross (1967) for streams in western Kansas.

Trout-perch (Percopsis omiscomaycus).— South Dakota status: declining.

In South Dakota, the trout-perch is missing from the Up-per Minnesota River drainage (Figure 38), where it was for-merly present (Bailey and Allum 1962). The species is still present in the Big Sioux River drainage (Dieterman and Ber-ry 1998, Hayer et al. 2006). The trout-perch typically occupies lakes and streams (Scott and Crossman 1973, Trautman 1981, Becker 1983), but may also occupy large riv-ers and floodplain lakes (Smith 1979). In lakes, the species makes daily migrations from deep water in the day to shallow water at night and may migrate from lakes into tributary streams to spawn (Harlan and Speaker 1956, Scott and Crossman 1973, Becker 1983). In streams, trout-perch are associated with clean substrates and deep pools with cover for hiding during the day (Trautman 1981). The decline of trout-perch from the Upper Minnesota River drainage in South Dakota may be related to

Figure 37. Flathead catfish present and historical distri-bution by drainage in South Dakota.

Figure 38. Trout-perch present and historical distribu-tion by drainage in South Dakota.


the loss of deep stream pools for cover or the loss of spawning habitat for lake populations.

Burbot (Lota lota maculosa).— South Dakota status: declining.

In South Dakota, the bur-bot is missing from the Ver-million River and Cheyenne River drainages (Figure 39), where it was formerly present (Bailey and Allum 1962). The species is still present in the lower Missouri River valley (Berry and Young 2004) and upper Missouri River valley (Johnson et al. 1998) drain-ages. The burbot occurs in large rivers, lakes, and reservoirs (Cross 1967, Smith 1979, Becker 1983). It is associated with cold-deep waters (Trautman 1981). It may ascend tributaries to spawn (Harlan and Speaker 1956, Eddy and Underhill 1979). The species may not have been a permanent inhabitant of the Vermil-lion or Cheyenne rivers, making migrations only when conditions were optimal (e.g., high flows) as suggested by Cross (1967) for the Kansas River. Dams on tributary rivers may impact spawning migrations.

Northern Plains killifish (Fundulus kansae).— South Dakota status: re-stricted native range.

In South Dakota, the northern plains killifish is present only in the Cheyenne River drainage (Hampton and Berry 1997, USGS 2001, 2002, 2004, Duehr 2004, Fig-ure 40), which is the extent of its known range in the state. Some researchers consider the northern plains killifish to be nonnative in the Cheyenne River drainage (e.g., Miller 1955, Kreiser et al. 2001), but we disagree with this view because there is no evidence of introduction and there is biogeographical support for the native presence of the species (Hoagstrom 2006). For a time, this species was grouped with the south-

Figure 39. Burbot present and historical distribution by drainage in South Dakota.

Figure 40. Northern plains killifish present and historical distribution by drainage in South Dakota.


ern plains killifish, (Fundulus zebrinus) but presently it is considered a distinct species (Kreiser et al. 2001, Nelson et al. 2004).

Plains topminnow (Fundulus sciadicus).—South Dakota status: declining.

In South Dakota, the plains topminnow is missing from the Vermillion River and lower Missouri River valley drain-ages (Figure 41), where it was formerly present (Bailey and Al-lum 1962). The species is still present in the Big Sioux River (Morey personal communica-tion), James River (Blausey 2001), Niobrara River (Cun-ningham et al. 1995, USGS 2002, 2003, Harland and Ber-ry 2004), White River (Cun-ningham et al. 1995, USGS 2003), and Cheyenne River (Hampton and Berry 1997, Duehr 2004, Hoag-strom 2006) drainages. The plains topminnow typically occupies small streams and wetlands with abundant aquatic vegetation and clean substrates (Miller 1955, Baxter and Stone 1995). The species has declined throughout much of its range (Harlan and Speaker 1956, Bailey and Allum 1962, Baxter and Stone 1995). Declines presumably are related to the loss of wetland habitats and degradation of stream channels. Kazmierski (1966) studied plains topminnow of Say Brook, a Ver-million River tributary, and concluded the species was largely confined to one sampling station due to degraded habitat conditions elsewhere.

Iowa darter (Etheostoma exile).— South Dakota status: declining.

In South Dakota, the Io-wa darter is missing from the Bois de Sioux River, lower Missouri River valley, Crow Creek, and Little Missouri River drainages (Figure 42), where it was formerly present (Bailey and Allum 1962, Bich and Scalet 1977). The species is still present in the Upper Minnesota River (Dieterman and Berry 1994, USGS 2002, 2003), Big Sioux River (Diet-erman and Berry 1998, Blau-

Figure 41. Plains topminnow present and historical dis-tribution by drainage in South Dakota.

Figure 42. Iowa darter present and historical distribu-tion by drainage in South Dakota.


sey 2001, Milewski 2001, Hayer et al. 2006), Vermillion River (Schmulbach and Braaten 1993), James River (USGS 2003), Niobrara River (Cunningham et al. 1995, Harland and Berry 2004), White River (Cunningham et al. 1995, USGS 2003), Cheyenne River (Duehr 2004), Moreau River (Loomis et al. 1999), Grand River (Erickson personal communication), and upper Missouri River valley (Johnson et al. 1995) drainages. The Iowa darter is typical of low gradient streams and wetlands with abun-dant aquatic vegetation (Smith 1979, Becker 1983) and inhabits cool waters (Trautman 1981). Wetland drainage, increasing turbidity, and siltation due to erosion impact this species (Scott and Crossman 1973, Smith 1979, Trautman 1981).

Slenderhead darter (Percina phoxocephala).— South Dakota status: missing.

In South Dakota, the slen-derhead darter is missing from the Upper Minnesota River drainage (Figure 43), where it was histori-cally collected (Bailey and Allum 1962). The slenderhead darter typi-cally occupies swift water habi-tats of relatively large, permanent streams that have clean substrate (Cross 1967, Trautman 1981, Becker 1983). The species is declining throughout its range (Becker 1983). Siltation of gravel and sand substrates have led to the decline of the species in Illinois (Smith 1979) and Ohio (Trautman 1981), which may explain the loss of slenderhead darter from South Dakota.

General trends

Status varied among declining fish species (Figure 44). Species varied both in the number of drainages from which they declined and in the number of drain-ages where they persisted. We used the combination of these two factors to rank declining and rare fish species by river drainage status (i.e., level of conservation concern, Table 2). Four extinct species that were historically present in more than one of the river drainages (silver lamprey, silverband shiner, northern hog sucker, black buffalo) ranked as highest conservation concern because they had declined from several drainages and were extinct. The second highest conserva-tion concern was for a group of four fishes that were missing from two or more river drainages, but were persistent in three or less river drainages (American eel, lake chub, hornyhead chub, blacknose shiner). Third highest conservation concern was for extinct species that historically occupied only one river drainage (bowfin, mooneye, blackchin shiner, slenderhead darter). Fourth was species

Figure 43. Slenderhead darter present and historical distribution by drainage in South Dakota.


that persisted in roughly the same number of river drainages from which they were missing (19 species). Fifth was rare species that have not declined but were restricted to a single river drainage (lake sturgeon, shoal chub, southern redbelly dace, mountain sucker, longnose sucker, northern plains killifish). Sixth (and last) was a group of species that declined from two or less river drainages and were widespread in recent surveys. River drainages of eastern South Dakota (Central Lowlands geomorphic province) and both sections of the Missouri River valley exhibited relatively high numbers of species losses (Figure 45). Poorly-sampled river drainages (Bois de Sioux River, Crow Creek, Little Missouri River) also appeared to have high num-bers of species losses, but this result may be an artifact of limited recent sampling. On the other hand, well-sampled river drainages of western South Dakota (Great Plains geomorphic province) had relatively few species losses. This suggests that habitat restoration and conservation may be most critical in the Central Low-lands and Missouri River valley. Most river drainages had eight or more declining or rare fish species present in recent surveys (Figure 45). Exceptions were poorly-sampled river drainages (Bois de Sioux River, Crow Creek, Little Missouri River) and the Ponca Creek and Bad River drainages. Reasons that the Ponca Creek and Bad River drainages supported few declining or rare species are uncertain, but likely have to do with the small size of these river drainages (Hoagstrom and Berry 2006) and human

Figure 44. Level of concern for fish species based on the combined rank among species of number of drainages where missing in recent collections and number of drainages where present. Species abbreviations are the first three letters of the generic epithet and first three letters of the spe-cies epithet, except for blackchin shiner Notropis heterodon (abbreviation = nothetdon) and blacknose shiner Notropis heterolepis (abbreviation = nothetpis). See Table 1 for generic and species epithets.


Table 2. Level of concern (1 through 6, highest to lowest) for declining South Dakota fishes and summary of impacts that have affected each species outside of South Dakota. Impacts are: 1-channel changes, 2-erosion, 3-dispersal barriers, 4-water withdrawal/drought, 5-wetland drainage.

FAMILIES, SPECIES,AND SUBSPECIES

LEVEL OFCONCERN IMPACTS

PETROMYZONTIDAE lampreysIchthyomyzon unicuspis silver lamprey 1 2, 3POLYODONTIDAE paddlefishesPolyodon spathula paddlefish 4 2, 3LEPISOSTEIDAE garsLepisosteus osseus longnose gar 4 2, 3AMIIDAE bowfinsAmia calva bowfin 3 1, 4, 5HIODONTIDAE mooneyesHiodon tergisus mooneye 3 2, 3ANGUILLIDAE freshwater eelsAnguilla rostrata American eel 2 3CYPRINIDAE carps and minnowsCouesius plumbeus lake chub 2 1, 2, 4Hybognathus argyritis western silvery minnow 4 1, 3H. hankinsoni brassy minnow 4 1, 2, 3, 4H. placitus plains minnow 4 1, 3, 4Macrhybopsis gelida sturgeon chub 4 1, 3, 4M. meeki sicklefin chub 4 1, 3, 4Nocomis biguttatus hornyhead chub 2 1, 2, 4Notemigonus crysoleucas golden shiner 6 1, 2, 5Notropis blennius river shiner 4 2N. heterodon blackchin shiner 3 2, 5N. heterolepis blacknose shiner 2 2, 5N. percobromus carmine shiner 4 1, 2N. shumardi silverband shiner 1 1, 2, 3Phenacobius mirabilis suckermouth minnow 4 2, 4Platygobio gracilis flathead chub 6 1Rhinichthys cataractae cataractae longnose dace 4 1, 2R. obtusus western blacknose dace 4 1, 2, 4CATOSTOMIDAE suckersCarpiodes velifer highfin carpsucker 4 2, 3Hypentelium nigricans northern hog sucker 1 1, 2, 3, 4Ictiobus niger black buffalo 1 1, 3, 4ICTALURIDAE North American catfishesIctalurus furcatus blue catfish 4 1Noturus flavus stonecat 6 2N. gyrinus tadpole madtom 6 1, 2, 5Pylodictis olivaris flathead catfish 6 ?PERCOPSIDAE trout-perchesPercopsis omiscomaycus trout-perch 4 1, 2GADIDAE codsLota lota maculosa Burbot 4 1, 3FUNDULIDAE topminnowsF. sciadicus plains topminnow 4 1, 5PERCIDAE perchesEtheostoma exile Iowa darter 4 2, 5Percina phoxocephala slenderhead darter 3 2


impacts. This suggests that conservation would be benefi-cial in all major river drainages of South Dakota. Based on our literature re-view, human impacts elsewhere that have led to declines of the species we studied in other portions of their range can be grouped into five broad cat-egories (listed in order of im-portance): (1) erosion, which causes siltation of aquatic sub-strates and increased turbidity due to suspended sediment; (2) channel alteration, which includes stream inundation, stream channelization, flow regime manipulation, and ri-parian vegetation removal; (3) dispersal barriers, which in-clude dams and road crossings; (4) water withdrawals/drought; and (5) wetland drainage. Erosion may negatively affect many fishes, but our literature review indicated that at least 23 of the declining fish species were susceptible to erosion elsewhere within their ranges (Table 2). Similarly, channel alteration likely has negative affects on many fishes, but our literature review indicated that at least 21 of the declining fish species of South Dakota were sensitive to channel changes in other regions (Table 2). According to our review, at least 15 species that have declined from South Dakota are migratory, which makes them susceptible to the impacts of dispersal barriers (Table 2). However, barriers may affect non-migratory fishes as well (Wall and Berry 2004). Ultimately, water withdrawals and drought affect all fishes, but our literature review suggests that 11 of the declining South Dakota fishes are very sensitive to reduced water supplies (Table 2). Wetland drainage also has many impacts on streams, but presumably was most detrimental to seven declining South Dakota fishes that typically inhabit wetlands (Table 2). The decline of many of the fishes we studied was likely attributable to a suite of impacts, but our analysis suggests that erosion control and river channel conservation and restoration should be a high priority for native fish conservation in South Dakota, followed by improved fish passage, increased base flows, and wetland restoration.

Figure 45. Number of species missing from each major South Dakota river drainage (top) and number of de-clining or rare species persisting in major South Dakota river drainages (bottom) based on recent (post-1990) fish surveys.


CONCLUSIONS

This analysis considers only river drainage scale declines and is not intended to represent all potential conservation concerns. Particularly, status assessments at smaller spatial scales, such as Gap analysis, provide additional perspectives on the status of each fish species (Sylvester 2004). For example, flathead chub and longnose dace have declined from some South Dakota river drainages, but they remain widespread in the state and are abundant throughout many river drain-ages (Kral and Berry 2005, Hoagstrom 2006). Thus, even though they have exhibited declines at the river drainage scale, their overall distribution and their abundance within river drainages persist suggest the level of concern for these species should be relatively low. Nonetheless, we believe that our ranking of fish species by level of concern (Figure 2, Table 2) provides a useful summary of fish species status at the river drainage scale and may be useful for determining which fishes require additional study or legal protection. For example, species in the level of concern group two may require specific attention because they have declined from multiple drainages and their present distribution is limited. Further, recent studies in river drainages where these species are present indicate that they are not locally abundant. Studies focusing on the local distribution and abundance of these species (e.g., Wall et al. 2004) would be beneficial because they better delineate the status of the species of concern and provide justification either for or against legal protection, depending on findings.

ACKNOWLEDGEMENTS

Federal Aid in Sport Fish Restoration funds administered by the South Da-kota Department of Game, Fish and Parks (Project Number F-21-R and F-15-R) supported this research. The South Dakota Cooperative Fish and Wildlife Re-search Unit is jointly supported by the South Dakota Department of Game, Fish and Parks, U.S. Geological Survey, U.S. Fish and Wildlife Service, and South Dakota State University. We thank Jack Erickson and Jeff Shearer of the South Dakota Department of Game, Fish and Parks and Greg Wanner of the U.S. Fish and Wildlife Service for providing us up-to-date information on fish collections throughout the state and Steven Herrington and Jeff Shearer for providing edito-rial comments.

LITERATURE CITED

Abell, R. 2002. Conservation biology for the biodiversity crisis: a freshwater follow-up. Conservation Biology 16:1435-1437.

Altenbach, C.S., R.K. Dudley, and S.P. Platania. 2000. A new device for collect-ing drifting semibuoyant fish eggs. Transactions of the American Fisheries Society 129:296-300.


Anderson, K.A., T.L. Beitinger, and E.G. Zimmerman. 1983. Forage fish as-semblages in the Brazos River upstream and downstream from Possum King-dom Reservoir, Texas. Journal of Freshwater Ecology 2:81-88.

Bailey, R.M. and Allum, M.O. 1962. Fishes of South Dakota. Ann Arbor, MI: Miscellaneous Publications, Museum of Zoology, University of Michigan, No. 119.

Baxter, G.T., and M.D. Stone. 1995. Fishes of Wyoming. Wyoming Game and Fish Department, Cheyenne.

Becker, G.C. 1983. Fishes of Wisconsin. Madison, WI: University of Wis-consin Press.

Berry, C.R., Jr., W.G. Duffy, R. Walsh, S. Kubeny, D. Schumaker, and G. Van Eeckout. 1993. The James River of the Dakotas. In L.W. Hesse, C.B. Stalnaker, N.G. Benson, and J.R. Zuboy, editors, Proceedings of the sym-posium on restoration planning for rivers of the Mississippi River ecosystem. National Biological Survey Report 19, U.S. Government Printing Office, Washington, D. C. pp 70-86.

Berry, C.R., Jr., and B. Young. 2004. Fishes of the Missouri National Recre-ational River, South Dakota and Nebraska. Great Plains Research 14:89-114.

Bich, J.P., and C.G. Scalet. 1977. Fishes of the Little Missouri River, South Da-kota. Proceedings of the South Dakota Academy of Sciences 56:163-177.

Blausey, C.M. 2001. The status and distribution of the Topeka shiner Notropis topeka in eastern South Dakota. M.S. thesis, South Dakota State University, Brookings. 113 p.

Bonner, T.H. and Wilde, G.R. 2000. Changes in the Canadian River fish assemblage associated with reservoir construction. Journal of Freshwater Ecology 15:189-198.

Braaten, P.J. 1993. The influence of habitat structure and environmental vari-ability on habitat use by fish in the Vermillion River, South Dakota. M.S. Thesis. South Dakota State University, Brookings. 99 p.

Braaten, P.J., and C.R. Berry, Jr. 1997. Fish associations with four habitat types in a South Dakota prairie stream. Journal of Freshwater Ecology 12:477-489.

Brown, C.J.D. 1971. Fishes of Montana. Big Sky Books, Montana State Uni-versity, Bozeman.

Burr, M.J., Teller, R.W. and Neitzert, K.M. 1999. Water resources data, South Dakota, water year 1998. Washington, D.C.: U.S. Geological Survey, Wa-ter-Data Report SD-98-1.

Cross, F.B. 1967. Handbook of fishes of Kansas. Miscellaneous Publication No. 45. Museum of Natural History, University of Kansas, Lawrence.

Cross, F.B. and Moss, R.E. 1987. Historic changes in fish communities and aquatic habitats in plains streams of Kansas. In: Matthews, W.J. and Heins, D.C., editors. Community and evolutionary ecology of North American stream fishes. Norman, OK: University of Oklahoma Press. pp. 155-165.

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 River. Proceedings of the South Dakota Academy of Science 74:55-64.


Dieterman, D., and C.R. Berry, Jr. 1994. Fishes in seven streams of the Min-nesota River drainage in northeastern South Dakota. Proceedings of the South Dakota Academy of Science 73:23-30.

Dieterman, D., and C.R. Berry, Jr. 1998. Fish community and water quality changes in the Big Sioux River. Prairie Naturalist 30:199-224.

Doorenbos, R.D. 1998. Fishes and habitat of the Belle Fourche River, South Dakota. MS Thesis. Brookings: South Dakota State University. 79 pp.

Duehr, J.P. 2004. Fish and habitat relations at multiple spatial scales in Chey-enne River basin, South Dakota. MS Thesis. Brookings: South Dakota State University. 79 pp.

Eddy S., and J.C. Underhill. 1974. Northern fishes, with special reference to the upper Mississippi Valley, third edition, revised and expanded. University of Minnesota Press, Minneapolis.

Eisenhour, D.J. 1999. Systematics of Macrhybopsis tetranema (Cypriniformes: Cyprinidae). Copeia 1999:969-980.

Eisenhour, D.J. 2004. Systematics, variation, and speciation of the Macrhybopsis aestivalis complex west of the Mississippi River. Bulletin of the Alabama Museum of Natural History 23:9-48.

Evermann, B.W., and U.O. Cox. 1896. Report upon the fishes of the Missouri River basin. In Report of the U.S. Commissioner of fish and fisheries for 1894. U.S. Government Printing Office, Washington, D.C. pp 325-429.

Fausch, K.D. and Bestgen, K.R. 1997. Ecology of fishes indigenous to the central and southwestern Great Plains. In: Knopf, F.L. and Samson, F.B., editors. Ecology and conservation of Great Plains vertebrates. New York, NY: Springer-Verlag, Ecological Studies 125. pp 131-166.

Fryda, D.D. 2001. A survey of the fishes and habitat of the White River, South Dakota. Master’s thesis, South Dakota State University, Brookings.

Gilbert, C.R.. 1978. Hiodon tergisus Lesueur, mooneye. In D.S. Lee, C.R. Gilbert, C.H. Hocutt, R.E. Jenkins, D.E. McAllister, and J.R. Stauffer, Jr. Atlas of North American Freshwater Fishes. North Carolina State Museum of Natural History, Raleigh. pp. 75.

Gilbert, C.R., and R.M. Bailey. 1962. Synonymy, characters, and distribution of the American cyprinid fish Notropis shumardi. Copeia 1962:807-819.

Hampton, D.R., and C.R. Berry, Jr. 1997. Fishes of the mainstem Cheyenne River in South Dakota. Proceedings of the South Dakota Academy of Sci-ence 76:11-25.

Harlan, J.R., and E.B. Speaker. 1956. Iowa fish and fishing, third edition. Iowa State Conservation Commission, Des Moines.

Harland, B.C. 2003. Survey of the fishes and habitat of western South Dakota streams. Master’s thesis, South Dakota State University, Brookings. 167 p.

Harland, B., and C.R. Berry, Jr. 2004. Fishes and habitat characteristics of the Keya Paha River, South Dakota-Nebraska. Journal of Freshwater Ecology 19:169-177.

Hayer, C.A., S.S. Wall, and C.R. Berry. 2006. Evaluation of aquatic GAP analysis fish distribution models with emphasis on rare fish species in South Dakota. South Dakota State University, Brookings.


Hesse, L.W., G.E. Mestl, and J.W. Robinson. 1993. Status of selected fishes in the Missouri River in Nebraska with recommendations for their recovery. In: L.W. Hesse, C.B. Stalnaker, N.G. Benson, and J.R. Zuboy, editors. Res-toration planning for the rivers of the Mississippi River ecosystem. Biologi-cal Report 19, U.S. Department of the Interior, National Biological Survey, Washington, D.C. pp. 327-340.

Hesse, L.W., J.C. Schmulbach, J.M. Carr, K.D. Keenlyne, D.G. Unkenholz, J.W. Robinson, and G.E. Mestl. 1989. Missouri River fishery resources in relation to past, present, and future stresses. Pages 352-371 in D.P. Dodge, editor, Proceedings of the international Large River Symposium (LARS), Canadian Special Publication of Fisheries and Aquatic Sciences 106, Depart-ment of Fisheries and Oceans, Ottawa.

Hoagstrom, C.W. 2006. Fish community assembly in the Missouri River basin. Ph.D. Dissertation. South Dakota State University, Brookings. 329 pp.

Hoagstrom, C.W. and Berry, C.R., Jr. 2006. Island biogeography of native fish faunas among Great Plains drainage basins: basin scale features influence composition. American Fisheries Society Symposium 48:221-264.

Hubbs, C.L. 1951. The American cyprinid fish Notropis germanus Hay inter-preted as an intergeneric hybrid. American Midland Naturalist 45:446-454.

Isaak, D.J., W.A. Hubert, and C.R. Berry, Jr. 2003. Conservation assessment for lake chub, mountain sucker, and finescale dace in the Black Hills Nation-al Forest, South Dakota and Wyoming. U.S. Department of Agriculture, Forest Service, Custer, South Dakota.

Johnson, B., J. Lott, and W. Nelson-Stastny. 1998. Annual fish population and sport fish harvest surveys on Lake Oahe, South Dakota 1997. Annual Report No. 98-8. South Dakota Department of Game, Fish and Parks, Pierre.

Johnson, B., J. Lott, and J. Riis. 1995. Annual fish population and sport fish harvest surveys on Lake Oahe, South Dakota 1994. Annual Report No. 95-7. South Dakota Department of Game, Fish and Parks, Pierre.

Kazmierski, J.A. 1966. Composition and seasonal changes of the Macrofauna of an intermittent prairie stream. M.S. Thesis. University of South Dakota, Vermillion. 71 pp.

Kelsch, S.W. 1994. Lotic fish-community structure following transition from severe drought to high discharge. Journal of Freshwater Ecology 9:331-341.

Kirby, D.J. 2001. An assessment of the channel catfish population in the Big Sioux River, South Dakota. MS Thesis. Brookings: South Dakota State University. 110 pp.

Kral, J.G., and C.R. Berry, Jr. 2005. Fishes at randomly selected sites on wade-able streams in South Dakota. Proceedings of the South Dakota Academy of Science 84:305-313.

Kreiser, B.R., J.B. Mitton, and J.D. Woodling. 2001. Phylogeography of the plains killifish, Fundulus zebrinus. Evolution 55:339-350.


Lee, D.S. 1978. Anguilla rostrata (Lesueur), American eel. In D.S. Lee, C.R. Gilbert, C.H. Hocutt, R.E. Jenkins, D.E. McAllister, and J.R. Stauffer, Jr. Atlas of North American Freshwater Fishes. North Carolina State Museum of Natural History, Raleigh. pp. 59.

Lee, D.S., and S.P. Platania. 1978. Carpiodes velifer (Rafinesque), highfin carpsucker. In D.S. Lee, C.R. Gilbert, C.H. Hocutt, R.E. Jenkins, D.E. McAllister, and J.R. Stauffer, Jr. Atlas of North American Freshwater Fishes. North Carolina State Museum of Natural History, Raleigh. pp. 369.

Lehtinen, S.F., and J.B. Layzer. 1988. Reproductive cycle of the plains min-now Hybognathus placitus (Cyprinidae), in the Cimarron River, Oklahoma. Southwestern Naturalist 33:27-33.

Lein, G.M., and D.R. DeVries. 1998. Paddlefish in the Alabama River drain-age: population characteristics and the adult spawning migration. Transac-tions of the American Fisheries Society 127:441-454.

Lott, J., D. Fielder, B. Johnson, J. Riis, C. Stone, and G. Wickstrom. 1994. Annual fish population surveys on South Dakota Missouri River reservoirs, 1993. Progress Report No. 94-8. South Dakota Department of Game, Fish and Parks, Pierre.

Loomis, T.M., C.R. Berry, Jr., and J. Erickson. 1999. The fishes of the upper Moreau River Basin. Prairie Naturalist 31:193-214.

Lott, J., W. Nelson-Stastny, and K. Potter. 2004. Annual fish population and angler use, harvest and preference surveys on Lake Oahe, South Dakota, 2003. Annual Report 04-22. South Dakota Department of Game, Fish and Parks, Pierre.

McKinney, M.L. 2002. Do human activities raise species richness? Contrasting patterns in United States plants and fishes. Global Ecology and Biogeogra-phy 11:343-348.

McPhail, J.D., and C.C. Lindsey. 1970. Freshwater fishes of northwestern Canada and Alaska. Bulletin 173. Fisheries Research Board of Canada, Ottawa.

Milewski, C.L. 2001. Local and systemic controls on fish and fish habitat in South Dakota rivers and streams: implications for management. Ph.D. dis-sertation, South Dakota State University, Brookings. 177 p.

Miller, R.R. 1955. An annotated list of the American cyprinodontid fishes of the genus Fundulus, with the description of Fundulus persimilis from Yu-catan. Occasional Papers of the Museum of Zoology No. 568. University of Michigan, Ann Arbor.

Miller, R.R., J.D. Williams, and J.E. Williams. 1989. Extinctions of North American fishes during the past century. Fisheries 14(6):22-38.

Moen, T. 1970. The occurrence of black buffalo, Ictiobus niger (Rafinesque), in Lake Mitchell, South Dakota. Proceedings of the South Dakota Academy of Sciences 49:42-45.

Morey, N.M., and C.R. Berry, Jr. 2004. New distributional record of the northern redbelly dace in the northern Great Plains. Prairie Naturalist 36:257-260.

Moyle, P.B., and J.E. Williams. 1990. Biodiversity loss in the temperate zone: decline of the native fish fauna of California. Conservation Biology 4:275-284.


Nelson, J.S., E.J. Crossman, H. Espinosa-Pérez, L.T. Findley, C.R. Gilbert, R.N. Lea, and J.D. Williams. 2004. Common and scientific names of fishes from the United States, Canada, and Mexico, sixth edition. American Fisheries Society Special Publication 29, Bethesda, Maryland.

Netsch, N.F., and A. Witt, Jr. 1962. Contributions to the life history of the longnose gar, (Lepisosteus osseus) in Missouri. Transactions of the American Fisheries Society 91:251-262.

Paukert, C.P., and W.L. Fisher. 2001. Spring movements of paddlefish in a prairie reservoir system. Journal of Freshwater Ecology 16:113-124.

Patton, T.M., F.J. Rahel, and W.A. Hubert. 1998. Using historical data to assess changes in Wyoming’s fish fauna. Conservation Biology 12:1120-1128.

Pflieger, W.L., and T.B. Grace. 1987. Changes in the fish fauna of the lower Missouri River, 1940-1983. In: Matthews, W.J. and Heins, D.C., editors. Community and evolutionary ecology of North American stream fishes. Norman, OK: University of Oklahoma Press. pp. 166-177.

Pister, E.P. 1999. Professional obligations in the conservation of fishes. Envi-ronmental biology of fishes 55:13-20.

Pittenger, J.S., and G. Schiffmiller. 1997. Ichthyofaunal decline in Ute Creek, South Canadian River drainage, New Mexico. Southwestern Naturalist 42:112-117.

Platania, S.P., and C.S. Altenbach. 1998. Reproductive strategies and egg types of seven Rio Grande basin cyprinids. Copeia 1998:559-569.

Quist, M.C., Hubert, W.A. and Rahel, F.J. 2004. Relations among habitat characteristics, exotic species, and turbid-river cyprinids in the Missouri River drainage of Wyoming. Transactions of the American Fisheries Society 133: 727-742.

Ricciardi, A., and J.B. Rasmussen. 1999. Extinction rates of North American freshwater fauna. Conservation Biology 13:1220-1222.

Scheurer, J.A., Fausch, K.D. and Bestgen, K.R. 2003. Multiscale processes regulate brassy minnow persistence in a Great Plains river. Transactions of the American Fisheries Society 132: 840-855.

Schmulbach, J.C., and P.J. Braaten. 1993. The Vermillion River: neither red nor dead. In L.W. Hesse, C.B. Stalnaker, N.G. Benson, and J.R. Zuboy, editors. Proceedings of the symposium on restoration planning for rivers of the Mississippi River ecosystem. National Biological Survey Report 19, U.S. Government Printing Office, Washington, D. C. pp. 57-69.

Scott, W.B., and E.J. Crossman. 1973. Freshwater fishes of Canada. Bulletin 184. Fisheries Research Board of Canada, Ottawa.

Shearer, J.S., and C.R. Berry, Jr. 2003. Fish community persistence in eastern North and South Dakota rivers. Great Plains Research 13:139-159.

Shuman, D.A., R.A. Klumb, and S.T. McAlpin. 2005. Pallid sturgeon popula-tion assessment and associated fish community monitoring for the Missouri River: segments 5 and 6. U.S. Army Corps of Engineers, Northwest Divi-sion Kansas City and Omaha Districts, Yankton, South Dakota.

Shute, J.R. 1978. Ictiobus niger (Rafinesque), black buffalo. In D.S. Lee, C.R. Gilbert, C.H. Hocutt, R.E. Jenkins, D.E. McAllister, and J.R. Stauffer, Jr. Atlas of North American Freshwater Fishes. North Carolina State Museum of Natural History, Raleigh. pp. 406.


Sinning, J.A. 1968. Fishes of the Big Sioux River. MS Thesis. Brookings: South Dakota State University. 74 pp.

Smith, P.W. 1979. The fishes of Illinois. University of Illinois Press, Urbana.SDGFP (South Dakota Department of Game, Fish and Parks). 2006. Threat-

ened, endangered and candidate species of South Dakota. South Dakota Department of Game Fish and Parks, Pierre.

Springman, D.J., and R.L. Banks. 2005. Range extension of the southern red-belly dace into South Dakota. Prairie Naturalist 37:175-176.

Sylvester, R.M. 2004. Upper Missouri River Basin aquatic GAP fish distribu-tion model accuracy assessment and white sucker, Catostomus commersonii, population characteristics in the upper Missouri River basin. MS Thesis. Brookings: South Dakota State University. 180 pp.

Trautman, M.B. 1981. The fishes of Ohio with illustrated keys, revised edition. Columbus, OH: Ohio State University Press.

Underhill, J.C. 1957. The distribution of Minnesota minnows and darters in relation to Pleistocene glaciation. Occasional Papers No. 7. Minnesota Mu-seum of Natural History, University of Minnesota Press, Minneapolis.

USGS (U.S. Geological Survey). 2001. Environmental monitoring and as-sessment program activities in South Dakota, annual progress report. U.S. Geological Survey, Rapid City, SD.




Wall, S.S., and C.R. Berry, Jr. 2004. Road culverts across streams with the en-dangered Topeka shiner, Notropis topeka, in the James, Vermillion, and Big Sioux river basins. Proceedings of the South Dakota Academy of Science 83:125-135.

Wall, S.S., C.R. Berry, Jr., C.M. Blausey, J.A. Jenks, and C.J. Kopplin. 2004. Fish-habitat modeling for gap analysis to conserve the endangered Topeka shiner (Notropis topeka). Canadian Journal of Fisheries and Aquatic Sciences 61:954-973.

Welker, T.L., and D.L. Scarnecchia. 2004. Habitat use and population structure of four native minnows (family Cyprinidae) in the upper Missouri and lower Yellowstone rivers, North Dakota. Ecology of Freshwater Fish 13:8-22.

Wickstrom, G. 1997. Annual fish population survey of Lewis and Clark Lake, 1996. Annual Report No. 97-16. South Dakota Department of Game, Fish and Parks, Pierre.

Wickstrom, G. 2004. Annual fish population survey of Lewis and Clark Lake, 2003. Annual Report No. 09-09. South Dakota Department of Game, Fish and Parks, Pierre.


Wilcove, D., M. Bean, and P.C. Lee. 1992. Fisheries management and bio-logical diversity: problems and opportunities. In R.E. McCabe, editor. Transactions of the fifty-seventh North American Wildlife and Natural Resources Conference. Wildlife Management Institute, Washington, D. C. pp. 373-383.

Winston, M.R., C.M. Taylor, and J. Pigg. 1991. Upstream extirpation of four minnow species due to damming of a prairie stream. Transactions of the American Fisheries Society 120:98-105.

Wood, R.M., R.L. Mayden, R.H. Matson, B.R. Kuhajda, and S.R. Layman. 2002. Systematics and biogeography of the Notropis rubellus species group (Teleostei: Cyprinidae). Bulletin of the Alabama Museum of Natural His-tory 22:37-80.


PERENNIAL-WARMWATER FISH COMMUNITIESOF THE CHEYENNE RIVER DRAINAGE:

A SEASONAL ASSESSMENT

Christopher W. Hoagstrom, Austin C. DeWitte and Nathan J.C. GoschDepartment of Wildlife and Fisheries Sciences


South Dakota Cooperative Fish and Wildlife Research Unit


ABSTRACT

We surveyed fish faunas of perennial-warmwater streams of the Cheyenne River drainage, South Dakota, downstream from Angostura Dam and the Belle Fourche Diversion Dam to describe each fish fauna, including the population structure of characteristic fish species, and to summarize broad-scale spatial and temporal trends in fish distributions. We selected our sample stations based on previous data on habitat conditions and fish distributions and sampled five stations monthly from May through October 2004. We determined the char-acteristic fish species of each sampling station based on five criteria: persistence, dominance, habitat association, population structure, and tagged recaptures. There was much variation in species composition among streams. Of 24 char-acteristic species, only stonecat was characteristic of all five stations and only longnose dace was characteristic of four stations. Nine species were characteristic of only one station. This indicates a strong relation between stream type and fish faunal composition. We also documented range contractions of large-river fishes (western silvery minnow, plains minnow, sturgeon chub, flathead chub), which were absent from the Upper Cheyenne River, presumably due to drought and perhaps exacerbated by Angostura Dam upstream. In contrast, plains topmin-now and smallmouth bass increased in the Upper Cheyenne River, presumably because they were favored by stable flows. The Cheyenne River drainage down-stream from Angostura Dam and the Belle Fourche Diversion Dam in South Dakota has a diverse fish community that is a product of habitat diversity and the absence of physical barriers to dispersal, which allows fishes to respond to changing conditions and locate suitable habitats via dispersal.

Keywords

Fish distributions, population structure, faunal turnover, seasonal change, recruitment


INTRODUCTION

In South Dakota, the distribution of stream fishes is being increasingly well documented (e.g., Hampton and Berry 1997, Doorenbos 1998, Duehr 2004), as is the pattern of species change (or lack thereof ) over time, as conditions change (Schmulbach and Braaten 1993, Dieterman and Berry 1994, 1998, Shearer and Berry 2003). However, these studies rely upon data from broad surveys that, in most cases, visited each sample location once. Aside from catch per effort data, these studies usually provide limited information regarding fish populations and they do not incorporate seasonal changes. This is largely because detailed population studies and seasonal (rather than single visit) surveys require relatively high effort. However, as a result, little is known about the population structure of most South Dakota stream fishes. There have been very few detailed studies of stream fish faunas in South Dakota. A notable exception is Kazmierski (1966), who studied the fishes of Say Brook of the Vermillion River drainage between April and October 1965 and, in so doing, provided much more information for that stream and time than is available for any other South Dakota stream. One obvious advantage that Ka-zmierski had was that he studied a very small stream that was easy to characterize. It would be beneficial to have detailed information on fish faunas of larger stream systems, but it is difficult to adequately sample such systems intensively due to their size. In this study, we characterize fish faunas of perennial warmwater streams within the Cheyenne River drainage of South Dakota. Our goal was to provide a detailed summary of fish populations that incorporated seasonal variations and corresponded to differences in habitat conditions. We used previous informa-tion on habitat conditions and fish distributions (summarized by Duehr 2004) to select a small number of sample stations that would allow us to sample in-tensively and yet make generalizations about patterns throughout the study area. Our goal was to accurately represent each fish fauna, including the population structure of each characteristic fish species, and to summarize broad-scale spatial and temporal trends in fish distributions.

METHODS

The Cheyenne River watershed lies entirely within the Great Plains and has an area of 65,398 km2, primarily in Wyoming and South Dakota, U.S.A. There are two main forks of the Cheyenne River, the Belle Fourche River and Upper Cheyenne River. Both begin in highlands of the western Powder River Basin in northeastern Wyoming and they encircle the Black Hills near the South Dakota–Wyoming border with the Belle Fourche River to the north and Upper Cheyenne River to the south. Downstream, the forks join to form the Lower Cheyenne River. Major dams and reservoirs were constructed on both forks near the Black Hills and additional dams are present on tributary streams within the Black Hills, but there are no major dams downstream from the Black Hills (Fig-ure 1). Thus, Black Hills streams, the two forks, and the Lower Cheyenne River


constitute two undammed stream segments. Both segments are approximately 360 km in length. In a recent study, Duehr (2004) grouped streams of the Cheyenne River drainage into four major categories. We studied the major stream types that were present along both undammed stream segments of the Cheyenne River drainage. The north segment included Whitewood Creek, the Belle Fourche River, and the Lower Cheyenne River. The south segment included Beaver Creek (near Buf-falo Gap, South Dakota), the Upper Cheyenne River, and the Lower Cheyenne River. According to the classification of Duehr (2004) Whitewood Creek and Beaver Creek are ‘steep streams’, the Upper Cheyenne River is a ‘flat stream’, the Belle Fourche River is a ‘small river’, and the Lower Cheyenne River is a ‘large river’. Whitewood Creek is the first major tributary that joins the Belle Fourche River below the Belle Fourche Diversion Dam near Belle Fourche, South Da-kota. We sampled Whitewood Creek downstream from Whitewood, South Da-kota, where it had a width of less than 3 m and maximum depth of less than 2 m. Habitat was primarily a series of riffle-pool sequences and the presence of pools was facilitated by scour below inactive beaver dams. Beaver Creek is the second major tributary that joins the mainstem Upper Cheyenne River below Angostura

Figure 1. Map of the Cheyenne River watershed of South Dakota, USA with the 48 contiguous United States inset. Circles indicate fish sampling stations and triangles indicate U.S. Geological Survey gaging stations. Letters correspond to dams and reservoirs: A = Belle Fourche Diversion Dam, B = Orman Dam, C = Deerfield Dam, D = Pactola Dam, E = Sheridan Dam, F = Stockade Dam, G = Angostura Dam.


Dam, near Hot Springs, South Dakota. It is also depleted by upstream diver-sions (Burr et al. 1999). We sampled Beaver Creek near Buffalo Gap, South Dakota, where it had riffle-pool habitat, a width of less than 3 m, and maximum depth of less than 2 m. We sampled the Upper Cheyenne River east of Buffalo Gap, South Dakota. The habitat was a series of riffle-pool sequences, river width was roughly 5 to 8 m, and maximum depth was less than 2 m. We sampled the Belle Fourche River near Elm Springs, South Dakota. The habitat was riffle-run-pool, river width was 5 to 8 m, and maximum depth was roughly 2 m. Finally, we sampled the Lower Cheyenne River near Howes, South Dakota. The habitat was run-riffle-pool, river width was roughly 10 m, and maximum depth was less than 2 m, except during high flows in May. Our study was conducted during a drought period that began in 2001. Based on mean annual discharge in the Lower Cheyenne River (U.S. Geologi-cal Survey Gage 06438500), water years 2002 through 2004 were three of the eight driest years on record. During this drought period, mean discharge varied among the streams we studied based on an Analysis of Variance test of loge-transformed mean daily discharge data (F = 5013, df = 4, 5475, P < 0.001). Beaver Creek had lowest discharge, followed by Whitewood Creek, the Upper Cheyenne River, the Belle Fourche River, and the Lower Cheyenne River (Figure 2). Tukey HSD tests indicated all stream pairs had significantly different mean discharge (P < 0.001) except the Upper Cheyenne River and the Belle Fourche River were similar (P = 0.312). We sampled one station on each stream (Figure 1). Stream sites were se-lected to be representative of each stream type. Fish collections were carried

out monthly at all stations from May through October 2004. Fish sampling was conducted us-ing a mesohabitat approach in order to representatively sample fish species (Taylor et al. 1996; Vadas and Orth 1998). We vi-sually identified mesohabitats as regions with relatively uniform water depth and velocity (Jackson 1975) and sampled all available mesohabitats at each station. Monthly sampling was con-ducted over a 36-hour period at each station. During each monthly sample we conducted the maximum number of meso-habitat collections possible within the time period. We conducted night seining and electrofishing from May through July and de-termined that fish assemblage composition did not vary be-

Figure 2. Mean discharge with standard devia-tion (error bars) for each sampling station based on nearby U.S. Geological Survey gaging stations shown on Figure 1.


tween day and night (unpublished data). We used multiple capture methods to collect fishes, depending upon habitat conditions, because our goal was to ac-curately determine fish species composition (Smith and Hubert 1989; Bramblett and Fausch 1991; Quist et al. 2005b). Our primary capture method was a flat 3.0 mm mesh minnow seine, 3.7 m in length and fitted with weights every 15.2 cm along the lead line. A 3.0-mm mesh bag seine, 10.6 m in length with one weight every 30.4 cm was used in deep, sluggish pools. A backpack electrofisher (Smith-Root Model LR-24) was used to supplement seine collections from creek stations. Overnight sets of hoop nets with 25.4-mm mesh (bar measure), 0.8 m in diameter and 3.6 m in length, were used in pools and overnight sets of Gee minnow traps with 3.0-mm mesh were used along vegetated stream margins. Light fishing tackle (hook and line) with various live and artificial baits was used in deep, sluggish pools of river sta-tions. Finally, 20-hook trotlines (size 10/0 and 12/0 Circle Sea hooks) baited with chicken gizzards were used in deep, sluggish pools of river stations. Fish collection effort included at least four capture methods at each station (Table 1). All fishes were identified and measured to the nearest 1.0 mm standard length (SL). Except for voucher specimens, all fishes were released alive. We floy tagged all fishes that were at least 250 mm standard length so that we could identify recaptured individuals. We identified characteristic fish species of each stream using five criteria: (1) the species was persistent, being present in at least 5 of 6 monthly collections,

Table 1. Total sampling effort for each study station by gear type. Sampling gears included a 3.7-m long flat seine with 3.0-mm mesh and lead weights every 15.2 cm, a 10.6-m long bag seine with 6.0-mm mesh, a battery powered backpack electrofisher, 25.4-mm mesh bar measure mini hoop nets that were 3.6 m long with 0.8-m hoops, Gee minnow traps with 3.0-mm mesh, light fishing tackle with artificial lures and earthworms, and 20-hook trot lines, size 10/0 to 12/0 Circle Sea hooks, baited with chicken gizzards. Stations are ordered by mean discharge.

STUDY STATION

FLAT SEINE

(m2)

BAG SEINE

(m2)

ELEC-TRO-

FISHING (HOURS)

HOOP NET

(HOURS)

MIN-NOW TRAP

(HOURS)

FISHING POLE

(HOURS)

TROT LINE

(HOURS)

Beaver Creek 6,073.9 0.0 3.2 49.7 86.1 0.0 0.0

White-wood Creek

4,274.2 0.0 3.9 112.0 114.6 0.0 0.0

Upper Cheyenne River

9,376.3 7,466.8 0.0 299.0 364.7 52.4 57.0

BelleFourche River

3,732.8 2,428.5 0.0 415.3 498.9 43.5 165.0

Lower Cheyenne River

10,420.2 4,822.1 0.0 379.9 282.4 4.5 153.0


(2) the species was dominant within collections, representing at least 1% of all individuals collected, (3) habitat occupancy was predictable, that is, the presence of a species in a given habitat was consistent among sampling trips, (4) the popu-lation of each species included multiple length classes, indicating that multiple age-groups were present, and (5) tagged individuals were recaptured. We con-sidered a species to be characteristic of a station if it fit at least three of the five criteria. We calculated percent shared characteristic species among streams as the number of unshared species divided by the total number of species, multiplied by 100 (Russell 1998). We assembled a summary of the distribution and length structure of each characteristic species among our study streams. If character-istic species were represented by more than 50 individuals in multiple streams, we compared standard length structure using Analysis of Variance to compare means, with Tukey Honestly Significant Difference tests for post-hoc pairwise comparisons and we used two-sample Kolmogorov-Smirnoff tests to compare standard length distributions (Sokal and Rohlf 1995).


We collected a total of 18,690 fish of 28 species. White sucker Catostomus commersonii, stonecat Noturus flavus, and green sunfish Lepomis cyanellus were present in all five stations. Species richness increased with stream size (mean discharge, Figure 3). The number of characteristic species also increased with stream size (Tables 2– 6), except Whitewood Creek had fewer characteristic spe-cies (5) than Beaver Creek (7). High percent-unshared species (> 40%) among all streams indicated high species turnover (Table 7).

Figure 3. Fish species richness for each sampling station by mean discharge (Figure 2). Stations are (from left to right) Beaver Creek, Whitewood Creek, Upper Cheyenne River, Belle Fourche River, and Lower Cheyenne River.


Table 2. Characteristic fishes of Whitewood Creek near Whitewood, South Dakota. A total of 4,769 fish was collected. Five criteria were used to define characteristic fishes: persistence (per-cent presence > 80%), dominance (percent species composition > 1%), predictable habitat asso-ciation, presence of multiple length groups, recapture (recap.) of floy-tagged individuals (percent recapture rate). Species that fit three or more of the five criteria were considered characteristic inhabitants of Whitewood Creek. NA = not applicable (no fish tagged).

SPECIES

PRES-ENCE

(%)

DOMI-NANCE

(%) HABITATLENGTH-GROUPS

RECAP.(%)

Fathead minnowPimephales promelas 83 6 Pools 2 NA

Longnose dace Rhinichthyscataractae cataractae 100 41 All 2 NA

Creek chubSemotilus atromaculatus 100 49 All 3+ NA

White suckerCatostomus commersonii 100 1 Pools 3+ 4

StonecatNoturus flavus 100 < 1 Riffles 2 NA

Table 3. Characteristic fishes of Beaver Creek near Buffalo Gap, South Dakota. A total of 3,058 fish was collected. Five criteria were used to define characteristic fishes: persistence (percent presence > 80%), dominance (percent species composition > 1%), predictable habitat associa-tion, presence of multiple length groups, recapture (recap.) of floy-tagged individuals (percent recapture rate). Species that fit three or more of the five criteria were considered characteristic inhabitants of Beaver Creek. NA = not applicable (no fish tagged).

SPECIES

PRES-ENCE

(%)

DOMI-NANCE


RECAP.(%)

Longnose dace Rhinichthyscataractae cataractae 100 21 All 2+ NA

Creek chubSemotilus atromaculatus 100 60 All 3+ NA

White suckerCatostomus commersonii 100 13 Pools 4+ 16

Mountain suckerCatostomus platyrhynchus 83 1 All 2+ NA

StonecatNoturus flavus 100 1 All 2+ NA

Plains topminnowFundulus sciadicus 100 3 Pools 2 NA

Green sunfishLepomis cyanellus 100 2 Pools 3+ NA


Table 4. Characteristic fishes of the Upper Cheyenne River near Buffalo Gap, South Dakota. A total of 1,721 fish was collected. Five criteria were used to define characteristic fishes: persis-tence (percent presence > 80%), dominance (percent species composition > 1%), predictable habi-tat association, presence of multiple length groups, recapture (recap.) of floy-tagged individuals (percent recapture rate). Species that fit three or more of the five criteria were considered characteristic inhabitants of the Upper Cheyenne River. NA = not applicable (no fish tagged).

SPECIES

PRES-ENCE

(%)

DOMI-NANCE


RECAP.(%)

Goldeye Hiodon alosoides 100 3 Pools 1+ 10Red shinerCyprinella lutrensis lutrensis 100 20 Runs 2 NA

Common carp Cyprinus carpio 83 1 Pools 2 0Plains sand shiner Notropisstramineus missuriensis 100 25 Runs 2 NA

Shorthead redhorseMoxostoma macrolepidotum 100 22 All 3+ 0

Stonecat Noturus flavus 100 3 Riffles 2+ NAPlains topminnowFundulus sciadicus 100 5 Pools 2 NA

Smallmouth bassMicropterus dolomieu 100 18 All 3+ 9%

Table 5. Characteristic fishes of the Belle Fourche River near Elm Springs, South Dakota. A total of 5,609 fish was collected. Five criteria were used to define characteristic fishes: persistence (percent presence > 80%), dominance (percent species composition > 1%), predictable habitat association, presence of multiple length groups, recapture (recap.) of floy-tagged individuals (percent recapture rate). Species that fit three or more of the five criteria were considered char-acteristic inhabitants of the Belle Fourche River. NA = not applicable (no fish tagged).

SPECIES

PRES-ENCE

(%)

DOMI-NANCE


RECAP.(%)

Goldeye Hiodon alosoides 83 < 1 Pools 1 17Red shinerCyprinella lutrensis lutrensis 100 38 Runs 2 NA

Plains sand shiner Notropisstramineus missuriensis 100 33 Runs 2 NA

Flathead chubPlatygobio gracilis 100 5 Runs,

pools 2+ NA

Longnose dace Rhinichthyscataractae cataractae 100 4 Riffles 2 NA

Shorthead redhorseMoxostoma macrolepidotum 100 4 Runs,

Pools 3+ 0

Channel catfishIctalurus punctatus 100 9 Runs,

Pools 4+ 0

Stonecat Noturus flavus 100 2 Riffles 2+ NANorthern plains killifishFundulus kansae 100 1 Backwaters 2 NA


Table 6. Characteristic fishes of the Lower Cheyenne River near Four Corners, South Dakota. A total of 3,533 fish was collected. Five criteria were used to define characteristic fishes: persis-tence (percent presence > 80%), dominance (percent species composition > 1%), predictable habi-tat association, presence of multiple length groups, recapture (recap.) of floy-tagged individuals (percent recapture rate). Species that fit three or more of the five criteria were considered characteristic inhabitants of the Lower Cheyenne River. NA = not applicable (no fish tagged).

SPECIES

PRES-ENCE

(%)

DOMI-NANCE


RECAP.(%)

Red shinerCyprinella lutrensis lutrensis 100 3 Pools 1 NA

Plains minnowHybognathus placitus 100 5 Pools 2 NA

Sturgeon chubMacrhybopsis gelida 100 1 Riffles 2 NA

Plains sand shiner Notropis tramineus missuriensis 100 12 Runs 2 NA

Flathead chubPlatygobio gracilis 100 48 All 3 NA

Longnose dace Rhinichthyscataractae cataractae 100 10 Riffles 2+ NA

Northern river carpsuckerCarpiodes carpio carpio 100 1 Backwaters 2+ 0

Shorthead redhorseMoxostoma macrolepidotum 100 2 Pools 2+ 0

Channel catfish Ictalurus punctatus 100 14 Runs, pools 5+ 0

StonecatNoturus flavus 100 1 Riffles 2 NA

Table 7. Percent unshared characteristic species (unshared species / total species) among streams.

STREAM WHITEWOODUPPER

CHEYENNEBELLE

FOURCHELOWER

CHEYENNE

Beaver 50 85 86 87Whitewood 92 83 85Upper Cheyenne 58 71Belle Fourche 42


Characteristic Species Accounts

Goldeye Hiodon alosoides—The goldeye was a characteristic species of the Upper Cheyenne and Belle Fourche river stations (Tables 4 and 5), where it in-habited large pools. It was also present, but rare, in the Lower Cheyenne River station (n = 1, 339 mm SL). The goldeye was difficult to capture by seine or net, but was relatively easy to capture by hook and line (Hoagstrom 2006). Given that young-of-year goldeye are normally less than 170 mm total length (Harlan and Speaker 1951, Kennedy and Sprules 1967, Brown 1971, Smith 1979, Traut-man 1981, Becker 1983) all goldeye we collected were either juveniles or adults (Figure 4).

The goldeye is widespread throughout central North Amer-ica where it occupies large lakes and rivers (Scott and Crossman 1973, Trautman 1981, Becker 1983). It is widespread in the Cheyenne River drainage being present in large river and small stream habitats (Duehr 2004). The goldeye was the only char-acteristic fish species for which there was no sign of local spawn-ing and recruitment. No young-of-year were collected and nei-ther ripe females nor males were observed. Goldeye commonly undertake spawning migrations into tributary streams of lakes and large rivers (Scott and Crossman 1973, Eddy and Underhill 1974, Trautman 1981, Nelson and Paetz 1992). Lake Oahe goldeye pri-marily spawn in tributary rivers

and goldeye eggs have been collected as far as 170 km upstream in the Cheyenne River (Nelson 1980). Elsewhere after spawning, adults reportedly continue to move upstream and feed (Scott and Crossman 1973). Perhaps the goldeye of the Upper Cheyenne River and Belle Fourche River were adults that moved upstream to feed after spawning further downstream in the Cheyenne River drainage or else their eggs and larvae were displaced downstream (June 1977). Nelson (1980) reported relatively high densities of larval goldeye in 1972-1975 from the mouth of the Cheyenne River. Red shiner Cyprinella lutrensis lutrensis—The red shiner was a characteristic species of the Upper Cheyenne, Belle Fourche, and Lower Cheyenne river sta-tions (Tables 4 through 6) and used all habitat types. It was absent in Beaver and Whitewood creek stations. Multiple age groups were present in all three riv-ers. Males in breeding coloration and gravid females were present in the Upper Cheyenne and Belle Fourche rivers. Young-of-year and adult red shiners were

Figure 4. Length frequency histograms (3 mm cat-egories) for the goldeye at sampling stations where it was a characteristic species. Triangles indicate mean standard length with standard deviation (error bars).


present at each river station, but population structure appeared to vary by river size, with larger adults present in the Upper Cheyenne River (Figure 5). Mean SL varied significantly among rivers (F = 187.7, df = 2, 2601, P < 0.01) with significant pair-wise differences between all riv-er station pairs (PMD > |4.4|, P < 0.01 for all comparisons). Standard length distributions were also significantly different for all pairwise comparisons (D > 0.25, P < 0.01). The red shiner is widespread throughout south-central North America (Matthews 1987). It is one of the most widespread and abundant species in the Chey-enne River drainage (Newman et al. 1999, Duehr 2004). It is a highly tolerant generalist species that is most successful in medium-sized streams (Cross 1967, Baxter and Stone 1995). Absence of red shiner from Bea-ver and Whitewood creeks and prevalence in the medium-sized Upper Cheyenne and Belle Fourche rivers, support these generalizations as well as those of previous studies (Hampton and Berry 1997, Doorenbos 1998, Duehr 2004). The relatively small size dis-tribution in the Lower Cheyenne River suggests larger river conditions were less favorable for red shiners. Common carp Cyprinus carpio—The common carp was a characteristic spe-cies in the Upper Cheyenne River station (Table 4), where it occupied pools. It was also captured from the Belle Fourche (n = 15, mean SL = 103 mm ± 126.3 mm SD) and Lower Cheyenne (n = 30, mean SL = 98 mm ± 146.1 mm SD) river stations and large adults were observed but never captured from the Beaver Creek station. The Upper Cheyenne River population included young-of-year and adults (Figure 6). The common carp is not native to North America, but is widespread due to introductions and subsequent range expansions (Trautman 1981, Becker 1983). The common carp is widespread in the Cheyenne River drainage, but is not typi-cally abundant (Duehr 2004). Several factors may account for this. The species is commonly associated with organic pollution (Cross 1967, Trautman 1981), which is relatively uncommon in the Cheyenne River drainage, due to the low density of the human population. Also, common carp spawn in shallow water

Figure 5. Length frequency histograms (3 mm catego-ries) for the red shiner at sampling stations where it was a characteristic species. Triangles indicate mean standard length with standard deviation (error bars).


with emergent vegetation, which provides suitable nursery habitat for the young (Becker 1983), but it is uncommon in the Cheyenne River drainage. Plains minnow Hybognathus placitus—The plains minnow was a characteristic species in the Low-er Cheyenne River station (Table 6) where it occupied pools. It was absent from all other stations. The Lower Cheyenne River popula-tion included young-of-year and adults (Figure 7). On the night of 26 May 2004 (between 2300 and 0100 hours), we captured 15 plains minnow and numerous pe-lagic eggs similar to those pro-duced by plains minnow (Sliger 1967) from a pool in the Lower Cheyenne River. This suggests that spawning was occurring at our sampling station and the sub-sequent presence of young-of-year suggests recruitment occurred as well. The plains minnow ranges throughout the Great Plains (Al-Rawi and Cross 1964). It is typi-cal of large plains rivers (Cross 1967), but has also been collected from small tributary streams of the Cheyenne River drainage (Duehr

2004). It was widespread in the Upper Cheyenne River and Belle Fourche River during 1996 and 1997 (Hampton and Berry 1997, Doorenbos 1998), prior to the drought. This species has experienced widespread declines (Cross and Moss 1987, Pflieger and Grace 1987, Hesse et al. 1993, Patton et al. 1998) that are related to dewatering, habitat degradation, and population fragmentation caused by river impoundments (Cross and Moss 1987, Winston et al. 1991, Wilde and Ostrand 1999, Bonner and Wilde 2000, Quist et al. 2004b). Disappearance from the Upper Cheyenne and Belle Fourche Rivers between 1997 and 2004 suggests that drought conditions decreased habitat suitability. Sturgeon chub Macrhybopsis gelida— Like plains minnow, the sturgeon chub was a characteristic species in the Lower Cheyenne River station (Table 6), where it was restricted to riffle and run habitats. It was absent from all other stations. The Lower Cheyenne River population included young-of-year and adults (Figure 8). We collected a gravid female sturgeon chub on 19 July 2004,

Figure 6. Length frequency histogram (3 mm categories) for the common carp at the sampling station where it was a characteristic species. Triangles indicate mean standard length with standard deviation (error bars).

Figure 7. Length frequency histogram (3 mm cat-egories) for the plains minnow at the sampling sta-tion where it was a characteristic species. Triangles indicate mean standard length with standard de-viation (error bars).


which along with the presence of young-of-year, suggests spawn-ing and recruitment occurred at our sampling station. The sturgeon chub is known from the Missouri River drain-age and the lower Mississippi River (Werdon 1993). In the Missouri River drainage, it is restricted to rivers (Cross 1967, Brown 1971, Reigh and Elsen 1979). It was present in the Upper Cheyenne River dur-ing 1996 and 1997 (Hamp-ton and Berry 1997), indicating that drought conditions caused a range contraction, similar to the range contraction of plains minnow. The sturgeon chub has declined from 45% of its native range (Bicknell 2001), presumably due to impoundments that inundate river valleys, fragment populations, alter downstream flow regimes, de-water downstream river reaches, and eliminate downstream sediment transport (Hesse et al. 1993, Kelsch 1994, Everett et al. 2004, Quist et al. 2004b, Welker and Scarnecchia 2004). Sturgeon chub disappeared from the Little Missouri during a drought, probably because a mainstem Missouri River reservoir (Lake Sakakawea) isolated the population (Kelsch 1994). If so, the sturgeon chub of the Cheyenne River could be at risk because they are isolated by Lake Oahe and their distribution has been reduced during the recent drought. Plains sand shiner Notropis stramineus missuriensis—We analyzed sub-sam-ples of collection for our sample stations and voucher specimens from collections throughout the Cheyenne River drainage (Table 8) to determine whether sand shiners fit descriptions of the plains subspecies (Bailey and Allum 1962, Metcalf 1966, Tanyolaç 1973). Average measurements fell within the ranges reported for plains sand shiner (Figure 9), but there was substantial variation of each character and some individuals were within ranges reported for the eastern sand shiner Notropis stramineus stramineus for each character. The plains sand shiner was a characteristic species in the Upper Cheyenne, Belle Fourche, and Lower Cheyenne river stations (Tables 4-6) and it used pri-marily riffle and run habitats. It was not present in the Beaver or Whitewood creek stations. Multiple age groups and gravid females were present in all three rivers. Young-of-year and adult plains sand shiners were present at each river station, but population structure appeared to vary by river size, with larger adults present in the Upper Cheyenne River, but more young-of-year in the Belle Four-che River (Figure 10). Mean SL varied significantly among rivers (F = 70.7, df = 2, 2711, P < 0.01). It was significantly higher in the Upper Cheyenne River than in the Belle Fourche and Lower Cheyenne rivers (PMD > |5.5|, P < 0.01 for both comparisons), but similar between the Belle Fourche and Lower Cheyenne

Figure 8. Length frequency histogram (3 mm catego-ries) for the sturgeon chub at the sampling station where it was a characteristic species. Triangles indi-cate mean standard length with standard deviation (error bars).


rivers (PMD = -0.8, P = 0.27). Standard length distributions were significantly different for all pairwise comparisons (D > 0.14, P < 0.01). The plains sand shiner inhabits shallow streams and rivers of the Great Plains that have permanent flow (Cross 1967, Brown 1971). It is widespread and abundant throughout the Cheyenne River drainage (Duehr 2004). High prevalence in the Upper Cheyenne and Belle Fourche rivers compared to the Lower Cheyenne River is consistent with previous studies (Hampton and Berry 1997, Doorenbos 1998) and similar to the red shiner. Absence from Whitewood Creek is consistent with former studies (Newman et al. 1999), but Whitewood Creek was heavily polluted prior to the earliest fish surveys (Evermann and Cox 1896). Plains sand shiners were historically present in Beaver Creek (Bailey and Allum 1962).

Table 8. Summary of Notropis stramineus missouriensis specimens by location within the Cheyenne River drainage, South Dakota.

NO. LOCATION DATE N

Standard Length (mm)

MIN. MEDIAN MAX.

1 Indian Creek1, Mud Butte 7-23-96 11 41 49 532 Horse Creek1, Newell 8-5-96 7 41 50 553 Willow Creek1, Newell 8-12-96 6 43 54 584 Spring Creek1, Sturgis 7-28-98 5 48 68 735 Spring Creek4, Bear Butte Creek 6-3-04 62 38 51 616 Alkali Creek1, Hereford 8-11-98 10 40 45 517 East Elm Creek1, Union Center 8-12-98 1 588 Elm Creek4, Hereford 5-18-04 1 429 Belle Fourche River3 7-15-96 6 52 55 58

10 Belle Fourche River, Elm Springs 5-24-04 54 25 38 5911 Beaver Creek4, Burdock 7-15-04 6 51 53 5612 Cottonwood Creek4, Edgemont 6-28-04 2 36 4213 Fall River, 4 Hot Springs 6-28-04 5 54 57 6414 Cheyenne River, Buffalo Gap 5-22-04 48 34 41 5315 Battle Creek4, Hermosa 6-27-04 55 39 44 5016 Rapid Creek4, Farmingdale 6-22-04 61 40 53 6117 Cheyenne River or Rapid Creek2 8-1-94 8 40 48 6718 Elk Creek1, New Underwood 7-29-98 9 44 56 6619 Elk Creek4, Elm Springs 5-25-04 39 24 50 6020 Sulphur Creek4, Castle Rock 5-30-03 3 50 52 5321 Cheyenne River, Howes 5-26-04 46 25 37 50

1Collections by J. Erickson, South Dakota Game Fish and Parks2Collections by Cunningham et al. (1995)3Collections by Doorenbos (1998)4Collections by Duehr (2004)


Fathead minnow Pimephales prome-las—The fathead minnow was a charac-teristic species in the Whitewood Creek station (Table 2) where it occupied pools. It was also present in the Upper Cheyenne (n = 1, 64 mm SL), Belle Fourche (n = 41, mean SL = 35 mm ± 8.2 mm SD) and Lower Cheyenne (n = 1, 42 mm SL) river stations. Multiple length classes were present in White-wood Creek including young-of-year (Figure 11). The fathead minnow is widespread throughout North America (Trautman 1981, Becker 1983) and is abundant throughout the Great Plains (Cross 1967). It is common throughout the Cheyenne River drainage and most abundant in small streams (Newman et al. 1999, Duehr 2004). However, it has never been taken from Beaver Creek (Bailey and Allum 1962, Duehr 2004). The fathead minnow is a pioneering species and is tolerant of degraded habitat conditions including flow in-termittence and pollution (Cross 1967, Trautman 1981). It is also dispersed by humans as a common baitfish, is com-monly introduced to lakes and ponds as forage for sport-fish, and is sometimes stocked to fishless waters for mosquito control (Eddy and Underhill 1974, Trautman 1981, Becker 1983). Re-portedly, the fathead minnow is a poor competitor and is usually most abun-dant in streams where few other species are present (Hubbs and Cooper 1936, Cross 1967). It is also highly suscep-tible to predation (Hubbs and Cooper 1936, Becker 1983). Flathead chub Platygobio gracilis—The flathead chub was a characteristic species of the Belle Fourche and Lower Cheyenne river stations (Tables 5 and 6), using all habitat types. It was absent from all other stations. Both popula-tions included multiple length classes,

Figure 9. Mean and total range (error bars) of counts and measures from sand shiners of the Cheyenne River drainage (Table 8). The division between sand shiner subspecies is shown for each character.


including young-of-year, but there were more large adults in the Belle Fourche River (Fig-ure 12). Mean SL varied sig-nificantly between stations (F = 21.0, df = 1, 2007, P < 0.01). It was significantly higher in the Upper Cheyenne River than in the Belle Fourche and Lower Cheyenne rivers (PMD = 6.0, P < 0.01). Standard length distributions were also signifi-cantly different (D > 0.19, P < 0.01). The flathead chub is wide-spread across the Great Plains where it occupies permanent streams (Olund and Cross 1961, Brown 1971, Nelson and Paetz 1992). In the Cheyenne River drainage, it is most abun-dant in rivers, but is sometimes present in smaller streams (Bai-ley and Allum 1962, Duehr 2004). The species is unknown from Whitewood and Beaver creeks, but was formerly pres-ent throughout the Upper Cheyenne River (Bailey and Al-lum 1962, Hampton and Berry 1997). The flathead chub has declined throughout much of its range due to habitat modi-fications caused by dams and channelization (Cross and Moss 1987, Pflieger and Grace 1987, Hesse et al. 1993, Quist et al. 2004b, Welker and Scar-necchia 2004). The apparent range contraction we observed between 1997 and 2004 sug-gests that the Upper Cheyenne River was less suitable for the flathead chub during drought.

Longnose dace Rhinichthys cataractae cataractae—The longnose dace was a characteristics species in Beaver Creek, Whitewood Creek, the Belle Fourche River, and the Lower Cheyenne River stations (Tables 2, 3, 5, and 6), but was absent

Figure 10. Length frequency histograms (3 mm cat-egories) for the plains sand shiner at sampling stations where it was a characteristic species. Triangles indi-cate mean standard length with standard deviation (error bars).

Figure 11. Length frequency histogram (3 mm cat-egories) for the fathead minnow at the sampling station where it was a characteristic species. Triangles indicate mean standard length with standard devia-tion (error bars).


from the Upper Cheyenne River station. In the creeks, longnose dace used all habitats, but in the rivers, they were mostly re-stricted to riffles. All four popu-lations included young-of-year and adults, but the Whitewood Creek population appeared to have less length-class diversity compared to other stations (Fig-ure 13). Spawning coloration was observed at all four stations. Mean SL varied significantly be-tween stations (F = 120.7, df = 3, 3201, P < 0.01). It was sig-nificantly higher in Beaver Creek than all other stations (PMDs > |6.9|, P < 0.01), but mean SL of the Belle Fourche River station was similar to Whitewood Creek (PMD = -0.5, P = 0.89) and the Lower Cheyenne River (PMD = -2.0, P = 0.12). Mean SL in the Lower Cheyenne River was higher than in Whitewood Creek (PMD = -2.5, P < 0.01). Nevertheless, standard length distributions varied among all stations (D > 0.14, P < 0.01). The longnose dace is one of the most widely distributed fishes in North America and typically inhabits swift streams or turbulent habitats of larger rivers (e.g., riffles) and lakes (e.g., wave swept shorelines; Scott and Crossman 1973, Trautman 1981, Becker 1983). It is common through-out the Cheyenne River drainage and most abundant in steep streams (Newman et el. 1999, Duehr 2004). The longnose dace was present in the Upper Cheyenne River near our sam-pling station as recently as 1996 and 1997 (Hampton and Berry 1997). This suggests a range contraction that parallels apparent range contractions for plains minnow, sturgeon chub, and flathead chub and is presumably due to drought conditions.

Figure 12. Length frequency histograms (3 mm cat-egories) for the flathead chub at sampling stations where it was a characteristic species. Triangles indi-cate mean standard length with standard deviation (error bars).

Figure 13. Length frequency histograms (3 mm categories) for the longnose dace at sampling stations where it was a characteristic species. Triangles indicate mean standard length with standard deviation (error bars).


Creek chub Semotilus atromaculatus—The creek chub was a characteristic species in the Beaver Creek and Whitewood Creek stations (Tables 2 and 3), where it primarily occupied pools. Multiple length groups, including young-of-year, were present at each station. Young-of-year dominated the Whitewood Creek station whereas multiple length groups were important in Beaver Creek. As a result, mean SL was significantly higher in Beaver Creek (F = 1133.9, df = 1, 4145, P < 0.01; PMD = -25.3, P < 0.01) and standard length distributions varied between stations (D = 0.58, P < 0.01). The creek chub was also present, but rare, in the Upper Cheyenne (n = 5, mean SL = 49 mm ± 9.3 mm SD) and Belle Fourche (n = 3, mean SL = 58 mm ± 8.5 mm SD) river stations. The creek chub is widespread and common in small streams throughout eastern North America (Cross 1967, Scott and Crossman 1973, Trautman 1981, Becker 1983). It has a patchy but broad distribution in the Cheyenne River drainage (Newman et al. 1999, Duehr 2004). The creek chub is able to disperse rapidly and dispersing individuals are often present as strays in atypical habitats such as large rivers (Cross 1967) and lakes (Trautman 1981). This behavior pre-sumably accounts for the creek chubs we collected from our Upper Cheyenne and Belle Fourche river stations. The creek chub is highly tolerant of pollution (Ellis 1914, Becker 1983) and stream channelization (Trautman 1981).

Northern river carpsucker Carpiodes carpio carpio—The northern river carpsucker was a characteristic species in the Lower Cheyenne River station (Table 6). Multiple length classes were present but we pri-marily collected young-of-year (Figure 15), which were mostly restricted to slackwater habitats along shore. The northern river carpsucker was also present in the Upper Cheyenne (n = 13, mean SL = 297 mm ± 27.9 mm SD) and Belle Fourche (n = 25, mean SL = 132 mm ± 58.3 mm SD) river stations. The northern river carpsuck-er is widespread throughout south-central North America where it typically occupies rela-tively large streams and rivers (Trautman 1981, Becker 1983).

The species is migratory (Trautman 1981). Like the goldeye, it migrates into tributaries from Lake Oahe (Beckman and Elrod 1971), which may explain the low abundance of adults in our collections. The northern river carpsucker ranges throughout the Upper Cheyenne and Belle Fourche rivers (Hampton and Berry 1997, Doorenbos 1998). It also inhabits low gradient tributaries, but is

Figure 14. Length frequency histograms (3 mm categories) for the creek chub at sampling stations where it was a characteristic species. Triangles indi-cate mean standard length with standard deviation (error bars).


most prevalent in riverine habitats (Duehr 2004). Given that the northern river carpsucker favors riverine habitat, it is possible that drought conditions during our study limited the distribution of the species, similar to plains min-now, sturgeon chub, and flathead chub. The species has declined from some portions of its native range (Patton et al. 1998). White sucker Catostomus commersonii—The white sucker was a characteristic species in the Beaver Creek and Whitewood Creek stations (Tables 2 and 3), where it was mostly restricted to pools. In both creeks, it was present as multiple length classes, including young-of-year (Figure 16). Mean SL was similar between creeks (F = 1.2, df = 1, 567, P = 0.28; PMD = -6.4, P = 0.28), but standard length distributions varied between stations (D = 0.15, P < 0.01). The white sucker was also present, but rare, in the Upper Cheyenne (n = 15, mean SL = 73 mm ± 17.1 mm SD), Belle Fourche (n = 35, mean SL = 135 mm ± 55.4 mm SD), and Lower Cheyenne (n = 1, 138 mm SL) river stations. This dis-tribution was similar to the creek chub. The white sucker is wide-spread and abundant throughout northern North America, east of the Rocky Mountains (Scott and Crossman 1973, Trautman 1981, Becker 1983). It is widespread throughout the Cheyenne River drainage and is most abundant in smaller streams (Newman et al. 1999, Duehr 2004). Elsewhere, the species is migratory (Eddy and Underhill 1974, Trautman 1981, Becker 1983) and its presence in Lake Oahe (Lott et al. 2004) pres-ents the possibility that migra-tions occur in the Cheyenne River drainage. However, the rarity of white sucker in riverine habitats suggests that white sucker popula-tions are non-migratory and largely restricted to tributary streams, as they are in the southern Great Plains (Cross 1967, Sublette et al. 1990). In Lake Oahe, the

Figure 15. Length frequency histogram (3 mm categories) for the northern river carpsucker at the sampling station where it was a characteristic spe-cies. Triangles indicate mean standard length with standard deviation (error bars).

Figure 16. Length frequency histograms (3 mm cat-egories) for the white sucker at sampling stations where it was a characteristic species. Triangles indi-cate mean standard length with standard deviation (error bars).


white sucker is associated with small intermittent tributaries rather than larger tributaries such as the Cheyenne River (Beckman and Elrod 1971). The white sucker is highly tolerant of pollution (Ellis 1914, Trautman 1981, Becker 1983). Mountain sucker Catostomus platyrhynchus—The mountain sucker was a characteristic species in the Beaver Creek station (Table 3). It was present as

multiple length-classes including young-of-year (Figure 17) and was found primarily in pool habitats. The mountain sucker was also present in the Whitewood Creek station (n = 6, mean SL = 56 mm ± 13.3 mm SD), but was absent from riverine stations. The mountain sucker is present throughout the south-ern Rocky Mountains and typi-cally inhabit mountain streams and lakes (Scott and Crossman 1973, Baxter and Stone 1995). In the Cheyenne River drainage it is mostly restricted to streams of the Black Hills (Bailey and Allum 1962, Isaak et al. 2003). The spe-cies has declined from the Upper Cheyenne River below Angostura Dam (Isaak et al. 2003) and from streams elsewhere (Patton et al. 1998). Shorthead redhorse Moxos-toma macrolepidotum—The short-head redhorse was a characteristic species in the Upper Cheyenne, Belle Fourche, and Lower Chey-enne river stations where it occu-pied a variety of habitats (Tables 4-6), but the species was absent from creek stations. All popula-tions included multiple length-classes, but population structure was highly variable with young-of-year abundant only in the Upper Cheyenne River and the Lower Cheyenne River only containing relatively large individuals (Figure 18). The Belle Fourche River pop-ulation was intermediate between the other two. Mean SL varied

Figure 17. Length frequency histogram (3 mm categories) for the mountain sucker at the sam-pling station where it was a characteristic species. Triangles indicate mean standard length with standard deviation (error bars).

Figure 18. Length frequency histograms (3 mm categories) for the shorthead redhorse at sam-pling stations where it was a characteristic spe-cies. Triangles indicate mean standard length with standard deviation (error bars).


significantly between stations (F = 132.0, df = 2, 610, P < 0.01). It was highest in the Lower Cheyenne River followed by the Belle Fourche River and the Upper Cheyenne River (PMDs > |34.8|, P < 0.01) and standard length distributions varied among all stations (D > 0.62, P < 0.01). The shorthead redhorse has a wide range in east-central North America where it occupies larger streams and rivers and sometimes lakes (Scott and Crossman 1973, Becker 1983, Baxter and Stone 1995). It is widespread in the Cheyenne River drainage, being most abundant in riverine habitats, but present in smaller tributaries as well (Newman et al. 1999, Duehr 2004). The shorthead redhorse is a migratory species that spawns in clear streams (Smith and Hubert 1989, Nelson and Paetz 1992). Populations in Lake Oahe are associated with large tributaries such as the Cheyenne River, which they use for spawning (Beckman and Elrod 1971). Abundance of young-of-year in the Upper Cheyenne River suggested con-ditions there were better for spawning than in other river stations. Channel catfish Ictalurus punctatus—The channel catfish was a characteristic species in the Belle Fourche and Lower Cheyenne river stations (Tables 5 and 6), where it was mostly found in pools. Multiple length-classes including young-of-year were present at both stations and relatively small individuals (< 200 mm SL) dominated both populations (Figure 19). Mean SL varied significantly between stations (F = 14.0, df = 1, 1002, P < 0.01, PMD = -16.5, P < 0.01) and standard length distributions varied (D = 0.25, P < 0.01). The channel catfish was also present in the Upper Cheyenne River sta-tion (n = 21, mean SL = 314 mm ± 111.1 mm SD), but was absent from both creek stations. The channel catfish is wide-spread throughout southern North America where it occupies larger streams, rivers, and lakes (Trautman 1981, Becker 1983). It is widespread in the Cheyenne River drainage and is most abun-dant in riverine habitats (New-man et al. 1999, Duehr 2004). In 1996 and 1997, it was more abundant in the Belle Fourche River than the Upper and Lower Cheyenne rivers (Doorenbos et al. 1999). The channel catfish is highly migratory (Trautman 1981, Becker 1983). It is abundant in Lake Oahe (Doorenbos et al. 1999, Lott et al. 2004). The Cheyenne River drainage presum-ably provides important spawning habitat for the Lake Oahe population (June 1977), which is supported by the abundance of young-of-year and juveniles in our collections. Important spawning locations are unknown, but may be limited

Figure 19. Length frequency histograms (3 mm categories) for the channel catfish at sampling sta-tions where it was a characteristic species. Triangles indicate mean standard length with standard de-viation (error bars).


in distribution and, in the Great Plains, may be most prevalent in tributaries (Smith and Hubert 1989, Gerhardt and Hubert 1990). Stonecat Noturus flavus—The stonecat was the only species that was charac-teristic of the fish faunas of all five sampling stations (Tables 2-6). Multiple age classes, including young-of-year, were present at all stations (Figure 20). Like

the longnose dace, the stonecat occupied pools and riffles in the creeks but was mostly confined to riffles in the rivers (with a few exceptions). The species was particularly abundant in riffles of the Belle Fourche River station. Few large individu-als were collected from White-wood Creek. The stonecat is relatively difficult to collect and our samples likely underestimate their abundance relative to other fishes. The stonecat is widely dis-tributed in central North Amer-ica where it occupies permanent streams (Harlan and Speaker 1951, Trautman 1981, Becker 1983). It is present in all warm-water stream types of the Chey-enne River drainage (Newman et al. 1999, Duehr 2004) but is most abundant in large river habitats. The stonecat is toler-ant of pollution (Becker 1983), yet it has declined in some por-tions of its range due to habitat degradation, especially siltation (Eddy and Underhill 1974, Trautman 1981) and drought (Cross 1967). Northern plains killifish Fundulus kansae—The northern plains killifish was a characteristic species in the Belle Fourche River station (Table 5) where it occu-pied shallow-slackwater habitat. It was present in multiple length classes including young-of-year

(Figure 21). The northern plains killifish was also present in the Lower Cheyenne River station (n = 13, mean SL = 38 mm ± 8.7 mm SD), but was absent from all other stations.

Figure 20. Length frequency histograms (3 mm cat-egories) for the stonecat at sampling stations where it was a characteristic species. Triangles indicate mean standard length with standard deviation (er-ror bars).


The northern plains killifish is native to the central Great Plains and Ozark Plateau (Kreiser 2001, Kreiser et al. 2001). Many consider the species to be nonna-tive in South Dakota (e.g., Bailey and Allum 1962, Kreiser et al. 2001), but we consider them na-tive because there is no evidence of their introduction and their native presence in the state is consistent with the biogeography of the Missouri River drainage (Hoagstrom 2006). The spe-cies is widespread in the Chey-enne River drainage where it is most prevalent in smaller streams (Newman et al. 1999, Duehr 2004). The northern plains killifish characteristi-cally inhabits shallow habitats with relatively warm temperatures and is tolerant of high salinity (Griffith 1974). Northern plains killifish distribution and abun-dance may be highly variable (Fausch and Bestgen 1997). The northern plains killifish has declined from some portions of its native range (Cross and Collins 1995). Historically, it was present in the Upper Cheyenne River, downstream from Angostura Dam (Bailey and Allum 1962, Hampton and Berry 1997). It remains abundant upstream of Angostura Reservoir (Duehr 2004). Plains topminnow Fundulus sciadicus—The plains topminnow was a characteristic species of the Beaver Creek and Upper Cheyenne River stations (Tables 3 and 4) where it was found in shallow pools or along shallow shorelines with slackwa-ter habitat, often where algae or emergent vegetation were pres-ent. The species was present at both stations as multiple length-classes including young-of-year (Figure 22). Mean SL varied significantly between stations (F = 4.0, df = 1, 176, P = 0.05, PMD = -2.5, P = 0.05) and stan-dard length distributions varied (D = 0.30, P < 0.01). The plains topminnow was absent from all other sampling stations. The native range of the plains topminnow includes the central Great Plains and Ozark Plateau (Cross et al. 1986). The species

Figure 21. Length frequency histogram (3 mm cat-egories) for the plains killifish at the sampling sta-tion where it was a characteristic species. Triangles indicate mean standard length with standard devia-tion (error bars).

Figure 22. Length frequency histograms (3 mm cat-egories) for the plains topminnow at sampling sta-tions where it was a characteristic species. Triangles indicate mean standard length with standard devia-tion (error bars).


is restricted to upstream portions of the Cheyenne River drainage, where it is widespread, and is absent from riverine habitats, being restricted to smaller streams (Duehr 2004). The plains topminnow was unknown from Beaver Creek until our survey (Bailey and Allum 1962) and first collected from the Upper Cheyenne River in 1996-1997 (Hampton and Berry 1997). Elsewhere, it has declined from portions of its native range (Harlan and Speaker 1956, Patton et al. 1998). Green sunfish Lepomis cyanellus—The green sunfish was a characteristic species of the Beaver Creek station (Table 3). It was present as multiple length-

classes (Figure 23) and was pri-marily found in pools. The green sunfish was also present in the Whitewood Creek (n = 9, mean SL = 58 mm ± 25.5 mm SD), Up-per Cheyenne (n = 12, mean SL = 71 mm ± 13.2 mm SD), Belle Fourche (n = 4, mean SL = 45 mm ± 3.1 mm SD), and Lower Cheyenne (n = 1, 42 mm SL) sta-tions. The green sunfish is a wide-spread species of south-central North America and most typically occupies small streams, ponds, and lakes (Trautman 1981, Becker 1983). It is a pioneering species that tolerates flow intermittence

(Cross 1967). It is widespread in the Cheyenne River drainage and is most preva-lent in small streams (Newman et al. 1999, Duehr 2004). The green sunfish is tolerant of pollution and habitat degradation (Trautman 1981, Becker 1983). Smallmouth bass Micropterus dolomieu—The smallmouth bass was a char-acteristic species in the Upper Cheyenne River station (Table 4). It was present

as multiple length-classes includ-ing young-of-year (Figure 24) and occupied all habitats. It was ab-sent from all other stations. The smallmouth bass is wide-spread throughout east-central North America where it inhab-its streams and lakes (Trautman 1981, Becker 1983). The species is not native to the Cheyenne River drainage (Hoagstrom and Berry 2006). It commonly occu-pies cool streams with clean rocky substrate (Paragamian 1991). The smallmouth bass is sporadically distributed in upstream portions of the Cheyenne River drainage

Figure 23. Length frequency histogram (3 mm categories) for the green sunfish at the sampling station where it was a characteristic species. Trian-gles indicate mean standard length with standard deviation (error bars).

Figure 24. Length frequency histograms (3 mm categories) for the smallmouth bass at the sam-pling station where it was a characteristic species. Triangles indicate mean standard length with standard deviation (error bars).


(Duehr 2004) where it is associated with Angostura Reservoir on the Upper Cheyenne River and Orman Reservoir adjacent to the Belle Fourche River. The abundance of smallmouth bass in the Upper Cheyenne River downstream of An-gostura Dam increased between 1996-1997and 2004. We attribute this increase to stable flow conditions during the recent drought, when there have been no substantial releases from Angostura Dam (U.S. Geological Survey data). Else-where, smallmouth bass are relatively successful below dams (Paragamian 1991, Quist et al. 2005a) and low-discharge years favor smallmouth bass recruitment (Swenson et al. 2002).

Additional Species

Gizzard shad Dorosoma cepedianum—The gizzard shad was present as young-of-year in the Upper Cheyenne (n = 1, 64 mm SL), Belle Fourche (n = 6, mean SL = 30 mm ± 6.8 mm SD), and Lower Cheyenne (n = 8, mean SL = 39 mm ± 13.4 mm SD) river stations. These individuals were presumably strays from Angostura and Orman reservoirs. The gizzard shad was absent from 1996-1997 surveys of the Upper Cheyenne, Belle Fourche, and Lower Cheyenne rivers (Hampton and Berry 1997, Doorenbos 1998) and from recent surveys of tributary streams (Newman et al. 1999, Duehr 2004). It is not native to the Cheyenne River drainage (Hoagstrom and Berry 2006). Western silvery minnow Hybognathus argyritis—The western silvery min-now was present in the Belle Fourche (n = 32, mean SL = 107 mm ± 31.7 mm SD) and Lower Cheyenne (n = 18, mean SL = 74 mm ± 23.8 mm SD) river stations. This species is relatively widespread in the Cheyenne River drainage, being most abundant in riverine habitat (Duehr 2004). In 1996-1997 it was present throughout the Upper Cheyenne River. Its absence from our collections suggests that drought conditions have caused a range contraction. Black bullhead Ameiurus melas—The black bullhead was present, but rare, in the Belle Fourche (n = 2, mean SL = 74 mm ± 9.2 mm SD) and Lower Cheyenne (n = 1, 180 mm SL) river stations. It is widespread throughout the Cheyenne River drainage and most abundant in small streams (Duehr 2004). Brown trout Salmo trutta—The brown trout was present, but rare, in the Whitewood Creek station (n = 2, mean SL = 239 mm ± 76.4 mm SD). The individuals we captured were presumably strays from upstream, where habitat is more suitable for this species. The brown trout is not native to North America. Rock bass Ambloplites rupestris—The rock bass was present, but rare, in the Upper Cheyenne River station (n = 1, 185 mm SL). The source of this indi-vidual was presumably Angostura Reservoir, or the Cheyenne River further up-stream. The rock bass is not native to the Cheyenne River drainage (Hoagstrom and Berry 2006), where it is a rare inhabitant of streams (Duehr 2004). Orangespotted sunfish Lepomis humilis—The orangespotted sunfish was present in the Belle Fourche (n = 8, mean SL = 47 mm ± 7.7 mm SD) and Lower Cheyenne (n = 3, mean SL = 41 mm ± 3.2 mm SD) river stations. It is widely distributed in the Cheyenne River drainage, but is most abundant in small streams (Newman et al. 1999, Duehr 2004). Individuals we collected were presumably strays from populations in tributaries.


Bluegill Lepomis macrochirus macrochirus—The bluegill was present in the Upper Cheyenne River (n = 1, 98 mm SL), presumably as a stray from Angostura Reservoir or the river upstream. It is not native to the Cheyenne River drainage (Hoagstrom and Berry 2006) and is rarely present in streams therein (Duehr 2004). Sauger Sander canadensis—The sauger was present in the Belle Fourche (n = 3, mean SL = 364 mm ± 86.6 mm SD) and Lower Cheyenne (n = 4, mean SL = 334 mm ± 55.2 mm SD) river stations. Sauger in Lake Oahe congregate in the Cheyenne River arm and presumably migrate upstream to spawn (Beckman and Elrod 1971). Local ranchers reported that the sauger migrates up the Lower Cheyenne and Belle Fourche rivers in the fall, but we did not confirm it, perhaps due to drought.

Trends Among Species and Stations

High fish species turnover (Table 7) indicated that the faunas of each sampling station were unique. A summary of characteristic species by station supports this finding because, for example, nine species (common carp, plains minnow, stur-geon chub, fathead minnow, northern river carpsucker, mountain sucker, plains killifish, green sunfish, smallmouth bass) were characteristic of the fauna at only one station. Three of these species (plains minnow, sturgeon chub, smallmouth bass) were absent from all other stations, indicating strong segregation among streams types. However, trends in the distribution of characteristic species were variable. For example, the creek chub and white sucker were only characteristic in the Whitewood and Beaver creek stations, but other characteristic creek species (longnose dace and stonecat) were also characteristic of river stations. However, some species (red shiner, plains minnow, sturgeon chub, plains sand shiner, north-ern river carpsucker, shorthead redhorse, plains killifish, smallmouth bass) were restricted to river stations. In contrast, the plains topminnow was found only in the Beaver Creek and Upper Cheyenne River stations, which were geographically close and had high discharge stability (Hoagstrom 2006). Similarly, the flathead chub, channel catfish, and plains killifish were found only in the geographically close Belle Fourche River and Lower Cheyenne River stations, which also were relatively large streams that had low discharge stability (Hoagstrom 2006). In most cases, species that were characteristic of more than one station had variable length-class structure among them, indicating differing population status. Given these patterns, it appears that both habitat features and geographical proximity influence fish faunal composition and habitat diversity among stream types in the Cheyenne River drainage corresponds to fish species diversity. Even though the fish faunas of each sampling station were unique and persis-tent during our study, it is evident from historical collections that the fish faunas of each stream are dynamic. Some species collected in 1996-1997 from the Belle Fourche, Upper Cheyenne, and Lower Cheyenne Rivers (emerald shiner Notropis atherinoides, spottail shiner Notropis hudsonius, yellow bullhead Amei-urus natalis, northern pike Esox lucius, white bass Morone chrysops, largemouth bass Micropterus salmoides salmoides, black crappie Pomoxis nigromaculatus, yel-low perch Perca flavescens, freshwater drum Aplodinotus grunniens; Hampton and


Berry 1997, Doorenbos 1998) were absent from our collections. Most likely, these species were either strays from upstream reservoirs and stock ponds (yellow bullhead, northern pike, largemouth bass, black crappie, yellow perch) or were upstream migrants from Lake Oahe that may have found the Cheyenne River drainage unsuitable under drought conditions (emerald shiner, spottail shiner, white bass, freshwater drum). More specifically, the distribution of several fishes appeared to be reduced during our study than in 1996-1997, presumably due to drought. That is, large-river species that ranged near the Upper Cheyenne River or Belle Fourche River stations in 1996-1997 were absent, but still found in the Lower Cheyenne River station, suggesting reduced discharge caused a downstream retreat. The fish fauna of the Upper Cheyenne River station was the most dramatic example. Large-river species that were collected in the Upper Cheyenne River during 1996-1997 (western silvery minnow, plains minnow, sturgeon chub, flathead chub; Hampton and Berry 1997, Doorenbos 1998) were absent from our collec-tions, while the abundance of plains topminnow and smallmouth bass increased. This is likely related to Angostura Dam, which controls the flow regime, because it may amplify the impacts of drought by eliminating floods downstream. There were fewer cases of range contractions of large-river species from the Belle Four-che River station (plains minnow), perhaps because floods still occurred there. Nonetheless, range contractions of large-river fishes during drought are to be expected and have been documented elsewhere (Cross and Moss 1987). If con-ditions in the Lower Cheyenne River remain suitable for the large-river species, they may be able to recolonize upstream when the drought ends. However, the persistence of large-river species could be hampered by Lake Oahe, which limits the downstream retreat (formerly, species could have retreated into the mainstem Missouri River). The combination of downstream reservoirs and drought has been implicated in the disappearance of large-river fishes from upstream river reaches elsewhere (Winston et al. 1991, Kelsch 1994, Pittenger and Schiffmiller 1997, Luttrell et al. 1999, Wilde and Ostrand 1999). In summary, the un-dammed segments of the Cheyenne River drainage have dynamic fish communi-ties due to the ability of fishes to respond to changing environmental conditions and seek out suitable habitats, but downstream impacts of Angostura Dam and upstream impacts of Lake Oahe may be a threat to sensitive species, specifically, large-river fishes. The absence of physical barriers to fish movement between the Belle Fourche Diversion Dam, Angostura Dam, and Lake Oahe clearly enhances the fish fau-nas of the Cheyenne River drainage. Migratory species (goldeye, northern river carpsucker, shorthead redhorse, channel catfish, sauger) added to the diversity of the faunas and, in many cases, were characteristic species in our sampling sta-tions. At the same time, the ability of these species to use the Cheyenne River drainage presumably enhances their populations in Lake Oahe (Beckman and Elrod 1971, June 1977). Thus, it is appropriate to view the Cheyenne River drainage as an extension of Lake Oahe, and vice versa. In conclusion, different stream types we studied corresponded to different fish faunas. This is no surprise, given that the distributions of fishes in the Great Plains commonly correspond to habitat conditions (Ostrand and Wilde 2002,


Quist et al. 2004a, 2004b, Brunger Lipsey et al. 2005). Further, the importance of free dispersal for Great Plains stream fishes is well documented (Smith and Hubert 1989, Fausch and Bramblett 1991, Fausch and Bestgen 1997, Labbe and Fausch 2000, Scheurer and Fausch 2003). Thus, relatively large stream seg-ments and sub-watersheds that lack dispersal barriers (e.g., dams), but have high habitat diversity, are likely to sustain relatively diverse fish communities because as conditions change, which they often do in the Great Plains, fishes can respond by relocating as necessary, so long as some suitable habitat is present. At pres-ent, the Cheyenne River drainage downstream from Angostura Dam and the Belle Fourche Diversion Dam represents such a sub-watershed. The future of fish communities there will ultimately depend both on the interaction between climate change and human activities.

ACKNOWLEDGEMENTS

Federal Aid in Sport Fish Restoration funds administered by South Dakota Game, Fish and Parks (Project Number F-21-R and F-15-R) supported this research. The South Dakota Cooperative Fish and Wildlife Research Unit is jointly supported by the South Dakota Game Fish and Parks, U.S. Geological Survey, U.S. Fish and Wildlife Service, Wildlife Management Institute, and South Dakota State University. We thank J. Kral, M. Mangan, R. Sylvester, and R. Rasmus for field assistance. We also thank M. Barnes and J. Duehr for logis-tical support and J. Shearer for providing editorial comments. S. Wall prepared the map (Figure 1). This study was only possible through the generosity of five private landowners who allowed us access to their land.

LITERATURE CITED

Al-Rawi, A.H., and F.B. Cross. 1964. Variation in the plains minnow, Hybog-nathus placitus Girard. Transactions of the Kansas Academy of Science 67:154-168.

Bailey, R.M. and M.O. Allum. 1962. Fishes of South Dakota. Ann Arbor, MI: Miscellaneous Publications, Museum of Zoology, University of Michigan, No. 119.

Baxter, G.T., and M.D. Stone. 1995. Fishes of Wyoming. Wyoming Game and Fish Department, Cheyenne.

Becker, G.C. 1983. Fishes of Wisconsin. Madison, WI: University of Wis-consin Press.

Beckman, L.G., and J.H. Elrod. 1971. Apparent abundance and distribution of young-of-year fishes in Lake Oahe, 1965-69. In G.E. Hall, editor. Reservoir fisheries and limnology. American Fisheries Society, Washington, D.C. pp. 333-347.

Bicknell, W. 2001. Endangered and threatened wildlife and plants; 12-month finding for a petition to list the sicklefin chub (Macrhybopsis meeki) and the sturgeon chub (Macrhybopsis gelida) as endangered. Federal Register 66:19910-19914.


Bonner, T.H. and G.R. Wilde. 2000. Changes in the Canadian River fish assemblage associated with reservoir construction. Journal of Freshwater Ecology 15:189-198.

Bramblett, R.G. and K.D. Fausch. 1991. Fishes, macroinvertebrates, and aquatic habitats of the Purgatoire River in Piñon Canyon, Colorado. South-western Naturalist 36: 281-294.


Brunger Lipsey, T.S., W.A. Hubert and F.J. Rahel. 2005. Relationships of eleva-tion, channel slope, and stream width to occurrences of native fishes at the Great Plains-Rocky Mountains interface. Journal of Freshwater Ecology 20: 695-705.

Burr, M.J., R.W. Teller. and K.M. Neitzert. 1999. Water resources data, South Dakota, water year 1998. Washington, D.C.: U.S. Geological Survey, Wa-ter-Data Report SD-98-1.

Cross, F.B. 1967. Handbook of fishes of Kansas. Miscellaneous Publication No. 45. Museum of Natural History, University of Kansas, Lawrence.

Cross, F.B. and J.T. Collins. 1995. Fishes in Kansas, second edition, revised. Lawrence, KS: University Press of Kansas.

Cross, F.B. and R.E. Moss. 1987. Historic changes in fish communities and aquatic habitats in plains streams of Kansas. In: Matthews, W.J. and Heins, D.C., editors. Community and evolutionary ecology of North American stream fishes. Norman, OK: University of Oklahoma Press. pp. 155-165.

Cross, F.B., R.L. Mayden, and J.D. Stewart. 1986. Fishes in the western Mis-sissippi Basin. In C.H. Hocutt and E.O. Wiley, editors. The zoogeography of North American freshwater fishes. John Wiley and Sons, New York. pp. 363-412.

Dieterman, D.J., and C.R. Berry, Jr. 1994. Fishes in seven streams of the Min-nesota River drainage in northeastern South Dakota. Proceedings of the South Dakota Academy of Sciences 73:23-30.

Dieterman, D.J., and C.R. Berry, Jr. 1998. Fish community and water quality changes in the Big Sioux River. Prairie Naturalist 199-224.

Doorenbos, R.D. 1998. Fishes and habitat of the Belle Fourche River, South Dakota. MSc Thesis. Brookings: South Dakota State University. 79 pp.

Doorenbos, R.D., C.R. Berry, Jr., and G. Wickstrom. 1999. Ictalurids in South Dakota. American Fisheries Society Symposium 24:377-389.

Duehr, J.P. 2004. Fish and habitat relations at multiple spatial scales in Chey-enne River basin, South Dakota. MSc Thesis. Brookings: South Dakota State University. 79 pp.

Eddy S., and J.C. Underhill. 1974. Northern fishes, with special reference to the upper Mississippi Valley, third edition, revised and expanded. University of Minnesota Press, Minneapolis.

Ellis, M.M. 1914. Fishes of Colorado. University of Colorado Studies 11(1): 5-136.

Everett, S.R., D.L. Scarnecchia, and L.F. Ryckman. 2004. Distribution and habitat use of sturgeon chubs (Macrhybopsis gelida) and sicklefin chubs (M. meeki) in the Missouri and Yellowstone rivers, North Dakota. Hydrobio-logia 527:183-193.


Evermann, B.W., and U.O. Cox. 1896. Report upon the fishes of the Missouri River basin. In Report of the U.S. Commissioner of fish and fisheries for 1894. U.S. Government Printing Office, Washington, D.C. pp 325-429.

Fausch, K.D. and K.R. Bestgen. 1997. Ecology of fishes indigenous to the central and southwestern Great Plains. In: Knopf, F.L. and Samson, F.B., editors. Ecology and conservation of Great Plains vertebrates. New York, NY: Springer-Verlag, Ecological Studies 125. pp 131-166.

Fausch, K.D. and R.G. Bramblett. 1991. Disturbance and fish communities in intermittent tributaries of a western Great Plains river. Copeia 1991: 659-674.

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.

Griffith, R.W. 1974. Environment and salinity tolerance in the genus Fundulus. Copeia 1974:319-331.

Hampton, D.R., and C.R. Berry, Jr. 1997. Fishes of the mainstem Cheyenne River in South Dakota. Proceedings of the South Dakota Academy of Sci-ence 76:11-25.

Harlan, J.R., and E.B. Speaker. 1951. Iowa fish and fishing. Iowa State Con-servation Commission, Des Moines.

Harlan, J.R., and E.B. Speaker. 1956. Iowa fish and fishing, third edition. Iowa State Conservation Commission, Des Moines.

Hesse, L.W., G.E. Mestl, and J.W. Robinson. 1993. Status of selected fishes in the Missouri River in Nebraska with recommendations for their recovery. In: L.W. Hesse, C.B. Stalnaker, N.G. Benson, and J.R. Zuboy, editors. Res-toration planning for the rivers of the Mississippi River ecosystem. Biologi-cal Report 19, U.S. Department of the Interior, National Biological Survey, Washington, D.C. pp. 327-340.

Hoagstrom, C.W. 2006. Fish community assembly in the Missouri River basin. Ph.D. Dissertation. South Dakota State University, Brookings. 329 pp.

Hoagstrom, C.W. and C.R. Berry, Jr. 2006. Island biogeography of native fish faunas among Great Plains drainage basins: basin scale features influence composition. American Fisheries Society Symposium 48:221-264.

Hubbs, C.L., and G.P. Cooper. 1936. Minnows of Michigan. Bulletin No. 8. Cranbrook Institute of Science, Bloomfield Hills, Michigan.

Hubert, W.A. and F.J. Rahel. 1989. Relations of physical habitat to abundance of four nongame fishes in high-plains streams: a test of habitat suitability index models. North American Journal of Fisheries Management 9: 332-340.

Isaak, D.J., W.A. Hubert, and C.R. Berry, Jr. 2003. Conservation assessment for lake chub, mountain sucker, and finescale dace in the Black Hills Nation-al Forest, South Dakota and Wyoming. U.S. Department of Agriculture, Forest Service, Custer, South Dakota.

Jackson, R.G., II. 1975. Hierarchical attributes and a unifying model of bed forms composed of cohesionless material and produced by shearing flow. Bulletin of the Geological Society of America 86: 1511-1522.


June, F.C. 1977. Reproductive patterns in seventeen species of warmwater fishes in a Missouri River reservoir. Environmental Biology of Fishes 2:285-296.

Kazmierski, J.A. 1966. Composition and seasonal changes of the Macrofauna of an intermittent prairie stream. M.S. Thesis. University of South Dakota, Vermillion. 71 pp.

Kelsch, S.W. 1994. Lotic fish-community structure following transition from severe drought to high discharge. Journal of Freshwater Ecology 9:331-341.

Kennedy, W.A., and W.M. Sprules. 1967. Goldeye in Canada. Bulletin 161. Fisheries Research Board of Canada, Ottawa.

Kreiser, B.R. 2001. Mitochondrial cytochrome b sequences support recogni-tion of two cryptic species of plains killifish, Fundulus zebrinus and Fundulus kansae. American Midland Naturalist 146:199-209.

Kreiser, B.R., J.B. Mitton, and J.D. Woodling. 2001. Phylogeography of the plains killifish, Fundulus zebrinus. Evolution 55:339-350.

Labbe, T.R. and K.D. Fausch. 2000. Dynamics of intermittent stream habitat regulate persistence of a threatened fish at multiple scales. Ecological Ap-plications 10: 1774-1791.

Lott, J., W. Nelson-Stastny, and K. Potter. 2004. Annual fish population and angler use, harvest and preference surveys on Lake Oahe, South Dakota, 2003.

Luttrell, G.R., A.A. Echelle, W.L. Fisher, and D.J. Eisenhour. 1999. Declin-ing status of two species of the Macrhybopsis aestivalis complex (Teleostei: Cyprinidae) in the Arkansas River basin and related effects of reservoirs as barriers to dispersal. Copeia 1999: 981-989.

Matthews, W.J. 1987. Geographic variation in Cyprinella lutrensis (Pisces: Cyprinidae) in the United States, with notes on Cyprinella lepida. Copeia 1987:616-637.

Metcalf, A. L. 1966. Fishes of the Kansas River system in relation to zoogeog-raphy of the Great Plains. University of Kansas Publications, Museums of Natural History 17:23-189.

Nelson, J.S., and M.J. Paetz. 1992. The fishes of Albert, 2nd edition. Univer-sity of Alberta Press, Edmonton.

Nelson, W.R. 1980. Ecology of larval fishes in Lake Oahe, South Dakota. Technical Papers of the U.S. Fish and Wildlife Service 101.

Newman, R.L., C.R. Berry. and W. Duffy. 1999. A biological assessment of four northern Black Hills streams. Proceedings of the South Dakota Academy of Science 78: 185-197.

Olund, L.J., and F.B. Cross. 1961. Geographic variation in the North American cyprinid fish Hybopsis gracilis. University of Kansas Publications, Museum of Natural History 13:323-348.

Ostrand, K.G., and G.R. Wilde. 2002. Seasonal and spatial variation in a prai-rie stream-fish assemblage. Ecology of Freshwater Fish 11:137-149.

Paragamian, V.L. 1991. Stream sedimentation and abundance of smallmouth bass. In: Jackson, D.C. ed. The first international smallmouth bass sympo-sium. Mississippi State, MS: Mississippi Agricultural and Forestry Experi-ment Station, Mississippi State University. Pp. 55-60.


Patton, T.M., F.J. Rahel, and W.A. Hubert. 1998. Using historical data to assess changes in Wyoming’s fish fauna. Conservation Biology 12:1120-1128.

Pflieger, W.L., and T.B. Grace. 1987. Changes in the fish fauna of the lower Missouri River, 1940-1983. In: Matthews, W.J. and Heins, D.C., editors. Community and evolutionary ecology of North American stream fishes. Norman, OK: University of Oklahoma Press. pp. 166-177.

Pittenger, J.S., and G. Schiffmiller. 1997. Ichthyofaunal decline in Ute Creek, South Canadian River drainage, New Mexico. Southwestern Naturalist 42:112-117.

Quist, M.C., W.A. Hubert. and F.J. Rahel. 2004a. Elevation and stream size thresholds affect distributions of native and exotic warmwater fishes in Wyo-ming. Journal of Freshwater Ecology 19: 227-236.

Quist, M.C., W.A. Hubert. and F.J. Rahel. 2004b. Relations among habitat characteristics, exotic species, and turbid-river cyprinids in the Missouri River drainage of Wyoming. Transactions of the American Fisheries Society 133: 727-742.

Quist, M.C., W.A. Hubert. and F.J. Rahel. 2005a. Fish assemblage structure following impoundment of a Great Plains river. Western North American Naturalist 65: 53-63.

Quist, M.C., W.A. Hubert. and F.J. Rahel. 2005b. Hierarchical faunal filters: an approach to assessing effects of habitat and nonnative species on native fishes. Ecology of Freshwater Fish 14: 24-39.

Reigh, R.C., and D.S. Elsen. 1979. Status of the sturgeon chub (Hybopsis gelida) and sicklefin chub (Hybopsis meeki) in North Dakota. Prairie Natu-ralist 11:49-52.

Russell, G.J. 1998. Turnover dynamics across ecological and geological scales. Pages 377-404 in M.L. McKinney and J.A. Drake, editors, Biodiversity dynamics, turnover of populations, taxa, and communities. Columbia Uni-versity Press, New York.

Scheurer, J.A., K.D. Fausch. and K.R. Bestgen. 2003. Multiscale processes regulate brassy minnow persistence in a Great Plains river. Transactions of the American Fisheries Society 132: 840-855.

Schmulbach, J.C., and P.J. Braaten. 1993. The Vermillion River: neither red nor dead. In: Hesse, L.W., Stalnaker, C.B., Benson, N.G. and Zuboy, J.R., editors. Restoration planning for the rivers of the Mississippi River ecosystem. Washington, D.C.: U.S. Department of the Interior, National Biological Survey, Biological Report 19. pp. 57-69.

Scott, W.B., and E.J. Crossman. 1973. Freshwater fishes of Canada. Bulletin 184. Fisheries Research Board of Canada, Ottawa.


Smith, J.B. and W.A. Hubert. 1989. Use of a tributary by fishes in a Great Plains river system. Prairie Naturalist 21: 27-38.

Smith, P.W. 1979. The fishes of Illinois. University of Illinois Press, Urbana.Sokal, R.R., and F.J. Rohlf. 1995. Biometry, the principles and practice of sta-

tistics in biological research, third edition. W.H. Freeman and Company, New York.


Sublette, J.E., M.D. Hatch, and M. Sublette. 1990. The fishes of New Mexico. University of New Mexico Press, Albuquerque.

Swenson, W.A., B.J. Shuter, D.J. Orr, and G.D. Heberlin. 2002. The effects of stream temperature and velocity on first-year growth and year-class abun-dance of smallmouth bass in the upper Mississippi River. American Fisher-ies Society Symposium 31: 101-113.

Tanyolaç, J. 1973. Morphometric variation and life history of the cyprinid fish Notropis stramineus (Cope). Occasional Papers, Museum of Natural History, University of Kansas 12:1-28.

Taylor, C.M., M.R. Winston, and W.J. Matthews. 1996. Temporal variation in tributary and mainstem fish assemblages in a Great Plains stream system. Copeia 1996: 280-289.

Trautman, M.B. 1981. The fishes of Ohio with illustrated keys, revised edition. Columbus, OH: Ohio State University Press.

Vadas, R.L., Jr. and D.J. Orth. 1998. Use of physical variables to discriminate visually determined mesohabitat types in North American streams. Rivers 6: 143-159.

Welker, T.L., and D.L. Scarnecchia. 2004. Habitat use and population structure of four native minnows (family Cyprinidae) in the upper Missouri and lower Yellowstone rivers, North Dakota. Ecology of Freshwater Fish 13:8-22.

Werdon, S.J. 1993. Status report on sturgeon chub (Macrhybopsis gelida), a candidate endangered species. U.S. Fish and Wildlife Service, Bismarck, North Dakota.

Wilde, G.R., and K.G. Ostrand. 1999. Changes in the fish assemblage of an intermittent prairie stream upstream from a Texas impoundment. Texas Journal of Science 51:203-210.

Winston, M.R., C.M. Taylor, and J. Pigg. 1991. Upstream extirpation of four minnow species due to damming of a prairie stream. Transactions of the American Fisheries Society 120:98-105.


RECENT RANGE EXTENSIONS,NAME CHANGES AND STATUS UPDATESFOR SELECTED SOUTH DAKOTA FISHES

Cari-Ann HayerSouth Dakota State University

Brookings, SD 57006

Brandon C. HarlandIowa Department of Natural Resources

Des Moines, IA 50319


South Dakota Cooperative Research UnitSouth Dakota State University

Brookings, South Dakota 57007, USA

ABSTRACT

We present new distributional records for ten species of fish in South Dakota: silver chub, Macrhybopsis storeriana, Topeka shiner, Notropis topeka, northern redbelly dace, Phoxinus eos, southern redbelly dace, P. erythrogaster, shorthead redhorse, Moxostoma macrolepidotum, yellow bullhead, Ameiurus na-talis, northern pike, Esox lucius, Iowa darter, Etheostoma exile, johnny darter E. nigrum, yellow perch, Perca flavescens, and walleye, Sander vitreus. We also pres-ent several recently described or renamed fishes: shoal chub, Macrhybopsis hyos-toma, Carmine shiner, Notropis percobromus, western blacknose dace, Rhinichthys obtusus, and northern plains killifish, Fundulus kansae. The sand shiner, Notropis stramineus, is traditionally separated into two subspecies: eastern sand shiner, N. s. stramineus and plains sand shiner, N. s. missuriensis, both of which are present in South Dakota. There are also three species of carpsuckers in South Dakota: northern river carpsucker, Carpiodes carpio carpio, central quillback carpsucker, C. cyprinus hinei, and highfin carpsucker, C. velifer. Difficulties in distinguish-ing among these species obscure the status of the rarer quillback and highfin carpsuckers. Many species in South Dakota are easily misidentified; therefore we recommend preserving specimens for future examinations.

Keywords

Fish distribution, range extensions, misidentification, South Dakota


INTRODUCTION

The last comprehensive survey of South Dakota fishes was conducted by Bai-ley and Allum in 1962. Researchers at South Dakota State University compiled a comprehensive list on the fish fauna of South Dakota which included both his-torical and recent survey records, allowing researchers to examine changes in fish distributions across the state (Smith et al. 2002, Hayer et al. 2006, Hoagstrom 2006). Systematic sampling by several agencies in South Dakota (see acknowl-edgements) continues to produce new distributional records for many South Dakota fishes (Blausey 2001, Harland 2003, Duehr 2004, Morey and Berry 2004, and Hoagstrom 2006). Our objectives were to report on species that have demonstrated significant range extensions based on historical and recent sam-pling sites across South Dakota drainage basins (Figure 1). In addition, we will present an update on South Dakota species whose names have changed recently, and report on certain species that are either difficult to identify or that are often confused with others.

METHODS

Literature was used to determine historical fish species presence in South Dakota and additional data were gathered from several agencies across South Dakota and surrounding states and used to compile an updated list of South Dakota fishes and their distributions (Hayer et al. 2006, Hoagstrom 2006) by drainage basin (Figure 1). Fishes were collected by various methods (i.e. seining,

Figure 1. Map of South Dakota drainage basins.


electrofishing, hook and line) depending on the study. A total of 2,015 historic and recent sampling locations were compiled and used to examine potential range extensions based on 11 drainages (Figure 1). Taxonomic name changes were based on an American Fisheries Society special publication which updated previous taxonomic lists in an attempt to “achieve uniform use of common names and avoid confusion in scientific names” (Nelson et al. 2004). The list suggests scientific names to use based on current literature and views of special-ists on the various taxa.


Range Extensions

Eleven species demonstrated considerable range extensions from their histor-ical distributions (Table 1). Each species will be discussed below, beginning with where the new distribution occurred and the habitat in which it was collected. This will be followed by the nearest documented locality and a description of other drainage basins where the species has been previously collected in the state. We will then hypothesize why the range extension occurred and give the status of the species in South Dakota (South Dakota Game, Fish and Parks 2006) and in the surrounding states of Iowa, Minnesota, Nebraska, North Dakota, and Wyoming (Iowa DNR 2006, Minnesota DNR 2006, Nebraska Game and Parks Commission 2006, North Dakota Game and Fish Department 1994, Wyoming Game and Fish 2006).

Macrhybopsis storeriana (Kirtland) – silver chub One silver chub was col-lected from the Keya Paha River (Table 1, Figure 2) west of Clearfield, South Dakota in June 2002 (Harland and Berry 2004) by seining. The sampling site was comprised of 75% sand, 15% gravel, and 0.34 m/s velocities, which are common silver chub habitats (Harlan et al. 1987). The Keya Paha River is a tributary to the Niobrara River in Ne-braska. The nearest published collection of this species is located in the Niobrara River in Nebraska (228.5 km). This is the first verified record for silver chub west of the Missouri River in South Dakota. The silver chub is cat-egorized as a large-river species being restricted to the Missouri River and a few

Figure 2. Silver Chub, Macrhybopsis storeriana, point distributions, native range, and range extensions into the Keya Paha River drainage.


Table 1: List of species exhibiting range extensions in South Dakota, the source for the extension, and specific location for the new record.

SPECIES SOURCE BASIN STREAM

NAME UTM TRSSilver ChubMacrhybopsis storeriana

Harland and Berry 2004 Keya Paha Keya Paha River 420813E

4771278NT96N R77E

Sec. 32

Topeka shinerNotropis topeka

Blausey 2001; Wall et al. 2004

Vermillion, James andBig Sioux

NorthernRedbelly DacePhoxinus eos

Morey and Berry 2004 Grand River Stink Creek 328639E

5054981N

SouthernRedbelly DacePhoxinuserythrogaster

Springman and Banks 2005 Big Sioux Little Beaver

CreekT98N R49W

Sec. 34

ShortheadredhorseMoxostomamacrolepidotum

Harland and Berry 2004 Keya Paha

AntelopeCreek

KeyaPahaRiver

387310E 4787039N

410169E 4775627N 433595E

4765215N

T38N R26E Sec. 34

T96N R78E Sec. 17

T95N R76E Sec. 22

Harland 2003 Missouri Bull Creek 459801E 4836631N

T103N R73WSec. 36

Yellow bullheadAmeiurus natalis Hampton 1998 Cheyenne Cheyenne River 635300E

4805690N T8S R7E Sec. 6

Northern PikeEsox lucius


AntelopeCreek

WhiteWillow Creek

363361E

4789507N 420970E

4771463N

T38N R28E Sec. 19

T96N R77E Sec. 32

Duehr 2004 Moreau Thunder Butte Creek

269361E 5018930N

T15N R18E Sec. 5

Iowa darter Etheostoma exile Duehr 2004 Cheyenne Sulpher Creek 685306E

4968045NT10N R12E

Sec 10Johnny darter Etheostomanigrum

Duehr 2004 Moreau Moreau River 360637E 5015925N

T15N R27E Sec. 11

Harland 2003 Missouri

American Crow

Bull Creek

466945E 4850456N

459801E 4836631N

T104 R73W Sec 14

T103N R73W Sec. 36

Yellow perchPerca flavescens


WhiteWillow Creek Sand Creek

420970E 4771463N 409874E

4774390N

T96N R77E Sec. 32

T96N R78E Sec. 19

WalleyeSander vitreum Harland 2003 White Dog Ear Creek 419882E

4836084NT103N R77E

Sec. 36


watersheds within the James, Big Sioux, and Niobrara river basins (Figure 2, Bai-ley and Alum 1962, Hesse et al. 1979, Smith et al. 2002) and typically inhabits semi-turbid water in strong current over sand a gravel substrates (Harlan et al. 1987). Harland and Berry (2004) suggested that this collection was the result of silver chub using the Keya Paha for spawning activities. It is not a species of concern in South Dakota or surrounding states and this new record represents a notable range extension.

Notropis topeka (Gilbert) – Topeka shiner The Topeka shiner was recorded to occur in 24 streams before 1999 (Blausey 2001) and collections since 1999 have recorded its presence in 21 additional streams located throughout the Vermillion, James, and Big Sioux river drainages (Wall et al. 2004). The Topeka shiner is native to Big Sioux, Vermillion and James rivers in eastern South Dakota (Bailey and Allum 1962). These new local records may be the result of the Topeka shiner being listed as endangered by the U.S. Fish and Wildlife Service (Tabor 1998), which prompted surveys of many tributaries within their historical range (Blausey 2001, Wall et al. 2004). The To-peka shiner is considered a species of concern in South Dakota and threatened in Iowa; however, recent collections in Minnesota and South Dakota (Hatch 2001, Wall et al. 2004) at both historic and new locations suggest it is more persistent in the northern part of its range in Minnesota and South Dakota than the south-ern part of its range in Kansas (Wall and Berry 2004).

Phoxinus eos (Cope) – northern redbelly dace Thirteen northern red-belly dace were collected in June 2003 by pulsed-DC elec-trofishing (Table 1, Figure 3) in the Grand River drainage (Morey and Berry 2004). The sample reach consisted of a single, unconstrained channel that transected open prairie, with pools consisting mostly of silt substrate, filamentous algae, and rooted macro-phytes. The stream gradient was low (0.1%) and there was no apparent surface flow at the time of sampling. Mean wetted width was 1.33 m and mean depth was 23 cm. Water quality conditions consisted of water temperature (16.6° C), dissolved oxygen (14.6 mg/L), conductivity (748 µmhos.cm), and pH (7.69). Habitat conditions were similar to those reported for this species by others (Brown 1971, Eddy and Underhill 1976, Becker 1983, Bestgen 1989). The nearest published collection of this species is located in Blue Blanket Creek in the upper Missouri River

Figure 3. Northern redbelly dace, Phoxinus eos, point distributions, native range, and range extensions into the Grand River drainage.


drainage (approximately 76 km to the east). Northern redbelly dace also occur in the Big Sioux and Minnesota river drainages in South Dakota and historical populations are distributed throughout the western Great Plains, including the Niobrara River drainage, White River drainage, and several minor tributaries of the Missouri River (Figure 4, Bailey and Allum 1962, Cunningham et al. 1995, Smith et al. 2003, Morey and Berry 2004), where they are restricted to perennial streams with slow, clear water and abundant macrophytes and algae (Morey and Berry 2004). Additional sampling efforts in previously under-sampled drainages might provide new records of northern redbelly dace in South Dakota (Morey and Berry 2004). The northern redbelly dace is considered a species of concern in North Dakota and considered threatened in South Dakota and Nebraska. This collection represents a considerable range extension into the Grand River drainage and a more complete distribution of glacial relict populations in the northern Great Plains (Morey and Berry 2004).

Phoxinus erythrogaster (Rafinesque) - southern redbelly dace Springman and Banks (2005) collected 48 southern redbelly dace in Little Beaver Creek (Big Sioux River drain-age) in South Dakota (Table 1, Figure 4). The reach was located approximately 4 km upstream from the Big Sioux River confluence and ap-proximately 150 km upstream from the Big Sioux and Mis-souri River confluences. The stream reach habitat consisted of clear, moderately flowing water with undercut banks, overhanging vegetation, and a substrate mixture of sand, silt, clay, and gravel substrates, which are typical southern redbelly dace habitats (Springman and Banks 2005). The southern redbelly dace has not been previously reported in South Dakota; however, they have been noted to occur in the nearby Big Sioux River drainage in Minnesota (Bailey and Allum 1962, Lee et al. 1980). The southern redbelly dace is fairly widespread throughout its range in Minnesota where they are not considered a species of concern.

Figure 4. Southern redbelly dace, Phoxinus erythrogas-ter, point distributions, native range, and range exten-sions into the Big Sioux River drainage.


Moxostoma macrolepidotum (Lesueur) – shorthead redhorse Four shorthead redhorse were collected from Antelope Creek (Table 1, Figure 5) and three total were collected at two Keya Paha River locations in early summer 2002 by sein-ing (Harland and Berry 2004). The sample reach was a third-order stream characterized by 0.30 m3/s flow velocities and sand a gravel substrates which is typical shorthead redhorse habitat (Pflieger 1997). The nearest documented locality in South Dakota is the Little White River (approximately 121 km west). In addition, one shorthead redhorse was collected in July 2002 in Bull Creek, a tributary to Lake Francis Case, a Missouri River reservoir (Table 1, Figure 6; Harland 2003). The nearest documented locality in South Dakota to this record is Medicine Knoll Creek in the Missouri River drainage, approximately 87 km North. The shorthead redhorse is com-mon in all major drainages in South Dakota (Figure 5, Bailey and Allum 1962). Collection of this species may have been the result of sampling in a previously under sampled region. These two collections represent range extensions into the Keya Paha and Missouri River drainages. The shorthead redhorse is not a species of concern in South Dakota or surrounding states.

Ameiurus natalis (Lesueur) – yellow bullhead Two yellow bullheads were taken from two reaches in the upper Cheyenne River in Fall River and Custer counties, South Dakota in 1997 (Table 1, Figure 6, Hampton 1998). One reach was located just below Angostura Dam in Fall River County, and the other was located near the intersec-tion of Custer, Pennington, and Shannon counties, South Dakota. They were collected in slow-moving stretches with submerged aquatic vegetation and substrates of sand, silt and gravel. The nearest docu-

Figure 5. Shorthead redhorse, Moxostoma macro-lepidotum, point distributions, native range, and range extensions into the Keya Paha River drainage.

Figure 6. Yellow bullhead, Ameiurus natalis, point dis-tributions, native range, and range extensions into the Cheyenne River basin.


mented localities for this species in South Dakota are in the Big Sioux and James rivers, approximately 482.8 km to the east (Figure 6). The yellow bullhead is na-tive to the Bios de Sioux, James, and upper Minnesota river drainages in eastern South Dakota (Bailey and Allum 1962). This is the western most native range for this species; however they have been extensively introduced into numerous western states (Fuller et al. 1999). This range extension in the Cheyenne River may be the result of an introduction. The yellow bullhead is not a species of concern in South Dakota or surrounding states and this new record represents a substantial range extension in the state and its entire reported range (Scott and Crossman 1973).

Esox lucius Linnaeus – northern pike In May and June 2002, two northern pike were col-lected by seining from Ante-lope and White Willow creeks, both of which are tributaries to the Keya Paha River in south-central South Dakota (Table 1, Figure 7, Harland and Berry 2004). Northern pike were collected in reaches with aquatic vegetation and low flow velocities which are typical northern pike habitats (Pflieger 1997). The nearest documented locality in South Dakota is the Little White River, approximately 96.6 km west from this new collection. Harland and Berry (2004) suggested that high stream-flows prior to sampling created washouts from nearby stocked ponds allowing northern pike access to the tributaries, resulting in this new collection. In addition, two northern pike were collected by seining in June 2003 in Thunder Butte Creek located within the Moreau River drainage (Table 1, Figure 8, Duehr 2004). The nearest documented locality to this new record is the Little Missouri River, approximately 165 km west. Northern pike are pres-ent in most drainages in South Dakota (Figure 7). They are considered native in the Red, Minnesota, Big Sioux, and James Rivers (Hoagstrom 2006). They are considered non-native (introduced) in the Vermillion, Missouri, and drainages west of the Missouri River (Bailey and Allum 1962, Fuller et al. 1999). These new records are the first to be recorded in the Keya Paha and Moreau river drain-ages. The northern pike is not considered a species of concern in South Dakota or surrounding states.

Figure 7. Northern Pike, Esox lucius, point distributions, native range, and range extensions into the Keya Paha River basin. Basins where the species is not found are white.


Perca flavescens (Mitchill) – yellow perch Three yellow perch were collected (along with north-ern pike) at White Willow Creek and one yellow perch was collected at Sand Creek (Table 1, Figure 8, Harland and Berry 2004). White Wil-low and Sand creeks are both tributaries in the lower South Dakota portion of the Keya Paha River. All specimens were collected by seining in May 2002 in reaches with low flow velocities, instream veg-etation, and coarse substrates of sand and gravel. The near-est documented locality from these locations for yellow perch in South Dakota is the Little White River, approximately 120.7 km west (Figure 8). The yellow perch is native to Eastern drainages of South Dakota and introduced across the rest of the state (Bailey and Alum 1962, Fuller et al.1999). As previously stated for northern pike, high stream flows prior to sampling cre-ated washouts from nearby stock ponds which may account for their presence in the tributaries (Harland and Berry 2004). The yellow perch is not a species of concern in South Dakota or surrounding states.

Sander vitreus vitreus (Mitchill) – walleye One walleye was collect-ed in summer 2002 by hook and line from Dog Ear Creek (Table 1, Figure 9), 2 km upstream from its confluence with the White River (Har-land 2003). The specimen was collected in typical walleye habitat (Pflieger 1997) below a culvert crossing in a deep pool where old culverts, concrete rip-rap, and woody substrate comprised the stream bottom. The nearest documented col-lection of walleye in South Dakota is Lake Francis Case. Walleye were introduced to Montana (Brown 1971, Holton and Johnson 2003)

Figure 8. Yellow perch, Perca flavescens, point distribu-tions, native range, and range extensions into the Keya Paha River basin.

Figure 9. Walleye, Zander vitreus vitreus, point dis-tributions, native range, and range extensions into the White River basin.


and Wyoming (Simon 1946, Baxter and Simon 1970, Baxter and Stone 1995), and South Dakota, Nebraska, and North Dakota comprise the southwestern edge of the historical native range (Bailey and Allum 1962, Cross 1967, Cross and Collins 1995, Fuller et al. 1999). Walleye is reported as native to all major drainages of South Dakota except the Little Missouri River drainage (Hoagstrom 2006). Walleye typically spawn between March and April (Harlan et al. 1987) where they move from large rivers and reservoirs into tributary streams usually during increased flow events. This specimen may have been collected as a result of being trapped in this tributary it was using for reproduction during the 2002 drought (Harland 2003). Walleye is not considered species of concern in South Dakota or surrounding states. Etheostoma exile (Girard) – Iowa darter Thirty-five Iowa dart-er specimens were collected from Sulphur Creek (Table 1, Figure 10) in the Cheyenne River drainage (Duehr 2004) by seining. The sampling reach was located upstream of the Cherry Creek and Red Owl Creek confluence (238 km from the Cheyenne Riv-er) and was characterized by small substrate (predominately muck, <0.1mm) with a mod-erate slope (0.3 m/km). The Iowa darter prefers clear, slug-gish, or standing waters with submerged aquatic vegetation (Bailey and Allum 1962). The nearest documented locality for the Iowa darter in South Dakota is on the North Fork of the Moreau River, approximately 45 km north. The Iowa darter is native to all major drainages east of the Missouri River (Figure 10, Big Sioux, James, Minnesota, Red Vermillion, and eastern and southwestern tributaries to the Mis-souri) and the Niobrara in the west (Bailey and Allum 1962). This new record represents a considerable range extension of its historical range and may be the result of additional sampling in previously undersampled regions and may rep-resent a more complete distribution of glacial relict populations in the northern Great Plains. The Iowa darter is not a species of concern in South Dakota or surrounding states.

Figure 10. Iowa darter, Etheostoma exile, point dis-tributions, native range, and range extensions into the Cheyenne River basin.


Etheostoma nigrum Rafinesque – johnny darter Seven johnny darters were collected by seining from the mainstem of the Moreau River in summer 2003 (Table 1, Figure 11, Duehr 2004). The johnny darter is tolerant of clear and turbid waters with a variety of substrate types and prefers stream habitats, but can also be found in lakes . The nearest documented local-ity for johnny darter in South Dakota is Medicine Knoll Creek in the Missouri River drainage, approximately 110 km east. In addition, 48 johnny darters were collected in Bull and American Crow creeks, tributaries of Lake Francis Case (Table 1, Figure 11, Harland 2003). The johnny darter is considered native to all drainages east of the Missouri River (Big Sioux, James, Minnesota, and Vermillion) and to the Niobrara River drainage in the west (Bailey and Allum 1962, Hoagstrom 2006). It is considered non-native to the Grand, Moreau, and upper Missouri Valley drainages (Figure 11, Hoagstrom 2006). This species is able to tolerate broad environmental condi-tions (Trautman 1981) and are considered highly invasive (Kuehne and Barbour 1983), which could account for this range extension. These two collections rep-resent new collections in the Missouri and Moreau river drainages. The johnny darter is not considered a species of concern in South Dakota or surrounding states.

Taxonomic Name Changes And Potential Misidentifications

Macrhybopsis hyostoma (Gilbert) – shoal chub The prairie chub, Macrhybopsis australis, shoal chub, M. hyostoma, burrhead chub, M. marconis, and Arkansas river speckled chub, M. tetranema were all formerly considered the speckled chub, M. aestivalis, but have recently been rec-ognized as distinct species (Table 2, Eisenhour 1997, 1999, Nelson et al. 2004) with varying geographic distributions. The shoal chub is the only one of these species that is present in South Dakota, where it is restricted to the lower Mis-souri Valley drainage (Hoagstrom 2006). Shoal chub habitat consists of large, low gradient, small to large rivers with broad shallow riffles over sand or mud and in fast riffles over firm gravel, often in fast water over shifting sand. It is also tolerant of high turbidity and dissolved solids.

Figure 11. Johnny Darter, Etheostoma nigrum, point distributions, native range, and range extensions into the Moreau River basin.


Notropis percobromus (Cope) - carmine shiner The carmine shiner is native to the Minnesota drainage (Bailey and Allum 1962) and more recently (post-1990 collections) the Big Sioux River (Table 2, Hoagstrom 2006). This species was previously referred to as the rosyface shiner (Notropis rubellus) in South Dakota (Bailey and Allum 1962); however, recent examination of the rosyface shiner based on geographic variation of allozyme products has separated it into three species (highland shiner, N. micropteryx, carmine shiner, N. percobromus, and rosyface shiner) with differing geographic distributions (Wood et al. 2002). The Minnesota River drainage in South Da-kota represents the western tip of the carmine shiner distribution. The carmine shiner is considered a species of concern in both South and North Dakota (North Dakota Game and Fish Department 1994), though not a species of concern in Minnesota.

Rhinichthys obtusus Agassiz – Western blacknose dace Historically, eastern, southern, and western forms of blacknose dace (Rhinichthys atratulus) were considered subspecies (R. a. atratulus, R. a. obtusus, and R. a. meleagris, Respectively, Hubbs and Lagler 1958, Scott and Crossman 1973). Recently, Smith (1985) and Jenkins and Burkhead (1994) proposed to recognize R. a. atratulus as one species (eastern blacknose dace, R. atratulus Her-man) and R. a. meleagris and R. a. obtusus as another (Table 2, western blacknose dace, R. obtusus Agassiz). These suggestions were accepted by the American Fisheries Society (Nelson et al. 2004). Western blacknose dace occur in eastern South Dakota in the Big Sioux, James, Minnesota, and Vermillion river drain-ages. Recent collections (post-1990) also document the western blacknose dace in the White River drainage (Hoagstrom 2006). The western blacknose dace is often confused with the longnose dace (R. cataractae cataractae), which has never been verified to occur in the Big Sioux, James, Minnesota, and Vermillion river drainages. However, it is important that researchers recognize the utility of voucher specimens and photographs to avoid misidentifications. The western

Table 2: List of species recent name changes and the sources of the change

OLD NAME NEW NAME SOURCE

Speckled ChubMacrhybopsis aestivalis

Shoal ChubM. hyostoma

Eisenhour 1999, Nelson et al. 2004

Rosyface shinerNotropis rubellus

Carmine shinerN. percobromus

Wood et al. 2002, Nelson et al. 2004

Sand shinerN. stramineus

Plains sand shinerN. s. missuriensis

Eastern sand shinerN. s. stramineus

Bailey and Allum 1962, Tan-yolac 1973,

Nelson et al. 2004

Blacknose daceRhinichthys meleagrisatratulus

Western Blacknose daceR. obtusus

Burkhead 1994,Nelson et al. 2004

Plains killifishFundulus zebrinus

Northern plains killifishF. kansae

Nelson et al. 2004, Kreiser 2001, Kreiser et al. 2001


blacknose dace is not considered a species of concern in South Dakota or sur-rounding states.

Fundulus kansae Garman – northern plains killifish The taxonomy of the plains killifish, Fundulus zebrinus, has been debated for over a century (Rahel and Thel 2004). The plains killifish is now recognized as two species (Table 2, Nelson et al. 2004): the northern plains killifish, F. kansae, and the southern plains killifish, F. zebrinus (Nelson et al. 2004, Kreiser 2001, Kreiser et al. 2001). Northern plains killifish is considered to be potentially introduced in the Cheyenne River drainage (Miller 1955, Bailey and Allum 1962). However, Hoagstrom (2006) states that as a result of highly variable and patchy distributions, earlier surveys may have failed to detect the species, and thus he considers the northern plains killifish to be native to the Cheyenne River drainage in South Dakota. Recent collections in the Cheyenne River drainage indicate that the northern plains killifish are relatively common as they have been collected on the mainstem and tributaries of the Cheyenne and Bell Fourche riv-ers (Hampton 1998, Duehr 2004, Hoagstrom 2006). Northern plains killifish is not a species of concern in South Dakota or surrounding states.

Notropis stramineus stramineus (Cope) – eastern sand shiner N. s. missuriensis (Cope) - plains sand shiner The sand shiner (Notropis stramineus) in South Dakota is represented by two subspecies, the plains sand shiner, N. s. missuriensis, and eastern sand shiner, N. s. stramineus (Bailey and Allum 1962). The plains sand shiner inhabits the Great Plains and is present in all major drainages west and including the Missouri River (Table 3, Bad, Cheyenne, Grand, Moreau, Niobrara and White drainages), while the eastern sand shiner occupies the central lowlands, interior highlands and coastal plains and is present in eastern drainages of South Dakota (Table 3, Big Sioux, James, Minnesota, and Vermillion river). These two subspecies are dif-ficult to distinguish and are often not differentiated. Diagnostics differentiating the two subspecies include circumference scales, post-orbital diameter of head, orbital diameter, head width, and predorsal scale-rows (Table 3, Tanyolac 1973).

Table 3: Distinguishing features and distribution in South Dakota for two sand shiner subspecies: Plains sand shiner, Notropis stramineus missuriensis, and eastern sand shiner, N. s. stramineus. Characters were designated by Tanyolac 1973.

CHARACTERS PLAINS SAND SHINER EASTERN SAND SHINER

Drainages present in SDBad, Cheyenne, Grand,

Missouri, Moreau,Niobrara, White

Big Sioux, James,Minnesota, Vermillion

Circumference scales 27-30 22-25Postorbital length of head 13-14% of SL 11-12% of SLOrbital diameter <8% SL >8% SLHead width >14% SL <15% SLPredorsal scale rows 15-17 13-15


We recommend that specimens be preserved for accurate identification and fu-ture verification in future studies

Carpiodes – Carpsuckers South Dakota waters contain three carpsuckers, the northern river carpsuck-er, Carpiodes carpio carpio, the central quillback carpsucker, C. cyprinus hinei, and the highfin carpsucker, C. velifer, all of which are difficult to differentiate (Table 4). The northern river carpsucker is common to all major drainages in

South Dakota with the exception of the Minnesota and Red river drainages (Bai-ley and Allum 1962, Hoagstrom 2006). The northern river carpsucker is closely related to the highfin carpsucker, which is native to the Big Sioux and lower Missouri Valley river drainages (Bailey and Allum 1962; Hoagstrom 2006). The central quillback carpsucker is native to South Dakota drainages east of the Mis-souri River (Hoagstrom 2006). One highfin carpsucker was reported from the James River drainage in 2000 (Shearer 2001, Shearer and Berry 2002, 2003), but after examination of the preserved specimen, it was concluded that it was the central quillback carpsucker (Hoagstrom 2006). This previous misidentification of this specimen was based on its lack of a nipple on the lower lip (Pflieger 1997) and patterns of tuberculation (Huntsman 1967). As a result of the difficulty in correctly differentiating between these three species of carpsucker, particulalry young of the year, it is important to take pictures and if possible voucher speci-mens to avoid misidentification.

CONCLUSIONS

A total of ten species of fish are new to the Big Sioux (southern redbelly dace), Cheyenne River (yellow bullhead and Iowa darter), Grand (northern

Table 4: Distinguishing features and distribution in South Dakota of three carpsuckers, Northern river carpsucker, Carpiodes carpio carpio, central quillback carpsucker, C. cyprinus hinei, and highfin carpsucker, C. velifer.

CHARACTERS

CENTRAL QUILLBACK

CARPSUCKERHIGHFIN

CARPSUCKER

NORTHERN RIVER

CARPSUCKER

Drainages present in SD Drainages east of Missouri River

Big Sioux andlower Missouririver drainages

All drainagesexcept Minnesota

and Red riverNipple on lower lip Absent Present Present

Anterior dorsal rays Long as base of the fin

Elongated – as long as base of fin

More than 1⁄2 the base of the fin

Lateral line scales 37-40 33-37 33-37

Mouth location Anterior to nostrils Mostly posteriorto nostrils Posterior to nostrils


redbelly dace), Keya Paha (silver chub, shorthead redhorse, northern pike, and yellow perch), Moreau (johnny darter and northern pike), White (walleye) and Missouri (shorthead redhorse, johnny darter, and walleye) river drainages in South Dakota. As fish distributions commonly change through time, future collections may reveal further changes and movements. There are three possible reasons for the range extensions presented in this paper: 1) additional sampling efforts in under surveyed drainages (i.e. west river drainages) reveal new records (e.g. northern redbelly dace, southern redbelly dace and Iowa darter), 2) invasive species range expansions within a river system (e.g. johnny darter) and 3) newly introduced/stocked species (e.g. northern pike, yellow bullhead). In conclusion, it is important to document fish range extensions in our constantly changing aquatic systems in South Dakota. We emphasize the importance of preserving specimens for later examination in order to obtain a more detailed and adequate list of species distributions.

ACKNOWLEDGEMENTS

Funding was provided by Federal Aid in Sport Fish Restoration under D. J. Project #F-57-R-1. We thank Jeremy Duehr, Chris Hoagstrom, and Jason Kral for their assistance with field data collection, and landowners who made these surveys possible. We thank those who allowed us to add their fish collection data to ours for a complete fish distribution list of South Dakota:, Chris Hoagstrom, Jeremy Duehr, Sheila Thomson – SDSU, Steve Freeling – Vermillion River Wa-ter Development District, Jeff Shearer – SDGF&P, Nathan Morey – South Da-kota Department of Transportation, Jason Kral – U.S. Geological Survey EMAP project, and Ryan Sylvester and Steve Freeling – UMRBGAP. We also thank Chris Hoagstrom and Steve Herrington for helping to improve this manuscript. The South Dakota Department of Game, Fish, and Parks, US Geological Survey, Wildlife Management Institute, US Fish and Wildlife Service and South Dakota State University jointly support the South Dakota Cooperative Research Unit.

LITERATURE CITED

Bailey, R.M., and M.O. Allum. 1962. Fishes of South Dakota. Miscellaneous Publications, Museum of Zoology, University of Michigan, No. 119, Ann Arbor.

Baxter, G.T., and J.R. Simon. 1970. Wyoming Fishes, revised. Bulletin No. 4. Wyoming Game and Fish Department, Cheyenne.

Baxter, G.T., and M.D. Stone. 1995. The fishes of Wyoming. Wyoming Game and Fish Department, Cheyenne.

Becker, G.C. 1983. Fishes of Wisconsin. University of Wisconsin Press, Madi-son, Wisconsin.

Bestgen, K.R. 1989. Distribution and notes on the biology of Phoxinus eos (Cyprinidae) in Colorado. Southwestern Naturalist 34: 225-231.


Blausey, C.M. 2001. The status and distribution of the Topeka shiner Notropis topeka in eastern South Dakota. Masters thesis, South Dakota State Univer-sity, Brookings, SD.


Cross, F.B. 1967. Handbook of fishes of Kansas. University of Kansas, Museum of Natural History, Miscellaneous Publication No. 45, Lawrence.

Cross, F.B., and J.T. Collins. 1995. Fishes in Kansas, second edition, revised. University Press of Kansas, Lawrence.

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 River. Proceedings of the South Dakota Academy of Science 74: 55-63.

Duehr, J.P. 2004. Fish and habitat relations at multiple spatial scales in Chey-enne River Basin, South Dakota. Master’s thesis. South Dakota State Uni-versity, Brookings.

Eddy, S. and J.C. Underhill. 1976. Northern fishes. University of Minnesota Press, Minneapolis, Minnesota.

Eisenhour, D. J. 1997. Systematics, variation, and speciation of the Macrhybopsis aestivalis complex (Cypriniformes: Cyprinidae) west of the Mississippi River. Ph.D. dissertation, Southern Illinois University, Carbondale.

Eisenhour, D. J. 1999. Systematics of Macrhybopsis tetranema (Cypriniformes: Cyprinidae). Copeia 1999:969-980.

Fuller, P.L., L.G. Nico, and J.D. Williams. 1999. Nonindigenous fishes intro-duced into inland waters of the United States. American Fisheries Society, Special Publication 27, Bethesda, Maryland.

Hampton, D. R. 1998. A survey of the fishes and habitat of the Cheyenne River in South Dakota. M. S. Thesis. South Dakota State University, Brookings.

Harlan, J. R., E. B. Speaker, and J. Mayhew. 1987. Iowa fish and fishing. Iowa Department of Natural Resources, Des Moines.

Harland, B.C. 2003. Survey of the fishes and habitat of western South Dakota streams. Master’s thesis. South Dakota State University, Brookings.

Harland, B. and C.R. Berry, Jr. 2004. Fishes and habitat characteristics of the Kaya Paha River, South Dakota-Nebraska. Journal of Freshwater Ecology 19(2): 169-177.

Hatch, J.T. 2001. What we know about Minnesota’s first endangered fish spe-cies: The Topeka shiner. Journal of Minnesota Academy of Sciences 65:39-46.

Hayer, C. A., S. S. Wall and C. R. Berry. 2006. Evaluation of aquatic GAP analysis fish distribution models with emphasis on rare fish species in South Dakota. Brookings, South Dakota State University: 107pp.

Hesse, L.W., G. Zuerlein, R. Vancil, L. Koziol, B Newcomb, and L.A. Retelsdorf. 1979. Niobrara-Missouri River fishery investigations. Nebraska Game and Parks Commission Technical Series No. 5, 39 p.

Hoagstrom, C.W. 2006. Zoogeographic patterns and faunal change of South Dakota fishes. PhD dissertation. South Dakota State University, Brookings, SD.


Holton, G.D., and H.E. Johnson. 2003. A field guide to Montana fishes. Mon-tana Fish, Wildlife and Parks, Helena.

Hubbs, C.L., and K.F. Lagler. 1958. Fishes of the Great Lakes region. Bulletin number 26, Cranbook Institute of Science, Bloomfield Hill, Michigan.

Huntsman, G.R. 1967. Nuptial tubercles in carpsuckers (Carpiodes). Copeia 1967:457-458.

Iowa Department of Natural Resources. 2006. Endangered and threatened plant and animal species. http://www.iowadnr.com/other/files/chapter77.pdf. Assessed 14 June 2006.

Jenkins, R.E., and N.M. Burkhead. 1994. Freshwater fishes of Virginia. Ameri-can Fisheries Society, Bethesda, MD.

Kreiser, B.R. 2001. Mitochondrial cytochrome b sequences support the recogni-tion of two cryptic species of killifish: Fundulus kansae and Fundulus zebri-nus. The American Midland Naturalist 146:199-209.

Kreiser, B.R., J.B. Mitton, and J.D. Woodling. 2001. Phylogeography of the plains killifish, Fundulus zebrinus. Evolution 55: 339-350.

Kuehne, R.A. and R.W. Barbour. 1983. The American darters. University Press Kentucky, Lexington, 177 pp.

Lee, D.S., C.R. Gilbert, C.H. Hocutt, R.E. Jenkins, D.E. McAllister, and J.R. Staufer, Jr. 1980. Atlas of North American freshwater fishes. North Caro-lina State Museum of Natural History, Raleigh, North Carolina.

Miller, R.R. 1955. An annotated list of the American cyprinodontid fishes of the genus Fundulus, with the description of Fundulus persimilis from Yucatan. Occasional Papers of the Museum of Zoology, University of Michigan, No. 568, Ann Arbor.

Minnesota Department of Natural Resources. 2006. Endangered fish. http://www.dnr.state.mn.us/ets/fish.html. Accessed 14 June 2006.

Morey, N.M., and C.R. Berry, Jr. 2004. New distributional record of the northern redbelly dace in the northern Great Plains. Prairie Naturalist 36:257-260.

Nelson, J.S., E.J. Crossman, H. Espinosa-Perez, L.T. Findley, C.R. Gilbert, R.N. Lea, and J.D. Williams. 2004. Common and scientific names of fishes from the United States, Canada, and Mexico. American Fisheries Society, Special Publication 29, Bethesda, Maryland.

Nebraska Game and Parks Commission. 2006. Nongame wildlife list. http://www.ngpc.state.ne.us/wildlife/programs/nongame/list.asp. Assessed 14 June 2006.

North Dakota Game and Fish Department. 1994. Fishes of the Dakotas. North Dakota Game and Fish Department, Bismarck, ND. Jamestown, ND: Northern Prairie Wildlife Research Center Online. http://www.npwrc.usgs.gov/resource/fish/dakfish/dakfish.htm (Version 16JUL97).

Pflieger, W.L. 1997. The fishes of Missouri, revised edition. Missouri Depart-ment of Conservation, Jefferson City.

Rahel, F.J. and L.A. Thel. 2004. Plains killifish (Fundulus zebrinus): a technical conservation assessment. [Online]. USDA Forest Service, Rocky Mountain Region. Available: http://www.fs.fed.us/r2/projects/scp/assessments/plain-skillifish.pdf [March 27, 2006].


Scott, W. B. and E. J. Crossman, 1973. Freshwater fishes of Canada. Fisheries Research Board of Canada Bulletin. 184:1-966.

Shearer, J.S. 2001. Temporal change in fish communities and modification of the index of biotic integrity for the James River of the Dakotas. Master’s thesis, South Dakota State University, Brookings.

Shearer, J.S., and C.R. Berry, Jr. 2002. Index of biotic integrity utility for the fishery of the James River of the Dakotas. Journal of Freshwater Ecology 17:575-588.


Simon, J.R. 1946. Wyoming fishes. Wyoming Game and Fish Department, Bulletin No. 4, Cheyenne.

Smith, C.L. 1985. The inland fishes of New York State. The New York State Department of Environmental Conservation, Albany.

Smith, K. A. 2000. Factors influencing rainbow smelt entrainment through Oahe Dam, South Dakota. M. S. Thesis. South Dakota State University, Brookings.

Smith, V. J., J. A. Jenks, C. R. Berry, C. J. Koplin, and D. M. Fecske. 2002. The South

Dakota GAP Analysis Project: Final Report. USGS Gap Analysis Program. University of Idaho, Moscow.

Smith, V.J., J.A. Jenks, C.R. Berry, Jr., C.J. Kopplin, and D.M. Fecske. 2003. South Dakota GAP analysis project final report. United States Geological Survey, Brookings, South Dakota.

Springman, D.J. and R.L. Banks. 2005. Range extension of the southern red-belly dace into South Dakota. Prairie Naturalist. 37(3): 175-176.

South Dakota Game Fish and Parks. 2006. Threatened, endangered, or candi-date species of South Dakota. http://www.sdgfp.info/Wildlife/Diversity/in-dex.htm. Accessed 14 June 2006. Pierre, SD.

South Dakota GAP Analysis Project (SDGAP) Home Page. 2001. Fishes of South Dakota. 14 June 2004 <http://wfs.sdstate.edu/sdgap/fish/fishes.htm>. USDI (United States Department of the Interior, Fish and Wildlife Service). 1997. Fishes of the Little White River South Central South Da-kota. U.S. Fish and Wildlife Service, Pierre.

Tabor, V.M. 1998. Final rule to list the Topeka shiner as endangered. Federal Register 63 (240): 69 008 – 69 021. US Department of the Interior, US Fish and Wildlife Service, Manhattan, Kansas.

Tanyolac, J. Tanyolac J. Morphometric variation and life history of the cyprinid fish Notropis stramineus (Cope). 1973. Occasional Papers of the Museum of Natural History, The University of Kansas, Lawrence.

Trautman, M.B. 1981. The fishes of Ohio. Revised edition. Ohio State Uni-versity Press, Columbus, 782 pp.

Wall, S.S. and C.R. Berry, Jr. 2004. Threatened fishes of the world: Notropis to-peka Gilbert, 1984 (Cyprinidae). Environmental biology of fishes 70:246.


Wall, S. S., C. R. Berry, Jr., C. M. Blausey, J. A. Jenks and C. J. Kopplin. 2004. Fish-habitat modeling for gap analysis to conserve the endangered Topeka shiner (Notropis topeka). Canadian Journal of Fisheries and Aquatic Sciences 61: 954-973.

Wood, R. M., R. L. Mayden, R. H. Matson, B. R. Kuhajda and S. R. Laymen. 2002. Systematics and biogeography of the Notropis rubellus species group (Teleostei: Cyprinidae). Bulletin of the Alabama Museum of Natural History 22: 37-80.

Wyoming Game and Fish. 2006. http://www.gf.state.wy.us. Assessed 14 June 2006.


ECONOMIC PERFORMANCE INFORESTRY-DOMINATED LINCOLN

COUNTY, MONTANA: 1969-2003

Russell L. StubblesDepartment of Horticulture, Forestry, Landscape and Parks


ABSTRACT

This study reviewed economic performance in forestry-dominated Lincoln County, Montana, from 1969 to 2003. Seven economic categories were cho-sen for the review: household median income, overall total personal income, per capita income, full-time and part-time employment, employment growth, total industry earnings, and average earnings per job. It was found that Lincoln County’s growth did not keep pace with Montana and/or the nation. Possible reasons for this sluggish economic growth were given. Future economic growth scenarios for the next ten years were discussed.

Keywords

Kootenai National Forest, forest production, forest recreation, economic growth, tourism, Lincoln County, Montana.

BACKGROUND

Today, the Kootenai National Forest (KNF) in Mont. and Idaho (dedicated in 1906) consists of 2.2 million acres, of which 1,690,000 acres are in Lincoln County (LC), Mont., in the northwestern corner of the state (British Colum-bia is the northern border, and Idaho is the western border). Adjacent to the south of LC, the KNF has 428,500 acres in Sanders County and 49,100 acres in Flathead County adjacent to the east. In the panhandle of Idaho there are 39,200 acres of the KNF in Bonner County and 10,300 KNF acres in Boundary County (Frament 2006). Almost 93% of LC is classified as forested land. Approximately 77% of that forested land is the KNF (Lincoln County Forest Stewardship Guide 2006). As expected, this ownership by the KNF impacts LC’s available tax base, patterns of land development, and population densities (Kootenai National Forest 2004). In 1969 LC’s population was 17,585. By 2003 it had grown to 18,892, making it the tenth most populated county of Montana’s 56 counties (Bureau of Economic Analysis 2003). Note: LC’s population as a percent of the statewide total is slowly falling over time, and the population is aging as well. In 2000 the median age for LC was 42.1 yrs. In 1990 it was 34.7 yrs. There is an increase in


populations over 50 years of age and a decrease in populations of less than 25 years of age (Kootenai National Forest 2004).

LINCOLN COUNTY’S ECONOMICS

National forests and county governments sometimes share common bound-aries, with the forests comprising a very large share of the county lands, as in the case of LC and the KNF. The KNF lands directly impact the economic vitality of LC. This leads to two basic questions:

1. What has been the general economic health of Lincoln County? 2. What will be the economic vitality of Lincoln County within the next

ten years?

HISTORICAL ECONOMIC RECORD

An analysis of the historical economic record of Lincoln County from 1969-2003 came from an examination of selected data banks and reports, focusing on specific economic categories. These selected economic categories reflect not only the usual local economic impact of having a national forest throughout the county but also the financial assistance (revenues) and longevity of federal-impact-to-counties programs, such as the Payments in Lieu of Taxes, a national program implemented after 1976. And, in LC’s case, the selected economic categories will also reflect the economic consequences of the W.R. Grace mine shutdown in 1990 (due to asbestos contamination). It is now a federal Superfund Site. Economists use various categories when measuring economic growth in counties (Kwang-koo et al. 2005). This research used seven universally accepted categories to measure the economic growth in LC:

1. Median Household Income,2. Overall Total Personal Income, 3. Per Capita Income,4. Fulltime and Part-time Employment, 5. Employment Growth, 6. Total Industry Earnings, and7. Average Earnings Per Job.

Generally, Lincoln County experienced slow economic growth from 1969-2003 in all seven categories, falling behind Montana and/or the nation (Smith 2006). As an example of this historical trend, LC’s median household income in 2000 was $26,754, while adjacent Flathead County, Mont., had a household median income of $34,466. Mont. was at $32,045 and the nation at $41,994 (US Census Bureau 2006).


DISCUSSION

Thus, from 1969 to 2003 LC fell behind the economic growth of Mont. and the nation; the result of many complex and integrated factors and forces. For example, during that time period the federal forests began a trend of providing less and less timber to the nation’s mills and for export. Today, the vast majority of timber for American mills, including those in LC, comes from state forests, private forests, corporate-owned forests, and imports. In 2006 the KNF provides very little timber to the two operating mills in LC. Most of the KNF timber goes out of the county – some to mills in nearby counties and some to out-of-state mills (Rumelhart 2006). For the record, industry-owned forested lands in LC are as follows: Plum Creek Timber Company, 270,000 acres and Stimson Lumber Company, 30,000 acres. Montana’s Department of Natural Resources and Conservation manages about 165,000 acres of Montana School Trust Lands in the county (Lincoln County Forest Stewardship Guide 2006). Perhaps LC’s transportation costs, the costs of environmental regulations and safeguards, and the general cost of timber production and operations hurt its competitiveness? However, in a similar situation in the forest-driven economy of Siskiyou County, Oregon:

The impacts of environmental regulations have been less than those stem-ming from changes in timber demand, increased efficiency in timber pro-cessing, and gradual but constant change from a manufacturing to a service economy. The driving force behind manufacturing decline is globalization, not stronger environmental regulation (Norgaard 1997: 8).

It is moot. The fact remains: Since at least 1969 LC has fallen behind the economic growth of both Montana and the nation.

POSSIBLE FUTURE SCENARIOS

Counties are constantly trying to solve their economic problems – espe-cially those problems that lead to economic stagnation and slow growth. LC is no exception. The LC Commissioners, the KNF personnel, the Kootenai River Development organization, and other groups such as the Eureka Rural Development Partners and the LC Forest Stewardship Coalition are advocating continued use of the KNF for forest production and related industries. New communication, collaboration, and cooperation plans and programs between all interested parties are being pursued. This includes an emphasis on participation in the relatively new Forest Stewardship Contracting Program. An example of this program was the Treasure Interface Stewardship Pilot Project in 2002 on the Libby Ranger District of the KNF. It promoted reduced (through thinning) for-est fuels, while creating wildlife habitat diversity on 765 acres. The innovation was that the receipts from the thinning were then used within the project area to build a restroom and a picnic shelter (Lincoln County Forest Stewardship Guide 2006). Some of LC’s future economic growth will come from present and


new uses of the KNF and its array of natural resources. However, America’s ru-ral economy is moving from natural resource extraction activities and its related processing to services (tourism and/or retirement). A second economic scenario involves expanding the KNF’s forest recreation offerings as part of a new effort to increase the local tourism and commercial recreation industries. Federal forest recreation is already an important part of the local economy, including tourism into the wilderness area. Recent economic data suggest that a shift in the economic trend toward tourism is already under-way for LC. For example, the KNF receives approximately 1,300,000 visitors per year and that average expenditure per visitor is @$67 per visit. Note: Only visitor expenditures made within 50 miles of the recreation site are included in this computation of the average spending (Stynes and White 2006). Future recreational opportunities will be even more important to the LC economy. Included in this second scenario is that the local hunting and fishing indus-tries, as well as the B&B businesses, will expand. There will be new tourism sub-sectors: adventure tourism, avitourism, agritourism, ecotourism, and so-called extreme-tourism. Most of this new success will be from the direct and aggressive use of the Internet, networking in general, and through new and expanding co-alitions of like-minded businesses. A third economic scenario, related to the second scenario, is that even with limited promotion LC will experience a situation similar to that of economically booming Flathead County, Mont., the adjacent county east of LC. Flathead County’s population grew from 51,969 in 1980, to 59,218 in 1990 and to 76,184 in 2002 (Kalispell Chamber of Commerce 2003: 4). In 2003 Flathead County’s median household income was $37,431, while LC’s median household income was $29,262 (USDA Economic Research Service 2006). Flathead County is the recipient of outside-of-the-county investments in single-residen-tial homes, second homes, vacation homes and cabins, and tourism businesses. This economic growth scenario for LC will occur as investors and buyers find the Flathead Valley (Flathead County) too congested and/or expensive. They will seek new opportunities with similar counties. LC will witness land use development driven by natural resource-based amenities and out-of-the-area investments. LC will have an investment surge as new residents, vacationers and/or investors seek its amenities: Beauty, solitude, hospitality, inexpensive development, and present and future public outdoor recreation facilities.

FUTURE RESEARCH QUESTIONS

Based on this study’s brief review of the economic performance in LC from 1969 to 2003, future research should focus on at least four points. First, how much does LC depend on the KNF for its present economic condition? Such a study would include the multiple use spectrum of timber, mining, grazing, and recreation. Economic dependency can be measured and results compared to similar sites and situations. Knowledge of the extent of the general economic dependency is useful to planners, financial investors, and other parties.


A second research focus should be on how Flathead County is already in-fluencing LC’s economic performance. For instance, there is a spending drain from LC to Flathead County as some LC residents drive to the city of Kalispel in Flathead County to shop at their larger stores with highly recognized brand names and greater varieties of goods and services. Many cities and counties on the edge of metropolitan/suburban areas suffer from this same phenomenon. The third research area involves general economic development policies and subsequent actions. Are there certain policies LC could adopt or change that would ensure a better chance of a stronger future economy? Is Flathead County the example to be followed? Concurrently, is there a desire in LC to change? Finally, will a stronger economy change LC’s rural lifestyle and amenities? Will it alter and/or destroy what makes LC attractive to its present residents?

ACKNOWLEDGEMENTS

This work was supported by the South Dakota Agricultural Experiment Sta-tion and approved as Journal Article No. 3575. This study is relevant to those South Dakota counties that are tied to the socio-economic situation in the Black Hills National Forest.

LITERATURE CITED

Bureau of Economic Analysis. Regional Economic Accounts. 2003. Available at: www.bea.doc.gov/bea/regational/bearfacts/action.cfm?fips=30053&areatype=30053&yearin=2003. (verified 22 June 2006)

Frament, Ellen, 2006. personal email, April 25. [email protected] Relocation Guide, Kalispell Chamber of Commerce, @2003, 4. Kootenai National Forest. Social Assessment – 2003 Update. 2004. USDA-

FS. Available at: www.fs.fed.us/rl/kootenai/projects/planning/documents/soc_asses/2003/html/chl/l-2-2.shtml. (verified 22 June 2006).

Kwang-koo, Kim, D.W. Marcouiller, and S. Deller. 2005. Natural Attributes and Rural Development: Understanding Spatial and Distributional Attri-butes. Growth and Change. 36: 2 Spring, 273-97.

Lincoln County Forestry Stewardship Guide. 2006. Rough draft. Rich Lane & Associates, Missoula, Montana. April 4. Available at: www.lincolncuntymn.us/Misc_Pages/Lincoln%20CountyForestStewardshipGuide(Draft%201.pdf (verified 22 June 2006).

Norgaard, K.M. 1997. Timber and Community Well-Being in Siskiyou County: A Socioeconomic Assessment. Available from Kalamath Forest Alliance).

Rumelhart, Paul. 2006. Kootenai River Development, Libby, MT. Telephone conversation, July 6.

Smith, Gary W. 2006. Northwest Income Indicators Project. Available at http://www.Niip.wsu.edu. (verified 14 April 2006).


Stynes, Daniel and Eric White, 2006. Spending Profiles of National Forest Visitors, NVUM. Four Year Report, USDA-Forest Service Inventory and Monitoring Institute, Venture agreement #01-JV-11130149-203, 44 pp. Available at: http://www.prr.msu.edu/stynes/nvumNV4Year.pdf. (verified 22 June 2006).

US Census Bureau: 2000 Census. 2006. Available at: http:// www.epodunk.com/cgi- bin/genInfo.php?locIndex=22327. (verified 22 June 2006).

USDA Economic Research Service. 2006. Available at: www.ers.usda.gov/data/unemployment/RDList2.asp?ST=MT&SF=1D (verified 22 June 2006).

Abstracts of Senior Research Papers

presented at

The 91st Annual Meeting

of the


BACTERIAL SUCCESSION ON THE LEAVES OF PONDEROSA PINE, PINUS PONDEROSA

Crystal Hostetter and David BergmannBlack Hills State University

Rapid City, SD 57701

ABSTRACT

We examined bacterial succession on Ponderosa pine needles in the Black Hills of South Dakota. Samples of living, dead, and decomposing pine needles were collected in 2004-2005. Total genomic DNA was extracted and the 16S rRNA gene was amplified using polymerase chain reaction (PCR), and PCR products separated by denaturing gradient gel electrophoresis (DGGE). We identified the different bacterial 16S rRNA genes collected in the bands of the DGGE gel, each with the potential of being a different bacterial species. The bands were sliced out and each was re-amplified again with PCR, cloned into a plasmid, and transformed into Escherichia coli. Transformed colonies were cultured and DNA extracted for sequencing. Sequences were BLAST searched and results were collected into a tree showing relationships. New samples were collected in summer from mature and young pine needles. Samples were either homogenized or sonicated and then plated out for growth. Samples of phenotypical similarity were isolated, DNA extracted 16S rRNA gene, amplified by PCR, and Restriction Fragment Length Polymorphism s(RFLPs) analyzed. Numbers of bacteria on the surfaces of mature leaves were estimated at 1.3 X l04 colony forming units per g from culturing and 5.1 X l06 cells per g from fluorescence microscopy. Results showed many more species of bacteria (over 40) could be detected through PCR and DGGE than by culturing. Many DGGE bands were likely Sphingomonas species, and Streptococcus, Acidobacteriaceae, and Friedmanella species were also dected by DGGE. One Sphingomonas species and Pseudomonas species were identified from culturing.



SYNTHESIS OF 8-AZA-1,4-OXIDOBICYCLO[4.3.0]NON-2-ENE

ON A SOLID SUPPORT

Nandeo Choony, Jeremiah J. Gums, Amanda Pekny and Mark RanekDepartment of Chemistry

Mount Marty CollegeYankton, SD 57078

ABSTRACT

The research team conducting this experiment attempted to synthesize the cycloadduct 8-aza-1,4-oxidobicyclo[4.3.0]non-2-ene (1) on a solid support resin, triphenylchloromethane polymer (2), involving an intramolecular Diels-Alder cycloaddition reaction. The cycloadduct (1) was isolated from the resin (2) by deprotection with a dilute acid1. The resin was converted back to its chloride by stirring it with concentrated hydrochloric acid in dichloromethane for use again.

ACKNOWLEDGEMENTS

This work was supported by a grant from the South Dakota Biomedical Research Infrastructure Network (SDBRIN) through the National Institutes of Health: NIH-NCRR #2P20-PR01647.


ASSESSMENT OF RANDOM METALBINDING PEPTIDES AS MODELS FOR

ALLOWABLE MUTATIONS IN FUNCTIONALLYIMPORTANT METAL BINDING MOTIFS

Carrissa Pietz and Robert WebbMount Marty College

Yankton, SD

ABSTRACT

The display of random peptide sequences on the surface of bacteria has proven to have many applications in microbiology, biotechnology, and medi-cal research. The amino acid composition and spacing of naturally occurring, metal ion-binding, peptide motifs are under strict structural and functional requirements, whereas randomly derived peptides have the ability to be more variable and exhibit more selectivity and affinity when binding metals. The FliTrxTM Random Peptide Display Library (Invitrogen) was screened for pep-tide sequences with the ability to bind Zn2+. None of the 33 derived sequences showed similarity to known Zn2+-binding proteins, indicating that completely novel Zn2+-binding peptide sequences were isolated. These Zn2+-binding pep-tides could potentially be used for bioremediation purposes or for developing treatments for toxic zinc exposure and metal uptake transport diseases.


FUNGI IN THE DIET OF FLYINGSQUIRRELS (GLAUCOMYS SABRINUS)

CAPTURED FROM THE NORTHERN BLACKHILLS, JUNE TO OCTOBER 2005

A. Gabel, C. Ackerman and M. GabelBiology Department


E. Krueger and S. WeinsU.S. Forest Service

Northern Hills Ranger DistrictSpearfish, SD 57783

ABSTRACT

Flying squirrels (Glaucomys sabrinus) were live-trapped in coniferous/decidu-ous and coniferous habitats in the northern Black Hills. Twenty-five collections of scat were obtained, sixteen from the coniferous/deciduous habitat and nine from the coniferous habitat. Fungi comprised over 90% of the contents. Spores of six genera of hypogeous fungi (truffles and false truffles) in the scat were iden-tified using light and scanning electron microscopy. Rhizopogon was collected throughout the summer and frequency (% of total spores counted) ranged from 88-100%. Gautieria, Geopora, Hymenogaster and Hysterangium were collected throughout the summer at lower frequencies. Elaphomyces was observed late summer to fall. Sporocarps of hypogeous fungi were collected throughout the trapping period to validate spore identifications. Of the thirteen sporocarps col-lected, seven were Rhizopogon. Fewer sporocarps of Elaphomyces, Gautieria, Hys-terangium and Tuber were collected. Maps showing dominant vegetation, slope, aspect and location of the sites were constructed using GIS software. This is the first report of sporocarps of Elaphomyces, Gautieria, Rhizopogon and Tuber, and of Elaphomyces, Gautieria, Geopora, Hymenogaster, Hysterangium and Rhizopogon spores in the scat of flying squirrels from the Black Hills of South Dakota.


VARIATIONS IN RARE EARTH ELEMENT (REE)SIGNATURES AND UNIT CELL DIMENSIONS (UCD)

FOR PURPOSES OF STRATIGRAPHIC CORRELA-TION IN THE PIERRE SHALE, SOUTH DAKOTA

Doreena PatrickGeochemical Solutions, LLC

North Wales, PA 19454

Paul N. WegleitnerGeochemical Solutions, LLC

Fort Pierre, SD 57532

James E. MartinMuseum of Geology

South Dakota School of Mines and Technology Rapid City, SD 57701

ABSTRACT

Our recent research has involved various methods for stratigraphic correla-tion and paleoenvironmental interpretation including Rare Earth Element (REE) analysis of fossil bioapatites and now unit cell dimension analysis (UCDA). REE analysis and UCDA in fossil bioapatite uses variations in REE signatures and unit cell dimensions within fossil bioapatite from various stratigraphic units. The REE composition and UCD are dependent upon conditions of the original dia-genetic waters and thus dependent upon availability of REE and other species for substitution during the per mineralization of the bioapatite. The REE signature and the UCD variations identify distinct intervals within lithologic formations. Because these intervals represent an averaging of periods of certain depositional environments, these distinct intervals can be correlated over significant areas. Fossil vertebrate samples were obtained from the Pierre Shale, in an area be-tween Chamberlain and Pierre, South Dakota, at localities along the banks of the Missouri River in Brule, Buffalo, Hughes, and Hyde counties. Vertebrate samples were collected from the lower, middle and upper Sharon Springs, Gregory, Crow Creek, lower and upper DeGrey and Verendrye members of the Pierre Shale. REE signatures and UCD variations were found to be consistent within indi-vidual lithostratigraphic units but are significantly different between units. REE signatures and UCD act as markers for their units and can be used to discrimi-nate between units for purposes of stratigraphic correlation. The REE analyses and UCDA resulting from our research provides for a finer scale of resolution for stratigraphic correlation and a proxy for paleoenvironmental interpretation.


THE EFFECTS OF HYPOTHYROIDISMON SPONTANEOUSLY HYPERTENSIVE

HEART FAILURE RAT MODELS

Brenda Simon, Bassel Kisso and A. Martin GerdesMount Marty College and Cardiovascular Research Institute

University of South Dakota and Sioux Valley HospitalYankton, SD 57078

ABSTRACT

Recent studies have shown that thyroid dysfunction is an important risk factor associated with heart failure. Clinical case studies suggest that long term hypothyroidism alone can cause heart failure. However, no animal studies have clearly demonstrated this to date. In this study, hypothyroidism was induced in rats with genetic hypertension (SHHF). The major hypothesis being tested is that reduced thyroid function in hypertension will accelerate progression of myocyte remodeling and lead to an earlier onset of heart failure. Hypothyroid induced atrophy of cardiac myocytes and progression of heart failure was assessed in SHHF (spontaneously hypertensive heart failure) rat models. The effects of induced hypothyroidism through the addition of propyl-thiouracil, which blocks the production of thyroid hormone, were investigated on lean female SHHF rats. Left ventricular function was determined by echo-cardiography and hemodynamics. Whole tissue pathology and isolated myocytes size and number were assessed. PTU treatment caused an increase in LV diastolic chamber diameter of 14%, and an increase in LV systolic chamber diameter of 38%. Heart rate decreased by ~100 beats per minute, left ventricular pressure decreased by ~50mmHG, and there was a decrease in the measurement of contractility over time. Both body weight and ventricular weight decreased. Wall stress increased by 19%. PTU treatment of SHHF rats resulted in clinical hypothyroidism which re-sulted in hastening the onset of heart failure. Left ventricular function declined, and myocytes were lost.


GENE EXPRESSION IN ECOLOGICALLYMEANINGFUL CONTEXTS:

EVOLUTION OF PLANT DEFENSESIN COMPETITIVE ENVIRONMENTS

David H. Siemens, Riston Haugen and Lexi SteffesBiology Department

andRichard Gayle

Mathematics Department


ABSTRACT Plants in the wild that are attacked by herbivores and pathogens often grow next to other plants that represent potential competitors. Therefore, in some cases one would expect the simultaneous evolution of defense and competitive ability. However, the optimal defense hypothesis, currently the best framework we have to understand the simultaneous evolution of defense and competitive-ness, predicts a tradeoff between these factors. Plants that effectively grow and compete well against neighbors are expected to have lower defense levels either because growth diverts limited resources away from defense production, or be-cause less valuable tissue that is consumed is more readily replaced. In contrast to these predictions, some recent studies have found evidence that some plant species may be able to simultaneously compete and defend effectively. One hy-pothesis for this result is that some defensive traits, such as toxin concentration, have dual functions in defense and competition. In two growth room experiments in which we examined transcript profiles of a close wild relative of Arabidopsis thaliana, we tested (1) whether neighboring plants elicit defense responses, and (2) whether there was overlap in gene expres-sion patterns between herbivory and competition treatments that were either complementary or antagonistic. In one experiment involving three treatments (herbivory, competition, control) there were over 900 significantly differentially expressed genes, and evidence that competition elicited genes with know func-tion in defensive pathways.


THE FIRST UNEQUIVOCAL NORTHAMERICAN OCCURRENCE OF THE MOSASAUR

HAINOSAURUS (REPTILIA) FROM THECRETACEOUS PIERRE SHALE OF THE MISSOURI

RIVER TRENCH, SOUTHERN SOUTH DAKOTA

James E. Martin and Wayne A. ThompsonMuseum of Geology

Department of Geology and Geological EngineeringSouth Dakota School of Mines and Technology

Rapid City, SD 57701

David C. ParrisNatural History Office

New Jersey State MuseumTrenton, NJ 08625

ABSTRACT

One of the largest reptiles to occur in the seas during the end of the Cre-taceous was the mosasaur, Hainosaurus. The mosasaur attained lengths of over 10 meters with a meter-long skull and is a tylosaurine mosasaur, a top preda-tor. Hainosaurus bernardi was originally described in 1885 from Maastrichtian deposits in Belgium based upon a partial skeleton and a second skull. Other reports of Hainosaurus from the Old World have occurred, but its occurrence in North America has been questioned. The first report of the genus in North America, Hainosaurus pembinensis was from the early Campanian Pembina Member of the Pierre Shale in Canada. Later workers have doubted the assign-ment of the Canadian taxon to Hainosaurus, and many authors have considered the taxon as a species of the closely allied genus, Tylosaurus. A number of char-acters have been utilized to define Hainosaurus, but many have been found in other taxa. Perhaps the most definitive character of Hainosaurus is that the pineal opening is shared by the frontal and parietal bones, rather than being confined to the parietal. Until now, no other North American mosasaur, including the Canadian taxon, exhibits a clearly shared pineal opening. Now, a new specimen found by Mr. Paul Neumiller from the late Campanian DeGrey Member of the Pierre Shale along the shore of the Missouri River in Gregory County exhibits the shared pineal opening. In addition, other characters suggested as characteris-tic of Hainosaurus including relatively smooth surfaces on the teeth, interlocking premaxillary-maxillary suture, and a prominent internal process of the suprasta-pedial process of the quadrate occur on the newly discovered specimen. Other apomorphies suggest that the specimen represents a new species. Evidently, tylosaurine specimens from the early Campanian exhibit the initial divergence of Hainosaurus from Tylosaurus, and by the late Campanian and Maastrichtian, the transition had been completed.


WEST NILE VIRUS INFECTION RATES ANDCULEX TARSALIS POPULATION DYNAMICS

AT A FARM SITE IN THE JAMES RIVER VALLEYOF EAST-CENTRAL SOUTH DAKOTA

Ryan J. Beyer, Clayton J. Wulf, Jacob S. Schaeffer,Mitchel M. McKenzie and Michael B. Hildreth

Departments of Biology/Microbiology and Veterinary ScienceSouth Dakota State University

Brookings, SD 57006

Anne N. Rounds, Ragna A. Godtland,Rachel.A. Hoffman and Christopher D. Carlson

South Dakota Public Health LaboratoryPierre, SD

ABSTRACT

In 2005, 46.5% of the West Nile Virus (WNV) cases in South Dakota were located in the 9 counties bordering the James River. During that summer, a rural site was selected in Beadle County, South Dakota to monitor the summer build-up of WNV-infected Culex tarsalis populations within the James River Valley. From June-September, adult mosquitoes were collected biweekly using a CDC miniature light trap baited with carbon-dioxide. Culex tarsalis were separated and sent to the S.D. Department of Health for WNV testing. WNV RNA was extracted using an RNA III Isolation kit. The RNA was detected by RT-PCR on a LightCycler platform. Five potential larval habitat sites were selected near the farm-site, and.dip samples collected weekly in sites contain standing water. The farm site showed a predominance of Aedes vexans adults during the beginning of the summer, peaking at 6,216 adult females on June 21. By July 7, the number of C. tarsalis exceeded that of Ae. vexans, a trend that continued to the end of the summer. Culex tarsalis adult females peaked on July 21, and dropped below 500 by the second week of August and below 20 by September 9. Ten of the 11 WNV-infected mosquito pools occurred from the July 14 to August 4. No Cx. tarsalis larvae were found until June 14, but it became the predominant species in late July through late August. The highest number of Cx. tarsalis larvae were collected on July 19 and September 1. The farm-site contain numerous tires, but only 1 male Cx. tarsalis larvae was found in them during the summer.


ARCHAEA ASSOCIATED WITHTHE PORCINE ILEUM OF WEANED PIGS

Clayton L. Scofield and V. BrözelDepartment of Biology and Microbiology

S. Lindblom, A. Rosa, S. Vilain and S. GeorgeDepartment of Animal and Range Sciences

R. Kaushik and D. FrancisDepartment of Veterinary Science


ABSTRACT

The porcine digestive tract harbors copious bacterial flora which plays an important role in digestion. While much is now known about the eubacterial diversity present, little is known about the prevalence and diversity of Archaea. The aim of this project was to determine the diversity of the archaeal microflora present in the ileum using 16S rDNA sequences as indicators of diversity. Two five week old weaned pigs were chosen at random. Genomic DNA was extracted from ileal lumen and the PCR was done with Archaea-specific primers A571F and UA1204R. Amplicons were cloned in pGEM-T Easy, transformed into E. coli JM109 and white colonies selected at random. The DNA sequences were checked for chimeras using the Pintail program. After authentication, sequences were aligned using ClustalW and grouped according to relatedness into a phylo-genetic tree. The sequences of the closest related known species were also added to the tree. All archaeal amplicons were related to methanogens and no chimeras were found. Some of the PCR amplicons were not of archaeal origin; however, some sequences indicated pig or human DNA, showing that the primer set was not specific for Archaea. Those sequences were not included in the tree. The result positioned the sample sequences into three clusters. Cluster One grouped closely with Methanobrevibacter ruminantium, Cluster Two with M. sp. ZA-10 and SM9 and M. gottschalkii, and Cluster Three with M. smithii strain PS. These results indicate that the Archaea present in the ileum of young pigs are largely members of the genus Methanobrevibacter.


LECTIN BINDING PROFILE ON THESMALL INTESTINE OF 5-WEEK OLD PIGSIN RESPONSE TO USE OF ANTIBIOTICS

AS GROWTH PROMOTANTS

Sajan George, Yejin Oh, Sebastien Vilain and Volker BrözelDepartment of Biology and Microbiology

Stacy J. Lindblom and Artur J.M. RosaDepartment of Animal and Range Sciences

David Francis and Radhey S. KaushikDepartment of Veterinary Science


ABSTRACT

Antibiotics are traditionally used for growth promotion in the pork industry; however, their use in animal feed has recently been banned in Europe because of the potential threat to human health. The intestinal microbiota together with the diet and gut epithelium comprises a complex ecosystem and plays an impor-tant role in mediating many physiological functions such as digestion/absorp-tion which ultimately effects the growth performance and health of the animal. Lectins are carbohydrate binding proteins having specific affinities for accessible sugar residues on the cell membranes. The presence or absence of certain carbo-hydrates on the gut epithelium may influence the growth of host and suscep-tibility to different pathogens. The aim of this study was to identify the lectin binding profile on the small intestine of weanling pigs in response to feeding with chlortetracycline or germ free conditions. Eighteen half-sib piglets obtained by caesarian were divided into three groups (n=6) and maintained as antibiotic fed, control and gnotobiotic until 5-weeks of age. Glycoconjugate composition on the ileal surface was examined by histochemistry using 23 biotinylated lectins. Results obtained from three animals from each group were analyzed and data showed that the lectins DBA, RCA120 , GSL-II, DSL, LEL, STL, WGA, s-WGA, Jacalin, PSA, LCA, SNA and MAL bound differentially to both villus and dome epithelium. Crypt epithelial cells showed varied affinity for STL while domes showed differences for WGA, PSA, LCA and MAL. Corona and follicles showed differential staining for GSL-I, STL, WGA, Jacalin, PSA, LCA and MAL. These findings will enable us to further understand the role of gut microbiota and mechanisms of action of antibiotics as growth promotants in pigs.


PATHOLOGY AND DISSEMINATIONOF ENCEPHALITOZOON INTESTINALIS

IN GNOTOBIOTIC PIGLETS

Gopakumar Moorkanat, Aaron F. Harmon and YeJin OhDepartment of Biology and Microbiology

Larry Hollar, David Francis, Michael B. Hildreth and Radhey S. KaushikDepartment of Veterinary Sciences


ABSTRACT

The pathology and dissemination of Encephalitozoon intestinalis, an oppor-tunistic infection occasionally associated with AIDS patients, was studied using gnotobiotic piglets as a possible animal model for human infections. Piglets were separated (n=3-4) into control (C), infected (I), and immunosuppressed/infected (S/I) groups and were maintained under germfree conditions. Piglets in the S/I group received 15 mg/Kg Cyclosporine oral solution and 25 mg/Kg Methylpred-nisolone acetate intramuscularly daily starting at 2 days. At day 3, each piglet from the infected and S/I groups were inoculated with approximately 0.75 X 106

E. intestinalis spores. Various tissues were harvested 3-4 weeks post-infection and processed for histological examination and PCR analysis. DNA was extracted in a bead beater with ceramic beads and purified with Qiagen’s DNeasy plant tissue kit. PCR was performed using E. intestinalis specific primers. The 127bp prod-uct was visualized on a 3% agarose gel. Not all of the samples have been analyzed yet, but patterns are emerging. At 5 days post infection, PCR results from fecal samples showed that 2 of 4 piglets from the infected group were shedding E. intestinalis spores; whereas, all 4 control piglets were PCR negative. Some of the tissues (e.g. ileum, lung and brain) from the S/I piglets tested PCR positive for E. intestinalis. Thus far, E. intestinalis had not been found in histological sections of the S/I tissues. These findings indicate that E. intestinalis was infectious to gno-tobiotic piglets even without immunosuppression and extraintestinal infections of E. intestinalis could be found in immunosuppressed piglets. Thus gnotobiotic piglets may be used as a large animal model for studying human E. intestinalis in further studies.


KINETICS OF QUATERNARY AMMONIUM METHYL CARBONATES WITH DIFFERING CHAIN LENGTHS

Eric M. Villa, Jessica L. Paumen, Gary W. Earl and Duane E. WeisshaarDepartment of Chemistry


ABSTRACT

Quaternary ammonium compounds have a wide variety of household and commercial uses; therefore, the methylation of tertiary amines is an extremely important reaction in the chemical industry. Currently, this process involves hazardous materials like dimethyl sulfate and/or methyl chloride. This research investigates the use of dimethyl carbonate as a green methylating agent. Specifi-cally, it deals with the kinetics of this SN2 reaction, using three different tertiary amines: tributylamine, trihexylamine and trioctylamine. The reaction was run in a multi-vessel pressure reactor and the ratio of quaternary ammonium methyl carbonate to that of unreacted amine was determined by High Performance Liq-uid Chromatography. From this data, the rates of reactions and the activation energies have been determined and will be presented.


DIABETES POPULATION STUDIES IN THE PINE RIDGE NATIVE AMERICAN RESERVATION

Patricia Hammond, Chance Weston, Deig Sandoval and Durga MittintiOglala Lakota College

Kyle, SD 57762

ABSTRACT

Diabetes population studies were conducted district wise in the Pine Ridge Native Indian reservation in South Dakota. From the Data provided by the IHS (Indian Health Service) Pine Ridge office, it was observed that Wounded Knee district showed the highest percentage of diabetes registered cases at 18% population. This is a standardized percent by the district total population. Next to it the Pine Ridge district, 17% shows a incidence of registered cases in the res-ervation. In general, the percentage (%) level of diabetes population (9%) in the reservation is higher than the National average of the country (6%) as a whole. The females consistently showed a higher percentage (% ) registered cases over the males and it is suggested that the root cause of this finding needs further investigation. In general, this trend is unusual compared to the US trend in which males have a higher tendency to diabetes compared to the females. Amongst the age groups of 0-19, 20-39 and 40+ , the age group of 40+ showed a higher percentage (%) incidence of diabetes cases. In this age group also, the females are in a higher percentage ( %) over the males.


STATISTICAL SOFTWARE FORTIME SERIES ANALYSIS

Paul Marshall and Daniel SwetsAugustana College

Sioux Falls, SD

ABSTRACT

Images derived from satellite sensors can provide a unique vantage point for studying global seasonal dynamics that have implications for global change issues, such as urbanization, drought monitoring, habitat changes, and land cover/land use changes. These two-dimensional spatial images are taken at regularly spaced, points in time, making a three-dimensional product that can be used for study-ing changes at a regional, national, or global scale. Trend analysis of temporal images is a fundamental tool in these studies. Statistical software, such as S-Plus and SPSS provide for sophisticated statistics on lower dimensional datasets, but are limited in their ability to provide for trend analysis on these stacks of two-dimensional images. The data volume and the format of the satellite imagery make adapting the commercial software packages untenable. On the other hand, sophisticated image processing software such as ENVI or Imagine lack the ca-pability of sophisticated statistical tests. Statistical tests, like the Mann-Kendall test, provide a non-parametric approach to analyzing trends, such as increasing or decreasing trends in plant production as estimated from satellite imagery. In this project, we developed software to meet this need.Continuity of vegetation index data derived from moderate resolution satellite imagery is critical to providing a context to compare data from future missions. The software contributed to the assurance that the transition from satellite sys-tem to system will be transparent, i.e., there will be no abrupt shifts in the data values collected from vegetation imagery between the sensors.


POLYMERIZATION AND COPOLYMERIZATIONOF LINSEED OIL WITH STYRENE

AND DIVINYLBENZENE

Jay D. Heeren and Timothy R. HightowerBlack Hills State University

Spearfish, SD

ABSTRACT

Currently, the vast majority of commercial polymers are synthesized from polyunsaturated petroleum products. Due to the limitation of these petroleum products, much work is needed to develop future polymers that can be derived from renewable natural resources. The application of linseed oil as a renewable natural monomer, as well as the copolymerization of linseed oil with various amounts of styrene and divinylbenzene, have provided a variety of bioplastics with diverse physical properties ranging from hard to rubber-like plastics.


ISOMERIC SERIES OF NOVEL CYANATEESTERS AND POLYCYANURATE RESINS

Josiah Reams, Tsvetanka Filipova and David A. BoylesSouth Dakota School of Mines and Technology

ABSTRACT

A need exists for new materials having enhanced thermal and dimensional stability over currently used materials in the electronics and aerospace industries. Synthetic routes have been developed and will be presented that detail the syn-thetic procedures and characterization of intermediates and products for these new cyanate esters, namely o-tetraaryl bisphenol A dicyanate, m-tetraaryl bisphe-nol A dicyanate, and p-tetraaryl bisphenol A dicyanate. These resin precursors have been designed with the goal in mind of incorporating greater length-to-width ratios into their molecular structure. The synthesis of an isomeric series may afford compounds for the development of structure-property relationships. The synthesis and characterization of these novel cyanate esters and characteriza-tion of the thermally polymerized thermoset resins will be presented.


POLYETHERIMIDES INCORPORATINGHIGH ASPECT RATIO BISPHENOLATE:

MONOMER SYNTHESIS AND POLYMERIZATION

Kalub Hahne, Tsvetanka Filipova, and David BoylesSouth Dakota School of Mines and Technology

Chemistry DepartmentRapid City, SD 57701

ABSTRACT

Aromatic polyetherimide is well known as one of the good materials for use in high temperature applications due to its aromatic and heterocyclic structure. The incorporation of flexible aromatic ether linkages in polymer chains provide for lower softening and improved solubility. General Electric has been reported the new convenient synthetic method for preparation of polyetherimides by ni-tro displacement of activated aromatic nitro groups from disubstituted bisimides with dianion of bisphenols. Soluble polyetherimides synthesized by direct solu-tion polymerization of bisnitroimide monomer and bisphenolates were obtained. Bisnitroimide monomer was prepared in three steps reactions started from phthalic acid. High aspect ratio TABPA and AsyBPA were used to prepare the bisphenolate. The monomers were characterized by their percentage yield, melt-ing points, FTIR and NMR spectra. The thermal behavior of the synthesized polymers was evaluated by deferential scattering calorimetry.


EFFECT OF BUCKTHORN (RHAMNUSCATHARTICA) ON THE TREE COMMUNITY OF

AN EASTERN SOUTH DAKOTA WOODLOT

Dale L. Droge and Javeria JavedCollege of Arts and Sciences

Dakota State UniversityMadison, SD 57042

ABSTRACT

Buckthorn (Rhamnus cathartica) is a non-native woody shrub that invades woodlands and swamps of many regions of the northern United States. Stud-ies in other states have shown that Buckthorn crowds the forest understory and greatly affects the recruitment of native tree species. We examined the frequency and dominance of tree species in a 7 hectare woodlot located within a Waterfowl Production Area in Lake County, South Dakota. We summarized data collected by ecology classes at Dakota State University over a 10 year period. Tree abun-dance and size were estimated using point quarter plotless sampling procedures. Over the 10 year time period, Buckthorn replaced Green Ash (Fraxinus pennsyl-vanica) as the most abundant tree species. Because Green Ash trees have larger diameters, it is still the most dominant species. However, Buckthorn dominance is increasing through abundance and growth in basal diameter over time. Young Buckthorn trees are increasing, but very few small diameter native trees are found. We show that the presence of Buckthorn will have dramatic effects on the future species composition of the woodlot through a decrease in the replacement of the large diameter native species such as ash and hackberry (Celtis occidentalis). Minnesota and other nearby states have instituted plans to remove existing popu-lations and control the spread of Buckthorn. Our data demonstrate that South Dakota also needs to take action before the tree diversity of forests, shelterbelts and woodlots is severely reduced.


GROWTH AND PRODUCTION OF DIFFERENTVARIETIES OF BITTER GORT MILLON

Al Eastman, Brian Danner, Deig Sandoval and Durga MittintiOglala Lakota College

Kyle, SD 57762

ABSTRACT

Optimization of parameters were carried out in the growth and production of different varieties of Bitter Gourd Melon especially i) Surface Temperature ii) Soaking time for the seeds and iii) to determine the rate of plant growth (length of stem). Bitter Melon Balsam Pear, Bitter Gourd Baby Doll (three replicas of each) had all been tested and it was found that around 21 deg C of surface temperature and around 23 deg C, at a depth of 1 1/2” internal soil temperature. The opti-mum relative humidity was found to be 70% for the germination and growth of these plants. The soaking time for the seed had a significant effect on the yield of Balsam Pear leaves (8.33 gms per seed) as compared to Baby Doll (5.33 gms per seed) and Wild Puerto Rico (no yield). Both Balsam Pear and Baby Doll had been planted (6 seeds/variety) and the time of soaking in water was 30 hrs and both had germinated 5 plants each. Around 10 plants of the Wild Puerto Rico that had been soaked for 24 hrs, only three germinated. The Baby Doll showed that its stem has grown 2.5 cm per day, on the aver-age. The Balsam Pear showed a growth of 1.5 cm per day.


EVIDENCE FOR ENAMELOIDIN XENACANTHID SHARK TEETH

Wendy Stiernagle and Gary D. JohnsonDepartment of Earth SciencesUniversity of South Dakota

Vermillion, SD 57069

ABSTRACT

The Xenacanthida (Chondrichthyes: Elasmobranchii) are characterized in part by the lack of enameloid in their teeth. The cusps are composed of ortho-dentine, sometimes with an outer layer of pallial dentine which is penetrated by tubules from the orthodentine. In 2003, the junior author described a new late Paleozoic genus, Barbclabornia, with one species, B. (Xenacanthus) luedersensis that is endemic to North America. The cusps were described as containing “hy-permineralized pallial dentine” in the cristae, based on the presence of dentinal tubules which penetrate only slightly into the outer layer, unlike pallial dentine. Also in 2003, in a paper reviewing the Xenacanthida from the Carboniferous of the British Isles, it was stated that “X.” luedersensis tooth cusps contain enameloid, based on thin-section analysis, and therefore may not be a xenacanthid. Enameloid contains dentinal tubules that extend into that tissue from the orthodentine. True enameloid also contains crystallites which are not visible in thin section and are only made evident by scanning electron microscopy (SEM). Because of the taxonomic implications, the authors analyzed the cusps of Barbclabornia luedersensis with SEM, and also those of two species of Orthacan-thus, a well known xenacanthid genus, with serrated (O. texensis) and smooth (O. platypternus) carinae. All Orthacanthus teeth lack enameloid. The orthoden-tine, as seen with SEM and in thin-section, is similar in all three species. Identi-fication of the tissue in the Orthacanthus carinae on the basis of SEM is equivo-cal, but more significantly, the tissue in the B. luedersensis cristae is considerably different. However, when compared to published descriptions and illustrations of the crystallites in the enameloid of euselachian sharks, it is evident that the presence of crystallites in B. luedersensis teeth cannot be established. The “texture” of the tissue in the cristae has a coarse granular appearance as seen with SEM; its classification, i.e., hypermineralized dentine or some new form of enameloid, remains unknown.


AN INVENTORY OF AMPHIBIANS AND REPTILES OF THE BLACK HILLS OF SOUTH DAKOTA

Laurelin Cottingham and Brian E. SmithDepartment of Biology


ABSTRACT

The reptiles and amphibians of the Black Hills have been poorly studied. The last complete inventory of the Black Hills herpetofauna was completed in 1974. We conducted a herpetofaunal inventory of the Black Hills of South Dakota in 2004 using a survey method incorporating GIS technology that we developed to identify historically understudied areas of the Black Hills. A species occurrence index (SOI) was calculated for each sampling grid using historical data and expected occurrence of each species in each grid. This index was used to determine which grids were most poorly known herpetologically. This method allowed us to rank sampling grids on a priority basis, and we apportioned our efforts accordingly. Much of the study focused on the highest priority sites in the southern Black Hills and areas surrounding the uplift of the Black Hills. We found that the herpetofauna of the interior Black Hills had been reasonably well-sampled. We documented 1.58 specimens/hour in 380 hours of surveys. This is higher than other herpetofaunal inventories conducted in the Great Plains region. More importantly, the SOI for the Black Hills overall was increased by 11.4 percent.


GENETIC VARIANCE IN THE SMOOTHGREEN SNAKE, OPHEODRYS VERNALIS,

IN SOUTH DAKOTA

Laurelin Cottingham, Brian E. Smith, Cynthia Anderson and Shane SarverBlack Hills State University

Center for the Conservation of Biological ResourcesSpearfish, SD 57799

ABSTRACT

The smooth green snake, Opheodrys vernalis, is a wide-ranging species found throughout much of the northeastern United States and southeastern Canada. Twenty-four isolated populations are found throughout the Midwest, northern Plains, and Rocky Mountains of the United States. It appears to have declined throughout much of its range and is now protected in Indiana, Missouri, Mon-tana, North Carolina, Wyoming, and Utah. There are isolated populations in the Black Hills and Bear Lodge Mountains of western South Dakota and northeastern Wyoming, as well as a population on the northeastern plains of South Dakota. This species has been the subject of taxonomic debate, includ-ing controversies over generic classification and subspecific variation. We have sequenced the d-loop region of the mtDNA and used the Basic Local Align-ment Search Tool at NCBI’s website to examine similarity with other Opheodrys species. For this study we are examining genetic variation among the plains population in South Dakota and populations in the Black Hills and Bear Lodge Mountains. Additionally, we are comparing genetic variation in these popula-tions with specimens collected in other parts of the species range. Sequence data obtained from PCR amplified D-loop, ND4, ND2 and cytochrome B regions of mitochondrial DNA will be used for genetic comparisons. Microsatellite mark-ers are being developed for further studies.


EUBACTERIAL DIVERSITY OFTHE PORCINE ILEUM OF WEANED PIGS

L. Weyrich, C. Scofield and V. BrözelDepartment Of Biology And Microbiology

S. Lindblom, A. Rosa, S. Vilain and S. GeorgeDepartment Of Animal And Range Sciences

R. Kaushik and D. FrancisDepartment Of Veterinary Science


ABSTRACT

The porcine digestive tract supports the growth of a range of bacteria, with numbers increasing from the jejunum through the ileum to the colon. This diverse bacterial community is believed to contribute to the growth and health of young pigs. The aim of this project was to determine the diversity of the eubacterial microflora present in the ileum of weaned pigs using 16S rDNA se-quences as indicators of diversity. Two five week old weaned pigs were chosen at random. Genomic DNA was extracted from the contents of ileal lumen and the PCR was done with primers S-D-Bact-0008-a-S-20 and S-*-Univ-1492-a-A-19 specific for the Eubacterial domain. Amplicons were cloned in pGEM-T Easy, transformed into E. coli JM109 and white colonies were selected at random. The DNA sequences were checked for chimeras using the Pintail program. After authentication, sequences were aligned using ClustalW and grouped according to relatedness into a phylogenetic tree. The sequences of the closest related known species were also added to the tree. Several amplicons were found to be chimeric in origin and were therefore discarded. The community comprised largely of members of the low G+C bacteria, including various Lactobacillus and Clostridium. None of the Clostridium sequences obtained clustered closely with known pathogens. Other sequences obtained included Terracibacter and entero-bacteria.


A STUDY OF THE PRION PROTEIN (PRP) GENE: THE EVOLUTIONARY HISTORY AND SERIAL

TRANSMISSION TO UNRELATED SPECIES

Forrest Cain, Cynthia Anderson and Shane SarverBlack Hills State University

Spearfish, SD 57799

ABSTRACT

A mis-folded form of the prion protein encoded by the PRNP gene is the cause of Transmissible Spongiform Encephalopathy or TSE. Some of the more familiar forms are Bovine Spongiform Encephalopathy, also known as Mad Cow Disease in cattle, Chronic Wasting Disease or CWD in deer and elk and Scrapie in sheep. TSE’s are fatal neurodegenerative diseases that cause damage to neurons of the CNS and eventually lead to death. The main goal of this project was to PCR amplify and sequence the coding region of PRNP, from several local animal species including, coyote, badger, skunk, bobcat, raccoon, lynx and pronghorn antelope, most of which are carnivores. The reason for using carnivores in the project is that they regularly feed on animals that are known to contract the fatal disease. By understanding the similarities in the protein sequence encoded by PRNP we may gain insight into the possibility of serial transmission of the dis-ease. The DNA sequence data and the derived amino acid sequence data were analyzed to determine the similarity among the various taxa represented. Results from this analysis will be presented.


COST AND BENEFITS OF COMPOUNDSFUNCTIONING AS A DEFENSE AGAINST

HERBIVORES IN BOECHERA STRICTA

Shane Ziegenbein, Lexi Steffes and David SiemensBlack Hills State University

Spearfish, SD 57799

ABSTRACT

Plants have supposedly evolved chemical compounds that function in defense against pathogens and herbivores; however there exists a tremendous amount of variation in the concentration of these compounds within and among plant populations. The optimal concentration is thought to be a function of costs and benefits associated with the production of these compounds, which may vary over resource gradients. A recent cost/benefit optimization model predicts multiple optima: low and high stable optima are predicted at low and high resource availability respectively, but as resources vary across intermediate levels stable optima are not predicted. Instead, as resources vary from low to high, concentrations are predicted to jump from low to high levels. To test these predictions we conducted two experiments. In both experiments we grew 32 maternal families of the self-fertilizing Boechera stricta (Brassicaceae) with varying amounts of fertilizer. In the first experiment we used a combination of fertilizers, which resulted in a non-linear relationship between growth and fertilizer levels. In the second experiment we simplified the fertilizer treatment, which resulted in the desired linear relationship. Glucosinolate toxins were then measured in both experiments. In the first experiment there was little support for the predicted pattern, and genetic variation did not obscure this result. The results from the second experiment will be presented.


CHARACTERIZATION OF PARA-ASSYMETRICBISPHENOL A POLYCARBONATE BY GEL

PERMEATION CHROMATOGRAPHY

Nathan Roark, Tsvetanka S. Filipova and David A. BoylesDepartment of Chemistry

South Dakota School of Mines and TechnologyRapid City, SD 57701

ABSTRACT

para-Asymmetric bisphenol A polycarbonate (p-AsBPA PC) was synthesized by interfacial polycondensation of an asymmetric triBPA monomer. The mono-mer was prepared from commercially available BPA. Gel permeation chrom-atrography was used for determination of molecular weight (Mw) and change in refractive index per change in sample concentration (dn/dc). dn/dc Values are of particular interest in polymer science as they must be calculated precisely so that accurate values for Mw and other parameters are derived. Differential scanning carolimetry was used to calculate glass transition temperature.


COMPARISON OF CAPTURE METHODSAND HOME RANGE OF WHITE-TAILED

JACKRABBITS IN SELECTED FIELDSIN EASTERN SOUTH DAKOTA

Dustin Schaible and Charles DieterDepartment of Biology and Microbiology


ABSTRACT

The white-tailed jackrabbit (Lepus townsendii) typically inhabits short mixedgrass prairie where it fills an important niche throughout its range. In many areas, jackrabbits serve as an important food item for several species of predators. Capture methods and home range size of white-tailed jackrabbits has received little attention in the literature and has not been reported in South Dakota. We used live traps and a drive corral to compare capture rates of jack-rabbits in three selected fields in Brookings, Kingsbury, and Beadle counties in eastern South Dakota. There were 18 white-tailed jackrabbits captured using live traps with an additional 16 captured using the drive corral. We averaged 0.04 jackrabbits per trap night and 0.29 jackrabbits per man hour using live traps. Using the drive corral, we averaged 1.23 jackrabbits per trap attempt and 0.48 jackrabbits per man hour. We used the fixed kernel estimator to calculate home range size which ranged from 0.40km2 to 4.76km2. Over all sites, female aver-age home range size (1.35km2) did not differ from average male home range size (1.34km2).


GEOMETRIC LAYERED TRIANGULATIONOF LENS SPACES

Luc PatryNorthern State University

Aberdeen, SD 57401

ABSTRACT

Layered triangulations of lens spaces, L(p,q), are realized by tetrahedra whose boundaries are totally geodesic . The structure of the two-skeleton of each lens space which completely determines the space, is described using a path in a Stern-Brocot like binary tree. Nodes of the tree are labeled by members of SL2(N). To each node correspond 2 lens spaces whose parameters are of the form 2a+b where a and b are components of the label of the node. Further the com-ponents of the nodes on the path from the root of the tree to the node of a lens space label correctly generators and relations corresponding to the faces of the tetrahedron attached to the node independently of the node of the lens space.


PHOSPHORYLATION OF eIF-4E REGULATES p53 PROTEIN SYNTHESIS FOLLOWING DNA DAMAGE

Ying Zhang and Da-Qing YangDivision of Basic Biomedical Sciences

University of South DakotaSanford School of Medicine


ABSTRACT

The eukaryotic translation initiation factor 4E (eIF-4E) is essential for ef-ficient cap-dependent protein translation. eIF-4E is known to be regulated by its inhibitory binding proteins (4E-BPs) and by its own phosphorylation. However, the role of eIF-4E phosphorylation in protein translation is still un-clear. The tumor suppressor protein p53 plays a critical role in suppressing cell transformation and maintaining genetic integrity. These functions are achieved by accumulation of p53 protein after DNA damage and by p53-induced activa-tion of genes that mediate either cell growth arrest or cell death. Although there is clear evidence indicating that p53 induction is regulated by protein synthesis following DNA damage, the mechanism for the translational regulation of p53 is poorly understood. Our results show that etoposide treatment caused a rapid increase in eIF-4E phosphorylation as well as increased p53 protein synthesis. The addition of CGP57380 (CGP), a specific inhibitor of the eIF-4E kinase Mnk1, not only inhibited eIF-4E phosphorylation but also resulted in reduced synthesis of p53 protein. 35S-labelling experiments further demonstrated that the accumulation of p53 protein was accompanied by an increase in the de novo p53 protein synthesis following etoposide treatment, whereas addition of CGP led to a decrease of p53 induction. Our findings provide the first evidence that phos-phorylation of eIF-4E by Mnk1 is critical for increased p53 protein synthesis in response to DNA damage.


THE IDENTIFICATION OF AN INTERNALRIBOSOMAL ENTRY SITE IN THE

5’-UNTRANSLATED REGION OF p53 mRNAPROVIDES A NOVEL MECHANISM FOR

THE REGULATION OF ITS TRANSLATIONFOLLOWING DNA DAMAGE

Marie-Jo Halaby, Ying Zhang, and Da-Qing YangDivision of Basic Biomedical Sciences

University of South DakotaSanford School of Medicine


ABSTRACT

The tumor suppressor p53 plays a crucial role in maintaining the genetic integrity of the cell and in suppressing cell transformation. Its cellular levels are usually low and rise substantially in response to DNA damage. Although research on p53 induction following DNA damage has mainly focused on the posttranslational modification of p53 by Mdm2, it is known that protein trans-lation also contributes to p53 induction. However, the mechanisms underlying translational regulation of the p53 protein following DNA damage are still un-clear. We show that p53 synthesis increased dramatically in MCF-7 cells treated with etoposide. Interestingly, this increase was accompanied by an increase in the association of the translation initiation factor eIF-4E with its binding pro-tein 4E-BP1, an inhibitor of cap-dependent protein translation. We further identified an internal ribosomal entry site (IRES) located in the 5’-UTR of the p53 mRNA, that was capable of driving the cap-independent translation of a downstream cistron encoding Firefly luciferase in a dicistronic expression vector. Moreover, we found that the activity of the IRES element increased in response to etoposide-induced DNA damage in MCF-7 cells. Our results provide a novel mechanism for the regulation of p53 translation in response to DNA damage.


BIG FOOT RIDE

Sandra Byrd, Brandon Ferguson and Sylvio MannelOglala Lakota College

Martin, SD 57551

ABSTRACT

Tradition and Technology: Using Waypoints, GPS, and GIS to retrace and document the historical Big Foot Ride. The Big Foot Ride commemorates the route of the Minneconjou Lakota people who fled the Standing Rock Reservation in the 1890’s. The original memorial ride began in 1986. The ride was conducted in accordance with an Oglala Lakota medicine man’s vision to release the spirits and complete the griev-ing cycle for the victims of the Wounded Knee Massacre. Before the ride was retraced, it is said that a ceremony was held which revealed that a horse should be blindfolded. It was the blindfolded horse who led the riders through the original path that the fleeing Lakota took over 100 years ago. The route of this past year has changed from the original path due to un-foreseen circumstances. Further work is needed to map the original route and provide alternative routes to honor the traditional path as well land-ownership realities. An experiment was conducted during the December 2005 Big Foot Ride. One of the riders, Brandon Ferguson, used a Global Positioning System, a Garmin 72 unit, to mark waypoints on the trail. The waypoints were then downloaded by a GIS student, Sandra Byrd. Geographic Information System, ArcGIS 9, was used to transform the waypoints into a route and these processes were combined to create a map of the trail. A poster was generated using Oglala Lakota College’s GIS Lab 44 inch plotter.


LAKOTA LAND—MAPPING CULTURE,HISTORY AND RECREATION

Charles Comes Killing and Sylvio MannelOglala Lakota College

Kyle, SD 57552

ABSTRACT

In this pilot project we locate Lakota historical, cultural and recreational areas. Lakota Land includes establishing a geodatabase, setting up an online interactive map, and investigating links of Native sites with geospatial features, such as water or mountains. For the Lakota Land project we a) preserve sites from destruction, b) ad-vance the understanding of geospatial technology to a broad Native American public c) foster Native identity, d) provide a tool for tribal decision making, e) educate Native and non-native communities, f ) provide a database for research-ers of Native culture/history g) offer alternatives to socio-economic problems and h) research geospatial patterns to identify additional sites. This project has a broad impact. It already involves the community, tribal pre-college and undergraduate geoscience education. Volunteers and students interview community members and acquire GPS coordinates of Native Ameri-can sites and map them as well as research their importance. This program helps reservation residents to enter and advance in geoscience careers. This project has the intellectual merit of increasing the understanding of geoscience by tribal residents. We are developing unique procedures for docu-menting Lakota cultural sites: a working database and an interactive online map of sites approved for publication. This map is a tool to acquaint tribal and non-tribal members with geospatial technology, Lakota values and culture. Author notes: Charles Comes-Killing, Oglala Sioux Tribal member, assis-tant/student at the GIS lab at Oglala Lakota College, Piya Wiconi, Pine Ridge Indian Reservation in Southwest South Dakota.


EAGLE NEST BUTTE

Elvin Returns and Sylvio MannelOglala Lakota College

Kyle, SD 57552

ABSTRACT

Presently there is a need for programs to identify, protect and preserve his-torical areas on the Pine Ridge Indian Reservation. New technology available, such as GIS - Geographical Information Systems (making and analyzing maps) and GPS - Global Positioning Systems (getting coordinates for locations) allows us to gather and document this information. The Pine Ridge Indian Reserva-tion is located near the southwestern corner of South Dakota. The reservation covers all of Shannon County, the southern part of Jackson County and parts of Bennett County. The Badlands area stretches across the northern border of the reservation and into the north section of Shannon County. This map displays the reservation and a few historical sites. Some of the sites are Eagle Nest Butte, Porcupine Butte, Prairie Chicken Butte, Doorway of The Wind, Redshirt Table, Old Woman Hill, Slim Buttes and Wounded Knee. Our objective was to produce educational information, locate sites and docu-ment all information for the residents of the reservation. We researched literature found in books, library articles, videos and historical documents. Furthermore, we interviewed residents of the reservation who had historical information. We first identified all sites and then drove and walked to each site to get the coordi-nates using the GPS unit. In the future we will use this information to create an understanding and awareness of our tribal heritage.


ADAPTATION OF LEAVES INCYPRIPEDIUM CANDIDUM

Katie Krahn and Carol WakeBiology/Microbiology


ABSTRACT

Cypripedium candidum is a terrestrial orchid that is found on the South Dakota and other Midwest prairies. Non-destructive leaf samples were collected during the summer of 2003, preserved in FAA and stored at 4°C until further studies were conducted. Leaf samples were infiltrated, embedded in paraffin, cut into 8 micrometer sections, placed on microscope slides and photographed. Staining procedures followed an optimized Safranin O / Fast Green protocol (Murphy, Wake 2005). The leaf structures and tissues demonstrated several traits adapted to various prairie microenvironments. Several main survival traits were studied: large sub-stomata crypts, extensive leaf trichomes (hairs) formation and C3 type leaf morphology / vascular bundles. The C3 photosynthesis type of leaf vascular anatomy enables C. candidum to grow very quickly in the cool late spring, early summer. Since its seed capsules do not mature until late September, extensive development of leaf trichomes and numerous sub-stomata crypts may play vital roles in allowing photosynthesis to continue while protecting the plant from desiccation during the warmer, more arid conditions of late summer, early fall. By incorporating both mesomorphic and xeromorphic traits, Cypripedium candidum are able to survive the various microenvironments / microclimates of eastern South Dakota open, prairie meadows.

Murphy, B., Wake, C. 2005. Histotechnique Staining Methods for Cypripedium candidum roots and rhizomes. SD Acad Sci. Abst. p. 351

Keywords

Cypripedium candidum, sub-stomata crypts, leaf anatomy, trichomes, C3 plant


OVARY ANATOMY OF CYPRIPEDIUM CANDIDUM

Joann McNally and Carol WakeBiology/Microbiology


ABSTRACT

Cypripedium candidum (white lady slipper orchid) is a native plant found in eastern South Dakota. Immature capsules were collected during the sum-mer of 2003, fixed and stored under refrigeration. To study the anatomy of the developing ovules (immature seeds), ovary (capsule) cross-section samples were infiltrated, paraffin-embedded, sectioned (8-10 microns), mounted, stained and photographed (Murphy, Wake 2005). The inferior ovary contained three locules with axile placentation with many developing ovules attached to the placentas of each carpel. The pericarp of this maturing capsule contained six longitudinal vascular bundles: three in pronounced external ribs between the fused carpels and three in median bundles running the length of the carpels. The vascular tissues of the protruding ribs were surrounded by many layers of highly lignified sclerenchymatous cells. The developing ovaries were very pubescent and the individual trichomes observed were multi-cellular and unbranched. Since C. candidum seed capsules require several months during the summer and fall to produce millions of dust-like seeds, the thick covering of trichomes may protect them from insect predation and / or desiccation. The external ribs may provide structural support as the capsules dehisce to distribute the wind-blown seeds of Cypripedium candidum to other South Dakota prairie meadows.

Murphy, B., Wake, C. 2005. Histotechnique Staining Methods for Cypripedium candidum roots and rhizomes. SD Acad Sci. Abst. p. 351

Keywords

Cypripedium candidum, capsule, ovary anatomy, ovule formation, tri-chomes

proceedings of the south dakota academy of science volume 85 … · 2017. 7. 12. · proceedings of...

Documents