Causes of Regional and Temporal Variation in Paleoindian Diet in Western North America

Home , Bonfire Shelter, Emeryville Shellmound

Item Type text; Electronic Dissertation

Authors Hill, Matthew E.

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 29/09/2021 03:03:13

Link to Item http://hdl.handle.net/10150/196065

CAUSES OF REGIONAL AND TEMPORAL VARIATION IN PALEOINDIAN DIET

IN WESTERN NORTH AMERICA

Matthew Ervin Hill Jr.

______

A Dissertation Submitted to the Faculty of the

DEPARTMENT OF ANTHROPOLOGY

In Partial Fulfillment of the Requirements For the Degree of

DOCTOR OF PHILOSOPHY

In the Graduate College

THE UNIVERSITY OF ARIZONA

2007 2

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Matthew Ervin Hill, Jr.

entitled Causes of Regional and Temporal Variation in Paleoindian Diet in Western North America and recommend that it be accepted as fulfilling the dissertation requirement for the

Degree of Doctor of Philosophy

______Date: October 16, 2007 Mary C. Stiner

______Date: October 16, 2007 Vance T. Holliday

______Date: October 16, 2007 Steven L. Kuhn

Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

______Date: October 16, 2007 Dissertation Directors: Mary C. Stiner and Vance T. Holliday 3

STATEMENT BY AUTHOR

This dissertation has been submitted in partial fulfillment of requirements for an advanced degree at The University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that accurate acknowledgment of source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

SIGNED: ____Matthew E. Hill, Jr.______

ACKNOWLEDGEMENTS

Many people have contributed to the successful completion this project. I am very grateful for the opportunities to work with Mary Stiner and Vance Holliday, who over the last four and half years have provided wonderful guidance and constant encouragement in all phases of my work. Thank you both very much. My other committee members, Steven Kuhn, Jay Quade, and Dave Dettman, have also provided constant interest and feedback in the various stages of this project. Four people are responsible (and blamed) for me being in school and thus are the reason this project was even possible: Drs. Suzanne Fish, Paul Fish, William Longacre and Margaret Beck. The advice and support shown by these wonderful and caring people made a cranky man realize what he really wanted to do. In addition, I am extremely thankful for the financial support provided by Paul, Suzy, and Margaret, which made going to school a reality. I also want to acknowledge the support and friendship shown to me over the years by a number of friends and colleagues. These fine people include Charles Adams, Jeff Altschul, Doug Avann, Doug Bamforth, Sharon Bauer, James Bayman, Joe Beaver, Lee Bement, Michael Bever, Scott Brosowske, Simon Bruder, Kent Buehler, Ryan Byerly, David Byers, Ken Cannon, Mike Cannon, Regina Chapin-Pyritz, Destiny Crider, Rebecca Dean, Ned Gaines, Dan Garcia, Janet Griffith, Cherie Freeman, Ted Hartwell, Virginia Hatfield, Marlin Hawley, India Hesse, Jerome Hesse, Jack Hofman, Steve Holen, Jeff Homburg, Bruce Huckell, Stance Hurst, Eileen Johnson, Edward Knell, Ruthann Knudson, Marcel Kornfeld, Jason LaBelle, Stacey Legyel, Luc Litwinionek, Brad Logan, Rolfe Mandel, Teresita Majewski, Larry Martin, Brian McKee, David Meltzer, Laura Niven, Bruce Phillips, Todd Pitezel, Adrianne Rankin, Bill Reitze, Lauren Ritterbush, David Rapson, Gene Rogge, Dean Sather, Frederic Sellet, Britt Starkovich, Todd Surovell, Lawrence Todd, Andrew Ugan, Danny Walker, Nicole Waguespack, Robert Weber, Tim Weston, and Don Wyckoff. When one studies bison, like I do, you end up spending a lot of time with some real weirdoes who spend way too much of their time thinking about stuff no person in their right mind should. My two favorite weirdoes are Matthew Hill (“Good Matt”) and Chris Widga. Not only am I pleased to call Matt and Chris friends, but I also thank them for years of encouragement, intellectual stimulation, good humor, and support. I only hope I can someday repay them for all the help they have given me. Furthermore, I want to acknowledge their contributions to this dissertation. Both were instrumental in the analysis and writing of Chapter 4. They also generously shared with me unpublished data for Agate Basin, Big Bone Lick, Big Goose Creek, Casper, Cordero Mine, Clary Ranch, Frasca, Hawken, Hudson-Meng, Itasca, Lindenmeier, Lipscomb, Lodgepole Creek, Milburn, Logan Creek, Mavrakis-Bentzen-Roberts, Mill Iron, Rustad, Scoggin, Simonsen, and Spring Creek. I also want to extend special recognition to Ed Knell. Over last couple years our lunch “meetings” have been a useful forum for discussing many of the issues I present in this dissertation. 5

Much of the data used in this study came from visits I made to museums across the country, and I acknowledge the efforts of a number of people who facilitated my work with these collections. Special thanks to John Alexander of the American Museum of Natural History (Folsom), Logan Ivy of Denver Museum of Nature and Science (Folsom), Fred Lange, Diana Leonard, and Deborah Confer of University of Colorado Museum (Jurgens and Olsen-Chubbuck), Larry Martin of University of Kansas Museum of Natural History (12 Mile Creek and Burntwood Creek), George Corner of Nebraska State Museum (Lipscomb, Scottsbluff, Red Smoke), Eileen Johnson of Texas Tech Museum (Lubbock Lake), Jeff Indeck and Rolla Shaller of Panhandle-Plains Historical Museum (Rex Rodgers), and Ernest Lundelius and Melissa Winans of Texas Archaeological Research Laboratory and University of Texas Vertebrate Paleontology Laboratory (Blackwater Draw, Bonfire Shelter, Lubbock Lake, Milnesand, and Plainview), Judy Brown, George Frison, Marcel Kornfeld, Martha Rodgers, and Danny Walker of University of Wyoming (Finley), and Christopher Shaw of Page Museum at the La Brea Tar Pits (La Brea Tar Pit). Financial support for various parts of this research were provided by Carroll D. Clark Fund of the Department of Anthropology at the University of Kansas, Haury Fund of the Department of Anthropology at the University of Arizona, and George C. Frison Institute. The NSF- Arizona AMS Laboratory, Greg Hodgin, and Sarah LaMotta supported the radiocarbon dating of the Scottsbluff, Rex Rodgers, and Finley assemblages. Special thanks also goes out to Mike Cannon, Tina Bothe, India and Jerome Hesse, Jack Hofman, Chad Jones, and Matt Hill for providing me a place to stay over the years while I worked on this project. Portions of this dissertation benefited from earlier comments by Margaret Beck, Donald Grayson, Matthew G. Hill, Oskar Burger, Sebastian Payne, Mary Stiner, Chris Widga, and Michael Wilson. Lisa Raffensperger is also thanked for her editorial assistance with the text. Finally, this research would not have been possible without Margaret Beck. I owed her everything. She provides an endless supply of love, support, encouragement, money (unfortunately not endless), as well as editorial and technical advice. I cannot express the gratitude I have that Margaret is part of my life

TABLE OF CONTENTS

LIST OF FIGURES ...... 10

LIST OF TABLES...... 15

ABSTRACT...... 17

CHAPTER 1: INTRODUCTION...... 18 Is a New Perspective Needed?...... 20 Project Goals and Approach ...... 21 Project Area ...... 24 Human Behavioral Ecology Theory ...... 25 Data Sources ...... 27 Significance and Implications of Dissertation ...... 31 Organization of Dissertation...... 33

CHAPTER 2: THEORETICAL FRAMEWORK...... 37 Diet Breadth and Prey Choice...... 39 Key Predictions of Prey-Choice Models...... 45 Recent Innovations to Diet Breadth Models...... 47 Sexual Division of Labor and Differential Provisioning Strategies ...... 47 Risk and Prey Choice...... 51 Central Place Foraging Patch-Choice Model...... 53 Choosing a Patch and which Resource to Exploit in that Patch ...... 57 Resource Depression and Foraging Efficiency...... 58 Processes that Depress Prey Populations...... 59 Hunting (Harvesting) Pressure...... 59 Environmental Conditions ...... 60 Effects of Resource Depression on Prey Populations...... 62 Prey Characteristics that Affect Resource Depression ...... 64 Implications of Resource Depression for Foragers...... 65 Archaeology and Foraging Theory ...... 67 Summary and Predictions ...... 70

CHAPTER 3: BISON PALEOBIOLOGY ...... 74 Approached to Studying Bison Ethology and Physiology...... 75 Social Organization...... 78 Reproduction...... 81 Rut ...... 81 Gestation ...... 83 Parturition ...... 84 Timing of Parturition and Birth Synchronization ...... 85

TABLE OF CONTENTS - Continued

Factors Influencing Birth Synchronization...... 88 Mobility...... 90 Ethnohistoric Accounts of Bison Mobility ...... 90 Ecological Studies of Bison Mobility...... 92 Bison Physiology ...... 96 Dietary and Seasonal Variation ...... 97 Variation by Sex ...... 101 Bulls ...... 101 Cows ...... 102 Ontogenic Variation...... 105 Modeling Bison Procurement and Processing ...... 106 Kill Composition and Size...... 106 Season of Mortality...... 107 Distribution of Prey...... 108

CHAPTER 4: BISON DIMINUTION: EVALUATING THE ROLE OF HUMAN HUNTING VERSUS CLIMATE-CHANGE FROM OSTEOMETRIC, AGE STRUCTURE, AND STABLE CARBON ISOTOPES ...... 109 Explaining Body Size Change ...... 112 Human Prey Selection as Catalyst for Body Size Diminution ...... 112 Scenario #1: Regularly Occurring, Catastrophic Kills ...... 114 Scenario #2: Sporadic Catastrophic Kills ...... 116 Scenario #3: Regular Selective Hunting...... 116 Scenario #4: Sporadic Selective Hunting ...... 117 Summary...... 117 Environmental Factors...... 119 Other Environmental Factors Affecting Body Size...... 122 Expected Effects of Environmental Changes ...... 125 Materials and Methods...... 125 Limb Elements Measurements as Proxy for Body Size...... 126 Mortality Patterns as Measure of Predation Pressure and Climatic Stress .... 134 Results ...... 135 Body Size: General Trends ...... 136 Body Size: Regional Trends ...... 141 Mortality Patterns...... 152 Discussion...... 157 Temperature Change...... 160 Predator and Competition Release...... 161 Stable Carbon Isotope and Grassland Ecosystems ...... 163 Conclusion ...... 167

TABLE OF CONTENTS - Continued

CHAPTER 5: REGIONAL DIFFERENCES IN PALEOINDIAN LITHIC ASSEMBLAGES, OCCUPATIONAL INTENSITY, AND DURATION ...... 171 Mobility and Resource Distribution...... 175 Prior Views of Paleoindian Life Ways ...... 179 Mobile, High-Tech Foragers...... 179 Low-Mobility, Low-Tech Foragers ...... 181 Archaeological Evidence of Mobility...... 182 Paleoindian Land Use Record...... 184 Site Function ...... 186 Comparison of Artifact Contents...... 188 Patterning in Distribution of Site Function...... 193 Occupational Intensity ...... 195 Artifact Density...... 196 Density Trends through Time ...... 200 Percent Local ...... 202 Discussion...... 204 Regional Trends...... 204 Temporal Trends...... 207 Conclusions...... 208

CHAPTER 6: VARIATION IN PALEOINDIAN FAUNAL USE...... 209 History of Research...... 213 Paleoenvironment, Prey Abundance, and Hunting Pressure ...... 216 Methods for Measuring Diet Breadth and Inferring Prey Rank ...... 219 The Study Sample...... 221 Trends in Paleoindian Faunal Assemblages ...... 227 Large Game Abundance and Bias from Kill Sites...... 227 Temporal Trend in Body Size of Prey Used...... 231 Use of Small- and Medium-Sized Game and Evidence for Resource Depression...... 236 Discussion...... 242 Conclusion ...... 247

CHAPTER 7: BISON KILLS: BONE BED COMPOSITION, SEASONAL VARIATION, AND LONG-TERM TRENDS...... 249 Previous Interpretation...... 251 The Record for Paleoindian Bison Bone Beds ...... 255 Results ...... 256 Seasonality of Kill...... 256 Bison Abundance: MNI Values...... 262 Herd Composition...... 270 Discussion...... 272 9

TABLE OF CONTENTS - Continued

Conclusions...... 276

CHAPTER 8: NEW PERSPECTIVE AND IMPLICATIONS ...... 277 Summary of Results...... 278 Land Use and Residential Mobility ...... 278 Diet Breadth and Diversity in Faunal Use ...... 280 Hunting Pressure and Human Impacts on Fauna...... 282 Bison Hunting ...... 284 Implications...... 285 Communal Kills and Aggregation Sites ...... 285 Organization of Male and Female Labor...... 288 After the Paleoindian Period...... 291 Big Game Hunting after Paleoindian Period ...... 292 Southern Plains ...... 296 Central Plains...... 297 Northern Plains and Rocky Mountains...... 297 Long-Term Trends in Resource Intensification and Diet Breadth...... 298 Use of Plants ...... 298 Grease Extraction...... 300 Some Closing Thoughts...... 301

APPENDICES ...... 303 Appendix A: Summary Information for Assemblages Used in this Study...... 303 Appendix B: Summary Information on Bison Remains Used in this Study...... 313

REFERENCES CITED...... 319

LIST OF FIGURES

Figure 1.1. Project area of this dissertation...... 28 Figure 2.1. Graphic depiction of the relationships between body size and (A) energy gain/individual, (B) processing costs, and (C) food rank (adapted from Bayham 1979: Figure 1)...... 43 Figure 2.2. An estimate of the optimal number of food items in the Ache diet that compares net energetic return of resources i (E i/h i) to (E/t), where E i = calories available in a unit of resource i, h i = handling time per unit of resource i, E = total calories acquired foraging, and t = search time (adapted from Hawkes, Hill and O'Connell 1982:Figure 1)...... 46 Figure 2.3. Patch choice model for central place forager (adapted from Kelly 1995: Figure 3–5). The gain function provided in both patches A and B is represented and maximized energetic return per total (roundtrip) time (T a and Tb) spent in the patch is given by the dashed line tangential to the gain function beginning at the origin of the graph. Richer patches have taller gain functions and the patch that provides the highest overall rate of energy is the one that produces the steepest slope of the line tangential to the gain function. . 55 Figure 2.4. Distribution of Deer Index ( ∑NISP Deer/ ∑NISP Mammals) values by stratum for Emeryville Shellmound (adapted from Broughton 2002:70)...... 66 Figure 3.1. Idealized annual bison herd size (data from Berger and Cunningham 1993)...... 79 Figure 3.2. Distribution of bison birth frequency for modern bison herds (Sources: Antelope Island: Wolf and Kimball 1989: Table 1; Badland National Park (1988–1987): Berger and Cunningham 1994: Table 6.1; National Bison Range: Rutberg 1984: Figure 1; Wind Cave National Park: Green and Rothstein 1993: Figure 1; Yellowstone National Park (1902–1925): McHugh 1958: 30; Yellowstone National Park (1902–1925): Meagher 1973: 75)...... 87 Figure 3.3. Idealized variation in bison physiology...... 101 Figure 4.1. Expected responses of Bison populations to different hunting scenarios...... 115 Figure 4.2. Map of the continental United States and southern Canada with archaeological sites and paleontological localities mentioned in the text and tables. Site numbers: 1 Heron-Eden, 2 Rustad, 3 Itasca, 4 Interstate Park, 5 Mill Iron, 6 Big Goose Creek, 7 Mavrakis-Bentzen-Roberts, 8 Rourke, 9 Carter-Kerr/McGee, 10 Hawken, 11 Vore, 12 Horner II, 13 Bugas-Holding, 14 River Bend, 15 Glenrock, 16 Casper, 17 Finely, 18 Wardell, 19 Scoggin, 20 Agate Basin, 21 Lodgepole Creek, 22 Hudson-Meng, 23 Scottsbluff, 24 Clary Ranch, 25 Milburn, 26 Spring Creek, 27 Red Smoke, 28 Logan Creek, 29 Simonsen, 30 Lindenmeier, 31 Jurgens, 32 Frazier, 33 Lamb Springs, 34 Frasca, 35 Olsen-Chubbuck, 36 Burntwood Creek, 37, 12 Mile Creek, 38 Scott County Pueblo, 39 Cooper, 40 Certain, 41 Lipscomb, 42 Rex Rodgers, 43 Twilla, 44 Plainview, 45 Lubbock Lake, 46 Bonfire Shelter, 47 Folsom, 48

LIST OF FIGURES - Continued

Blackwater Draw, 49 Milnesand, 50 Garnsey, 51 Rancho La Brea, 52 Big Bone Lick...... 127 Figure 4.3. Relationship between body mass and calcaneal tuber length for 57 ungulate species from across the globe. Data based on Christiansen (2002:Table 1)...... 132 Figure 4.4. . Ratio diagram of dimensions of bison calcanei plotted against the mean value for the modern standard (0.0 vertical line). Points with positive values are larger than the modern bison standard; points with negative values are smaller than the Late Holocene (Glenrock) bison standard. Gray bar represents range of values for Late Holocene comparative standard...... 137 Figure 4.5. Ratio diagram of dimensions of bison humeri plotted against the mean value for the modern standard (0.0 vertical line). Points with positive values are larger than the modern bison standard; points with negative values are smaller than the Late Holocene (Glenrock) bison standard. Gray bar represents range of values for Late Holocene comparative standard...... 138 Figure 4.6. Smoothed trend line of ratio diagram of dimensions of bison calcanei plotted against the mean value for the modern standard (0.0 vertical line), with relationships to major Late Quaternary climatic periods and colonization of North America by human hunters...... 140 Figure 4.7. Smoothed trend line of ratio diagram of dimensions of bison humeri plotted against the mean value for the modern standard (0.0 vertical line), with relationships to major Late Quaternary climatic periods and colonization of North America by human hunters...... 141 Figure 4.8. Cross plot of calcaneal maximum length with 95% confidence interval over time for cows from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds)...... 147 Figure 4.9. Cross plot of calcaneal maximum length with 95% confidence interval over time for bulls from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds)...... 148 Figure 4.10. Cross plot of humeral distal articular breadth with 95% confidence interval over time for cows from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds)...... 149 Figure 4.11. Cross plot of humeral distal articular breadth with 95% confidence interval over time for bulls from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds)...... 150 Figure 4.12. Group means and observed mortality patterns in bison assemblages from Early Paleoindian, Late Paleoindian, Archaic, and Late Prehistoric/ Protohistoric periods...... 154

LIST OF FIGURES - Continued

Figure 4.13. Box Plot comparison of the proportion of young animals represented in four Early Paleoindian, seven Late Paleoindian, four Archaic, and nine Late Prehistoric/Protohistoric bison assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H = 2.83; p = .42) cannot reject null hypothesis for equality of population median among groups...... 156 Figure 5.1. Map of the Great Plains and Rocky Mountains with Sites Mentioned in this chapter...... 185 Figure 5.2. Box plot comparison of the proportion of number of projectile point to number of lithic artifacts from 42 camp, 56 kill/processing, and 10 special function assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H = 45.9, df = 2, p <.001) rejects the null hypothesis for equality of population median among groups...... 190 Figure 5.3. Box plot comparison of the proportion of number of non-projectile point tools to number of projectile points from 36 camp, 49 kill/processing, and 4 special function assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H = 42, df = 2, p < .001) rejects the null hypothesis for equality of population median among groups...... 191 Figure 5.4. Box plot comparison of the proportion of number of pieces of debitage to number of tools from 37 camp, 39 kills/processing, and 4 special function assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H= 25.7, df = 2, p <.001) rejects the null hypothesis for equality of population median among groups...... 192 Figure 5.5. Box plot comparison of the artifact density (# of artifacts/m 2) of multifunction camps and kill/processing assemblages from sites <50 km and >50 km from lithic sources. Kruskal-Wallis one-way analysis of variance by ranks test (H = 42, df = 2, p < .001) rejects the null hypothesis for equality of population median among groups...... 198 Figure 5.6. Map of the Great Plains and Rocky Mountains with site artifact density distribution...... 199 Figure 5.7. Box plot comparison of the artifact density (# of artifacts/m 2) for 3 plains/rolling hills, 15 alluvial valleys, and 9 foothill/mountain multifunction campsites located <50 km lithic sources. Kruskal-Wallis one-way analysis of variance by ranks test (H = 2.67; df = 2; p = .26) does not rejects the null hypothesis for equality of population median among groups...... 201 Figure 5.8 Scatter plot of log (artifact density) versus site age (calBP). Best-fit regression line drawn though 59 camp/workshop, kill/processing, and special function sites (r 2 < .001; F = .004; df = 1, 57; p = .70)...... 202 Figure 6.1. Sites with Paleoindian faunal remains used in this study...... 210 Figure 6.2. Box plot comparison of large fauna (size class 4) abundance index (AI) between specialized kill/processing localities and multifunction short- term camps. Kolmogorov-Smirnov two-sample test (D=.731; χ2 = 35.9; p<.0001) rejects null hypothesis that two samples have the same distributions. 228 13

LIST OF FIGURES - Continued

Figure 6.3. Box plot comparison of large fauna (size class 4) abundance index (AI) among alluvial valley, foothill/mountains, and plains/rolling hill with (A) both kill/processing localities and multifunction short-term camps, and (B) multifunction short-term camps only. Kolmogorov-Smirnov two-sample tests do not reject the null hypothesis that the distributions of large fauna AI values for sites in alluvial valley (D = .188; χ2 = 1.009; p > .99), foothill/mountain (D = .183; χ2 = 1.74; p = .84), and plains/rolling hills (D = .120; χ2 = .195; p > .99) are the same between samples that included both kills and camps and samples that included only camps...... 230 Figure 6.4. Evenness in the representation of three prey categories (large-, medium-, and small-bodied fauna) using the Reciprocal of Simpson's Index (3 = most even) versus radiocarbon age. Reciprocal of Simpson's Index from alluvial camp assemblages are significantly higher than those from foothill/mountains (Kolmogorov-Smirnov two-sample D = .609; χ2 = 12.309; p = .004)...... 233 Figure 6.5. Evenness in the representation of three prey categories (large-, medium-, and small-bodied fauna) using the Reciprocal of Simpson's Index (3 = most even) versus radiocarbon age for the foothill/mountain multifunction short-term camps of Medicine Lodge Creek and Mummy Cave...... 234 Figure 6.6. Evenness in the representation of three prey categories based on body size (large-, medium-, and small-bodied fauna) using the Reciprocal of Simpson's Index (3 = most even) by radiocarbon age for the foothill/mountain multifunction short-term camps of Allen and Wilson Leonard...... 235 Figure 6.7. Triangular graph of proportional representation of three prey categories based on forager harvesting strategies and prey body size and predator avoidance strategies for medium ungulate, small slow, small- difficult-to-catch taxa for alluvial valley and foothill/mountain multifunction short-term camps. The presence of large fauna was excluded for consideration for this graph. (MLC=Medicine Lodge Creek)...... 237 Figure 6.8. Plot of (A) % slow small fauna and (B) % difficult small fauna versus radiocarbon age for multifunction short-term camp assemblages in alluvial valley habitats with best-fit regression lines...... 241 Figure 7.1 Map of the Great Plains and Rocky Mountains with sites mentioned in this chapter...... 250 Figure 7.2 Bar chart of relative frequency of 0.1-year dental age group intervals for 66 Paleoindian bison assemblages. Chi-square analysis ( χ2 = 25.9; df = 9; p = .002; Cramer’s V = .13) suggests significant difference between this distribution and equal representation of all dental age intervals...... 258

LIST OF FIGURES - Continued

Figure 7.3 Bar chart of relative frequency of 0.1-year dental age group intervals for 23 Early (>11,500 calBP) and Late (<11,500 calBP) Paleoindian bison assemblages. Chi-square analyses suggest no significant difference between distribution for Early Paleoindian assemblages and equal representation of all dental age intervals ( χ2 = 11.9; df = 9; p = .22), but significant difference for Late Paleoindian assemblages ( χ2 = 22.1; df = 9; p = .009; Cramer V’s = .15).. 260 Figure 7.4 Bar chart of relative frequency of 0.1-year dental age group intervals for 9 alluvial valley, 12 foothill/mountain, and 45 plains/rolling hill Paleoindian bison assemblages. Due to small sample for these groups, statistical evaluation of these patterns is only possible with plains/rolling hill sites. Chi-square analysis ( χ2 = 22.8; df = 9; p = .00; Cramer's V = .15) suggests significant difference between distribution for plains/rolling hill assemblages and equal representation of all dental age intervals...... 261 Figure 7.5 Distribution of minimum number of individual (MNI) estimates for bison from 84 Paleoindian assemblages...... 265 Figure 7.6 Plot of Log (MNI estimate) versus Log(excavation area [m 2]) for 84 Paleoindian bison assemblages with best-fit regression line. Best-fit regression line drawn to illustrate trend in data. Spearman’s rho shows a moderate, but significant, association between MNI and excavation size (r s=.42; p = .004)..... 265 Figure 7.7 Box plot comparison of minimum number of individuals (MNI) between 34 Early (>11,500 cal BP) and 49 Late (<11,500 cal BP) Paleoindian bison assemblages. Kolmogorov-Smirnov two-sample test (D=.197; χ2 = 3.1; p = .42) cannot reject null hypothesis that two samples have the same distributions. Largest outliers not shown here...... 267 Figure 7.8 Box plot comparison of minimum number of individuals (MNI) between 13 alluvial valley, 15 foothill/mountain, and 55 plains/rolling hill Paleoindian bison assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H = 4.45; p = .11) cannot reject null hypothesis for equality of population median among groups...... 268 Figure 7.9 Cross plot of mean minimum number of individuals (MNI) with 95% confidence interval over eight seasonal intervals for Paleoindian bison assemblages. Only sites with seasonal estimates confined to a single season were used in this study...... 269 Figure 7.10 Plot of mean %cow/calf (closed circle) and mean %bull (open squares) by eight seasonal intervals...... 271 Figure 7.11 Plot of %males versus minimum number of individuals (MNI) for 34 Paleoindian bison assemblages. Spearman’s rho (r s= -.24; p = .22) shows weak negative relationship between %male and MNI for 28 Paleoindian bison assemblages with evidence for male animals present...... 272

LIST OF TABLES

Table 1.1. Important Hypotheses Concerning Paleoindian Subsistence-Settlement Strategies...... 23 Table 1.2. The Distribution of Cultural Complexes by Physiographic Section/Province for Assemblages used in this Dissertation...... 29 Table 1.3. Summary of Site Function by Physiographic Section/Province for Assemblages used in this Dissertation...... 31 Table 2.1. Strategic Goals, Key Assumptions, Cost-Benefit Criteria, and Major Constraints of Diet Breadth, Patch Choice, and Resource Depression Models (Adapted from Beaver 2007; Charnov 1976; Orians and Pearson 1979; Smith 1982; Winterhalder 2001)...... 38 Table 2.2. Net and Gross Energy Yields of Food Classes Consumed by Modern Foragers (Data from Kuhn and Stiner 2006:Table 1)...... 44 Table 2.3. Key Predictions Generated by Diet Breadth models (adapted from Kelly 1995; Smith 1982; Winterhalder 2001)...... 46 Table 2.4. Predictions Generated by Central Place Foraging Patch Choice model (adapted from Orians and Pearson 1979; Smith 1982; Winterhalder 2001)...... 56 Table 3.1. Modern bison research topics and their implications for archaeological interpretations...... 76 Table 3.2. Distributions for the Timing of Annual Birth for Modern Bison Herds...... 86 Table 3.3. Annual Home Range and estimated Foraging Radius for Modern Bison Populations...... 95 Table 4.1. Summary Information for Sites Used in this Analysis ...... 128 Table 4.2. Summary Statistics for Calcanei Greatest Length for Sites Used in this Study...... 143 Table 4.3. Summary Statistics for Humeral Distal Articular Breadth for Sites Used in this Study...... 145 Table 4.4. Three-Cohort System Results for Bison Kill and Processing Localities, based on Mandibular Tooth Eruption and Wear Sequence...... 153 Table 5.1. Sites used in this chapter...... 173 Table 5.2. Comparison of frequency of the major tool categories for Kill/Processing and Multi-function Camp localities...... 194 Table 5.3. Frequency of Archaeological Components by Site Function and Physiographic Region ...... 195 Table 5.4. Frequency of Archaeological Components by Site Function and Time Period...... 195 Table 5.5. Summary Statistics for Site Artifact Density (# Artifact/ m 2) by Site Type ...... 196 Table 5.6. Summary Statistics for Raw Material (%Local) by Physiographic Region...... 203 Table 5.7. Summary Statistics for Raw Material (%Local) by Site Type ...... 203 Table 6.1. Sites with Paleoindian Faunal Remains...... 211 16

LIST OF TABLES- Continued

Table 6.2. Summary of Life History, Predator Defense Strategy, and Capture Requirements for Common Medium and Small Prey in Paleoindian Assemblages...... 221 Table 6.3. Summary information for 69 Paleoindian zooarchaeological assemblages used in this chapter...... 222 Table 6.4. Assemblage Ubiquity, Total NISP, and NISP/Site for Identifiable Fauna Reported from the Sites Listed in Table 6.1...... 226 Table 6.5. Comparison of relative taxonomic abundance estimates based on estimated mass of meat as calculated from site minimum number of individual (MNI) estimates versus number of individual specimens (NISP) for three Paleoindian sites...... 246 Table 7.1. Summary of key predictions related to the Paleoindian bison kills composition and seasonality, as developed in chapter 3...... 249 Table 7.2. Julian and calendar day intervals associated with bison dental age groups...... 257 Table 7.3. Summary statistics for bison minimum number of individual (MNI) values for the eight seasonal estimates used in this study...... 269

ABSTRACT

This dissertation explores geographic and diachronic variation in Great Plains and

Rocky Mountain Paleoindian (12,500–7000 14 C years before present) forager exploitation of animal resources in order to explore how use of different habitats influenced land-use and subsistence strategies. To accomplish this goal, this study documented the full range of variability in the Paleoindian record using a combination of published data and new data. These patterns were then compared to explicit predictions derived from behavioral ecology and animal ethology and biology studies. The results, presented in this dissertation, allow the testing of several, often contradictory, important subsistence- settlement hypotheses in current Paleoindian research, specifically the ongoing debate about Paleoindian diet breadth and human causes of megafaunal extinction. Overall, there appears to be a covariance between environmental zone and forager land use. Paleoindian foragers structured their land use according to the presence and nature of a number of important resources within major environmental zones. Specifically, this study finds sites in grassland settings with low diversity of resources have lower artifact densities and are often dominated by exotic lithic raw materials. In these same areas prehistoric groups made almost exclusive use of large fauna. Sites in foothill/mountain or alluvial valley settings with ecologically high density and high diversity have higher proportions of short-term camps than do other areas and those camps have higher artifact density than do other types of sites. These sites exhibit a mixed use of small- and medium-sized game.

Overall this study shows Paleoindian hunters had only modest impact on prey species. 18

CHAPTER 1: INTRODUCTION

Since 1895, archaeologists and paleontologists have excavated hundreds of sites across the Great Plains and Rocky Mountains in an attempt to understand the economy, land use, and mobility of Paleoindian hunter-gatherers (e.g., Frison 1991; Hofman and

Graham 1998; Sellards 1952; Wormington 1957). Perhaps because of the varied nature of the Paleoindian archaeological record, views concerning Paleoindian subsistence and land use have differed over time. For much of the 20th century, archaeologists simply saw Paleoindians as little more than mobile big game hunters (e.g., Barbour and Schultz

1932a, 1932b, 1936; Schultz 1932, 1943; Sellards 1952; Wedel 1961; Willey 1966;

Wormington 1957). In the last couple of decades, a vigorous debate over two distinct models for Paleoindian subsistence strategies has appeared, each with different perspectives of the relationship between human demography, climatic change, and prey ecology.

One group of researchers views Paleoindians as independent bands of foragers who were heavily technology-dependent, highly mobile specialized hunters (Bement 1999;

Frison and Todd 1986, 1987; Hofman and Todd 2001; Waguespack and Surovell 2003;

Waguespack 2007; Wheat 1972). This group believes these people relied almost exclusively on mammoth and bison for their subsistence, and had home ranges that extended over vast areas. Such a high degree of residential mobility and little potential to meet up with other groups during their travels meant that Paleoindians had to carry a 19

flexible and highly reliable biface-based tool kit that would be useful for multiple

activities (Hofman 1991, 1992; Kelly and Todd 1988; MacDonald 1998).

According to these researchers Paleoindians were quite unique among the hunter- gatherers that have lived throughout North America. It was only approximately 8000 years ago when the landscape filled with people and large game became scarce that foragers living in the Great Plains and surrounding areas became more sedentary and more dependent on a broad spectrum of plants and small game, similar to hunter- gatherers living later in time.

The alternate perspective sees Paleoindians as less mobile and relying mostly on small game and plant resources for much of their subsistence (e.g., Bamforth 1988,

2002a, 2002b; Bonnichsen et al. 1987; Bousman et al. 2004; Cannon and Meltzer 2004;

Davis 1962; Greiser 1985; Johnson 1977, 1987). Although this group of researchers also considers Paleoindian population density to have been very low, they believe foragers occupied regions of high ecological diversity more intensively and repeatedly, often constructing thermal features and habitation structures as well (e.g., Irwin-Williams et al.

1973; LaBelle 2005). Because in this view foragers ranged over relatively small areas,

Paleoindian technology was not based on high-quality bifacial tools, but instead focused on informal reduction of locally available raw materials. Some even argue that by 10,000 years before present, population density and regionalism was so pronounced that distinct ethnic groups began to appear within the Paleoindian period (Frison 1992; Frison and

Grey 1980; Johnson 1989; Pitblado 1999). 20

This second group of researchers does not portray Paleoindian foragers as very different from later groups living in North America. Any differences in mobility, reliance on large game hunting, or use of bifacial lithic technology between Paleoindians and later populations are a matter of degree, not of kind.

IS A NEW PERSPECTIVE NEEDED?

As this brief review suggests, Paleoindian archaeologists have segregated themselves into two camps based on diametrically opposed theories about Paleoindian life ways. I argue in this dissertation that if archaeologists accept only one of two opposing models for understanding Paleoindian subsistence and land-use strategies, there is little possibility for increasing our understanding of much of the variation inherent in the

Paleoindian record.

I argue here that a more productive approach for understanding Paleoindian life ways is to (1) document the full extent of the variation in archaeological expression of the behavioral repertoire of Paleoindian foragers, (2) determine whether there is geographic and/or temporal patterning in the archaeological record, and (3) attempt to identify some of the underlying factors (e.g., environmental, demographic, social) responsible for these behavioral differences (Binford 1980, 1987, 2001). As LaBelle (2005:2) recently noted, a

“productive research strategy examines where this variability occurs, in what form it takes, and at what spatial scale it is visible.” 21

Modern hunter-gatherers have been repeatedly shown to utilize different strategies in response to the different structural properties of their environment (Binford 1980, 1982,

1983; Kelly 1995). Therefore, I see no reason that the seemingly contradictory behavioral repertoires outlined above could not represent end-member expressions of one or more adaptive continua. This idea follows Binford’s notion of the forager-collector continuum, whereby the specific subsistence-settlement strategy utilized by a hunting and gathering group is flexible and can change in response to the environmental conditions encountered by that group, despite the existence of some overall characteristics that limit and define the cultural adaptation. The approach used in this dissertation is not to develop a single model that explains the entirety of the Paleoindian record, but rather to try to test a series of hypotheses about the structural organization of Paleoindian foraging behavior.

PROJECT GOALS AND APPROACH

In this dissertation I explore geographic and diachronic variation in Great Plains and

Rocky Mountain Paleoindian (12,500–7000 14 C years before present) exploitation of animal resources. The main goal is to understand how the use of different habitats influenced land use and subsistence strategies and ultimately the character of the

Paleoindian archaeological record. Exploring these patterns is important for understanding how these hunter-gatherers reacted to the uneven distribution of resources across landscapes and to long-term environmental changes during the Late Pleistocene and Early Holocene. 22

While documenting the full range of variability in the Paleoindian record is an important first step, pattern recognition alone is not sufficient. No matter how good our pattern recognition skills may be, they alone cannot lead to an understanding of the organizational properties of the human behavior and the natural formation processes that created the patterns (Binford 1983, 1987). I move beyond pattern recognition here by comparing the empirical patterns in the archaeological record to predictions derived from behavioral ecology and animal ethology and biology studies (e.g., Smith 1982; Speth

1983; Winterhalder and Smith 1981).

I follow an approach outlined most recently by Broughton (2002:62), who recognized that while the archaeological record does not directly test predictions of the foraging models, archaeological data can be used to “leverage predictions” about how foraging efficiency and various archaeological indices should have varied in the past (see also

Binford 1968, 1987). Foraging theory and animal behavioral studies are used to create explicit predictions that will function more or less as null hypotheses to which trends in the quantitative analyses of archaeological data can be compared.

The results from these comparisons can be used to test several, often contradictory subsistence-settlement hypotheses in current Paleoindian research (Table 1.1). Ultimately

I attempt to draw correspondences between habitat settings (grasslands, woodlands, alluvial valley) occupied by Paleoindian foragers and their hunting strategies.

Specifically, I test models for geographic and seasonal correspondences between site occupation intensity, environment(s) utilized, relative use of small versus large faunal resources, and variation in bison hunting. 23

Table 1.1. Important Hypotheses Concerning Paleoindian Subsistence-Settlement Strategies Hypothesis Evidence For Evidence Against Chapter(s) • Biased selection of site • Dominance of large game types used in analyses in faunal assemblages • Overreliance on use of Paleoindians • Chipped stone tools made projectile points have high from distant raw • 5 residential Biface technology not materials mobility typical • Lithic technology based • Low quality, locally on well-made bifaces available lithic raw material common • Small site size and • Recovery and excavation at Paleoindian chipped stone many sites incomplete sites were assemblages • Dense accumulation of occupied for • Lack of storage features 5 artifacts at many sites brief periods and habitation structures suggest extended of time • Little differentiation in occupations site tool assemblages • Dominance of bison • Biased selection of site Paleoindian remains in many faunal types used in analyses subsistence assemblages • Evidence for use of small came • Evidence for year-round game 6 & 7 primarily hunting of bison • Some regions have no from hunting • Low richness and skewed evidence for use of large large game evenness in faunal game assemblage There is • Measures for diet breadth limited • Use of bison does not increase through time evidence for 4, 6, & 7 change through time • Bison undergo size resource reduction through time depression Bison use does not • Bison hunting occurs • Differential representation change year-round of males and females at kill according to • Pleistocene/Early sites 4 & 7 changes in • bison Holocene climates were Size of kills vary according physiology or less seasonal than today to season of mortality behavior Human • Megafauna extinctions • No evidence for human hunting correlate with appearance exploitation of most taxa caused of human hunters that went extinct 4 extinction of • Climate models do not • Bison survive the end of megafauna explain extinction events the Pleistocene 24

The data used in this study are drawn from published sources and recent work by the author from a large number of sites and over a broad area encompassing the northern and

Southern Plains and adjacent Rocky Mountains of western North America. Simple exploratory analyses are used to explore the dataset to identify factors that influence

Paleoindian dietary diversity and understand the scale in time and space in which these factors operated.

PROJECT AREA

The Great Plains and Rocky Mountains were chosen as the project area of this study because together they provide a diverse environmental setting that is very sensitive to even small-scale perturbations. The major habitats in this region provided hunter- gatherers with a diversity and abundance of key biotic and abiotic resources.

Knell (2007) recently showed that there were differences in the abundance and diversity of edible floral and faunal resources among major modern environmental zones

(e.g., alluvial valley, foothill/mountain, and plains/rolling hills). He shows that alluvial valley and foothill/mountain environments exhibit greater diversities of edible biotic resources than do plains/rolling hill environments, irrespective of season (Knell

2007:Table 3.2). In addition, because foothill/mountain and alluvial valley environments are distributed over significantly fewer hectares than are grassland environments, the resources in these areas are also more densely spaced. 25

The differences among environmental areas are not just limited to number of useable resources; there are also seasonal differences in their relative abundance, measured as total kilocalorie per hectare (kcal/ha), that indicate that foothill/mountain settings have the highest abundance (measured as kcal/ha) for edible biotic resources of any of the three environmental zones (Knell 2007:Table 3.6). This difference is most apparent during the fall and winter. Alluvial valleys also exhibit higher kcal/ha values than the plains/rolling hills during all seasons, although the difference is less pronounced during the fall.

The Great Plains and Rocky Mountains are also an appropriate study area because they have a long history of archaeological research that has produced much high-quality information about site age, chipped stone assemblage, bison utilization, and use of small- bodied animals for a large number of sites (Frison 1991; Hofman and Graham 1998).

HUMAN BEHAVIORAL ECOLOGY THEORY

As more fully discussed in the next chapter, this study uses human behavioral ecology models to create a testable analytical framework for studying Great Plains and Rocky

Mountain Paleoindians’ subsistence and land-use strategies. Human behavioral ecology represents a family of theories derived from microeconomics and ecology that attempts to understand the decision-making process of a rational individual to fulfill certain needs under conditions of limited resources and means (Charnov 1976; Bettinger 1987; Bird and O’Connell 2006; Emlen 1966; MacArthur and Pianka 1966; Orians and Pearson 26

1979; Smith 1982; Winterhalder and Smith 2000). The goal of this body of theory is to

create models that are general enough to apply to as many different species as possible

and yet specific enough to make detailed explanations about the behavior of an individual

forager (Smith 1982:626). In general, behavioral ecology models describe the actions of a hypothetical actor who must overcome certain environmental constraints by choosing, from a series of options, the strategy that provides the optimal returns on some relevant currency.

These models have been applied to a wide range of problems in anthropology (see reviews in Bettinger 1987; Cronk 1991; Winterhalder and Smith 2000). Anthropologists, including many archaeologists working in the American West, have tended to used behavioral ecology theories to explore issues such as (1) diet breadth or prey/food choice

(Butler and Campbell 2004; Broughton and Grayson 1993; Jochim 1988; Simms 1987),

(2) patch choice and settlement location (Bettinger 1991; Cannon 2000), (3) social relationships or group dynamics (Belovsky 1988; Bettinger and Baumhoff 1982), (4) processing behavior and resource transport (Barlow and Metcalfe 1996; Bettinger et al.

1997; Burger et al. 2005), (5) technological or behavioral innovation (Bright et al. 2002;

Ugan et al. 2003) and (6) resource sharing and exchange (Blurton Jones 1987;

Winterhalder and Smith 2000).

While it is likely that Paleoindian foragers were rational decision makers, they probably did not always make decisions that optimized their caloric returns. As a result, optimization models are most valuable because they are able to easily demonstrate when foraging groups are acting in non-economically-efficient ways relative to the modeled 27 assumptions (Cronk 1991). And it is these instances, when foragers respond in ways other than that which we would predict, that are most important to explain and potentially that provide the greatest insight into past behaviors.

DATA SOURCES

Data collected for the study were drawn from a total of 269 Paleoindian assemblages

(Appendix A; Figure 1.1). Most of the information used here comes from published information on fauna remains, chipped stone tool and debitage, and a series of site-level characteristics (e.g., geomorphic context, area excavated). These data were supplemented with data independently collected from the 17 zooarchaeological assemblages (e.g.,

Byerly et al. 2005, 2007; Hill 1996, 2002a, 2002b, 2006, 2007; Hill et al. 1992, 1993;

Hill and Hill 2002; Hill and Hofman 1997; Hofman et al. 1995) along with unpublished data provided to the author by Matthew G. Hill and Chris Widga.

While Clovis, Folsom, and Cody assemblages are well represented in this study

(Table 1.2), most other cultural complexes are not. Moreover, nearly 32% (n=85) of these sites could not be assigned to a specific cultural complex, but instead were classified as dating to either the Early (>10,000 14 C years BP) or Late (<10,000 14 C years) Paleoindian period. Assemblages were grouped by physiographic sections and provinces as defined by Fenneman (1928, 1931). Approximately 80% of the assemblages used in this study

Figure 1.1. Project area of this dissertation

Table 1.2. The Distribution of Cultural Complexes by Physiographic Section/Province for Assemblages used in this Dissertation. Dissected Middle Southern Alberta Black Colorado Edward High Missouri Osage Plains Raton Wyoming Cultural Complex Till Rocky Rocky Total Plains Hills Piedmont Plateau Plains Plateau Plains Border Volcanic Basin Plains Mountains Mountains Clovis 2 2 8 2 1 6 2 23 Goshen 1 1 4 1 7 Folsom 3 10 2 6 2 5 1 4 2 35 Indeterminate Early Paleo. 1 5 11 1 3 21 Plainview 12 3 15 Agate Basin 1 1 4 6 Hell Gap 1 1 1 5 2 10 Cody 3 1 7 1 1 9 6 6 1 1 1 3 40 Dalton 1 1 2 Milnesand 1 1 Merserve 2 2 Allen/Frederick 1 9 2 1 1 14 Angostura 2 1 5 3 1 12 Golondria 3 3 Indeterminate Late Paleo. 1 8 18 26 6 5 64 Lovell Constricted 4 4 Lusk 1 1 Pryor Stemmed 7 7 Wilson 2 2 Total 4 3 14 3 23 83 55 35 19 9 2 8 11 269

come from five major areas: the High Plains (n=83), Middle Rocky Mountains (n=55),

Missouri Plateau (n=35), Edward Plateau (n=23), and Osage Plains (n=35).

David Meltzer and others (Meltzer 1993, 2004; Cannon and Meltzer 2004; Hill 2007)

have suggested the composition and diversity of zooarchaeological assemblages are

expected to vary among sites that serve different functional roles in settlement and

subsistence systems (e.g., kills, residential camps, processing localities). As Cannon and

Meltzer (2004:1980) argue, zooarchaeological assemblages from kill and butchery sites are expected to consist of a small number of taxa killed at or near those sites; faunal assemblages from residential camps, on the other hand, may consist of a more diverse suite of taxa brought to the site from many different locations.

Given the possible influence site function has on the zooarchaeological assemblage composition, this dissertation examines general functional characteristics through the characteristics of the chipped stone assemblage at each locality and secondarily on non- faunal characteristics concerning site organization and structure (Table 1.3). The site function definitions used here are discussed in Chapter 5, and generally agree with prior definitions developed by Sellet (1999), Wheat (1978, 1979), Judge (1973), and LaBelle

(2005). These site definitions are based primarily on the characteristics of the chipped stone assemblages at each locality and secondarily on non-faunal characteristics that provide information on site organization and structure. While all site function designations are somewhat problematic and influenced by assumptions about foraging

organization and mobility, the designations used in this study are generally accepted by

most researchers in the field, and used here to group different assemblages together based 31

on particular site variables in order to facilitate comparisons among the composition of

zooarchaeological assemblages.

Table 1.3. Summary of Site Function by Physiographic Section/Province for Assemblages used in this Dissertation Site Function Physiographic Burial/ Kill/ Lithic Section/ Province Camp Quarry Unknown Cache Process Workshop Alberta Plains 3 1 Black Hills 3 Colorado Piedmont 1 1 5 5 2 Dissected Till 1 2 Plains Edward Plateau 13 1 1 8 High Plains 1 3 19 47 13 Middle Rocky 1 45 3 5 1 Mountains Missouri Plateau 25 5 3 2 Osage Plains 1 4 7 3 4 Plains Border 1 1 5 1 1 Raton Volcanic 1 1 Southern Rocky 6 2 Mountains Wyoming Basin 6 3 1 1 Grand Total 3 6 129 84 14 33

SIGNIFICANCE AND IMPLICATIONS OF DISSERTATION

The underlying question in this study is, how did human exploitation of animal resources change in response to rapid climatic and human demographic changes? While

the primary goal of this work is to increase our understanding of regional and long-term

variation in Paleoindian subsistence and land use, it is also hoped that the results will 32 contribute to larger anthropological questions about the nature of human-environment interactions.

In particular, my findings are relevant to the Paleoindian/Archaic transition.

Traditionally, archaeologists have assumed that the subsistence and land-use patterns associated with the Archaic period derived from the earlier Paleoindian system (e.g.,

Caldwell 1958; Jennings 1957; Willey 1966). But it seems that the transition from

Paleoindian to Archaic society represented a significant evolutionary shift from a highly mobile large-game hunting system to a more sedentary system heavily reliant on plants and small game (Bayham 1979).

While there is little reason to doubt that Archaic populations descended from

Paleoindian groups, I believe the traditional model presents an oversimplified dialectic between Archaic and Paleoindian adaptations. I explore here the underlying issues associated with understanding how much variation is to be expected in a single forager adaptive system. I specifically try to understand (1) the conditions under which small animals become a viable economic alternative to large game, (2) factors that trigger hunter-gatherers' decisions to increase exploitation intensity of a single, large-bodied prey species and/or broaden diet breadth through inclusion of new prey species, (3) effects hunting has on prey populations, and (4) the key factors underlying decision-making associated with resource scheduling and seasonal exploitation.

These issues are relevant to studies beyond North America; for example, they are important for the global study of the organization of the penecontemporaneous "broad spectrum revolution" that was occurring throughout the Americas and the Old World at 33

the beginning of the Holocene (Flannery 1969; Stiner 2001). Some of the socioeconomic

changes that we see in North American prehistory are mirrored by independent processes

elsewhere in the world and in other periods. Explanations citing cultural diffusion or

direct influences are out of the question. Understanding the details of one story therefore ultimately furthers understanding of similar changes occurring elsewhere in the world.

ORGANIZATION OF DISSERTATION

Because of the large-scale goals of this dissertation and the desire to use data from as

many sites as possible, the chapters of this dissertation do not follow a single research

thread. The reality is that not all sites contributed equally to each research question posed

by this study. In addition, there is value in matching the scale and scope of the analysis to

the available dataset and specific questions being posed. The reader should be aware that

the approach taken changes with each chapter, as does the temporal and spatial scope.

The analyses of chipped stone artifacts and the various types of faunal remains differ in

terms of the assemblages used and how the data is organized. As a result, each data

analysis chapter in this dissertation should be treated as a stand-alone study, which tries

to answer a unique set of questions about Paleoindian subsistence and land use.

Chapter 2 establishes the theoretical framework for this dissertation. This chapter

describes three closely related analytical categories used in behavioral ecology theory:

diet breadth, patch choice, and resource depression. Specifically, discussions for each

model will focus on (1) strategic goals of the theory, (2) key assumptions, (3), the cost- 34 benefit criteria used, and (4) the variables that constrain the actions of the forager. In addition, this chapter identifies key predictions generated from these models. The chapter concludes with a consideration of the inherent limitations in applying these models to archaeological problems, and presents a testable analytical framework for studying Great

Plains and Rocky Mountain Paleoindians’ subsistence and land-use strategies.

Chapter 3 uses modern studies of bison ethology and biology to develop an ecological foundation upon which to model seasonal variation in prehistoric utilization of bison. The first half of this chapter describes modern bison ethology, dealing with issues such as social organization, reproduction, and mobility. The rest of the chapter outlines the annual variation in bison physiology. The purpose is not to draw simple analogies between modern and prehistoric bison, but to instead examine the underlying conditions that affect bison behavior and physiology, in an effort to develop explanatory models that are as applicable to modern bison as they are to Late Pleistocene/Early Holocene herds.

Chapter 4 draws on data concerning body size and mortality age profiles from North

American Bison of various periods and published results from soil stable carbon isotope studies. The isotope data are used as proxy measures for understanding the relative effects of climate-driven environmental change and human hunting pressure on large terrestrial mammals of North America. However, body size variation and mortality age profiles are the primary foci of this analysis because they relate closely to many aspects of animal ecology and behavior, including responses to hunting pressure and environmental conditions. 35

Chapter 5 uses chipped stone assemblages to characterize Great Plains Paleoindian land-use strategies. This chapter begins with a brief review of how prior studies linked forager mobility and land use with environmental constraints. I then describe the expected material correlates to hunter-gatherer mobility and review the extant

Paleoindian archaeological record for the Great Plains and Rocky Mountain areas. An explicitly exploratory data analysis approach is used here, in which simple quantifiable measures of site function, chipped stone artifact density, and percent non-local lithic raw material are used to identify potential correspondences between occupational intensity, group mobility, and environmental settings. In addition, I examine how the patterns in the data vary with time.

Chapter 6 tests the following propositions: (1) including data from kill assemblages in faunal comparisons overestimates the contribution of big game hunting to the Paleoindian diet; (2) Paleoindians altered their diet based on differences in the environmental setting occupied at any given time and/or over time due to environmental change or human- induced resource depression; and (3) data on Paleoindian use of small game can be used as indicators of resource depression.

Chapter 7 examines how seasonality affects Paleoindian procurement strategies.

Three questions are asked about this relationship: (1) During which season(s) did most bison hunting occur? (2) Does seasonality influence the size of a kill? (3) Does the representation of male and female bison in the assemblage change according to the season of the kill? 36

Finally, Chapter 8 summarizes and integrates the many results of this dissertation and evaluates the validity of existing models for Paleoindian land use and subsistence strategies. This chapter concludes with a brief discussion of the implications of the results for our understanding of the Paleoindian-Archaic transition and use of faunal resources by later prehistoric populations in the project area. 37

CHAPTER 2: THEORETICAL FRAMEWORK

The following chapters use human behavioral ecology models to generate a series of testable hypotheses for comparison with Paleoindian patterns. In order to produce useful hypotheses, this chapter presents a review of the behavioral ecology literature in order to identify (1) the strategic goals of each of the foraging models used in this study,

(2) the underlying assumptions of the models, (3) the cost-benefit criteria presented in each model, and (4) the major variables assumed to constrain the actions of the forager

(Table 2.1). Throughout this dissertation, a forager is defined as any actor who acquires foodstuffs and other important resources through tactics other than agriculture and husbandry (Winterhalder 1981:16). Only when discussing Binford's (1980) forager- collector continuum is the term defined differently. In addition, brief discussions will outline how the use of these models has changed over the years.

Discussions here will focus on three interrelated themes: diet breadth, patch choice, and resource depression. Patch is defined as any discrete, localized concentration of resources (Winterhalder 2001:19). Although these are but three of a much broader suite of concepts included in foraging theory, these are the most commonly used models and the most relevant to the subject of this dissertation. 38

Table 2.1. Strategic Goals, Key Assumptions, Cost-Benefit Criteria, and Major Constraints of Diet Breadth, Patch Choice, and Resource Depression Models (Adapted from Beaver 2007; Charnov 1976; Orians and Pearson 1979; Smith 1982; Winterhalder 2001). Resource Diet Breadth Patch Choice Depression • Identify optimal patch use set • Identify effect • Identify the optimal Strategic • Identify optimal of predation set of resources to Goals duration of patch use pressure on be exploited • Identify optimal set of prey species resources within patch • Resources randomly • Resources randomly and sequentially and sequentially encountered encountered • Multiple patches • Foraging time utilized includes search time • Resource distribution and handling time "patchy" • Foragers encounter • Resources transported resource in Same as diet Key to residential base proportion to their choice and patch Assumptions • density (fine-grained Foraging causes choice models encounter) depleted resources along curve of • Simultaneous search diminishing returns for all prey types • Travel time is • Prey rank is unproductive measured along a • Patches not reoccupied single dimension of profitability until resources within it are rejuvenated • Marginal return rate • Return rates per unit Cost-Benefit within patch types and of search and Criteria average net return for handling time all patches • Efficiency ranking of • Search and pursuit patch types time • Prey Major • Habitat richness • Encounter rates with reproductive Constraints • Travel time between high-ranked rates patches resources • Resource depletion rate

DIET BREADTH AND PREY CHOICE

The most common problem of foraging theory concerns how foragers choose which foodstuff to exploit. This question is usually approached through diet breadth models

(also referred to as prey models or prey choice models). The seminal diet breadth models

were developed by MacArthur and Pianka (1966) and Emlen (1966), with later revisions

made by Eastabrook and Durham (1976), Heller (1980), Rapport (1971), and Schoener

(1971). While each of these models is unique, they all share a few key attributes. The

scope of this chapter does not allow full review of the mathematical aspects of any

particular model; I will discuss the key components, major assumptions and important

predictions inherent to most of the diet breadth models (see Beaver 2007 for discussion of

mathematic constructions of some of the models).

All diet breadth models have four components: (1) goal, (2) currency, (3) constraints,

and (4) decision or alternative set (Hames and Vickers 1982; Schoener 1971;

Winterhalder 1986). Diet breadth models assume that the goal of a forager is to optimize

some variable based on specific environmental (including physical and social

environments) conditions.

The optimization assumption is perhaps the most controversial aspect of diet breadth

models, because many critics believe it presupposes that foragers make explicit choices

to optimize or conduct complex calculations to determine which strategy is most

beneficial. In fact, numerous studies have shown that many different organisms, from

invertebrates to humans, optimize to varying degrees and there is no evidence that they 40 make a conscious decision to do so (see Emlen 1966; Krebs and Davies 1987; McNamara

1982; Oaten 1977). Instead, it is believed that natural selection will favor strategies with the capacity to efficiently solve fitness-related problems (Schoener 1971; Smith 1982).

Bird and O’Connell (2006) group optimization analyses under the term “phenotypic gambit.” At optimization’s most basic level, foragers simply rely on a specific strategy, informed by prior experience, which solves a particular problem in as near optimal manner as possible. Behavioral ecology theory makes no assumptions about the exact mechanism (e.g., instinct, learned behavior, conscience thought) underlying the decision.

If the goal is to optimize, then what is the unit or currency most important to the forager? Previous studies show that currencies can vary (e.g., Bayham 1979; Kelly 1995;

Kuhn and Stiner 2001; Jochim 1988). Traditionally, optimal diets are thought to either minimize search and handling costs or maximize post-encounter return rates (Emlen

1966; Eastabrook and Durham 1976; MacArthur and Pianka 1966; Schoener 1971).

Search costs are defined as the time it takes to locate a particular resource, while handling costs are the time it takes to harvest that resource (adapted from Kelly 1995:78). Post- encounter return rate is the energetic net yield per unit of time after encountering a resource (adapted from Kelly 1995:78). In a time minimizer strategy, "fitness is maximized when time spent feeding to gather a given energy requirement is minimized," while an energy maximizer strategy optimizes fitness "when net energy is maximized for a given time spent feeding" (Schoener 1971:376).

There are, however, other currencies. Time (e.g., more time for other activities) and risk-avoidance (i.e., reduce potential exposure to predators, injury, or climatic stress) are 41

also important currencies (Winterhalder and Smith 2000:54). In fact, almost any measure

of a resource's value, such as protein content, fat value, nutrients, or personal prestige,

can be a currency (e.g., Belovsky 1988; Hawkes et al. 1982; Hawkes and O'Connell

1992; Hockett and Haws 2003, 2005; Kelly 1995). As discussed later, because so many

variables can be considered currency, it is possible for members of the same group to use

different value systems in their diet choice decisions (see Jochim 1988; Hawkes et al.

1997; Kuhn and Stiner 2001; Wilderhalder and Smith 2000).

A forager's ability to optimize is limited by a series of endogenous (internal) and

exogenous (external) constraints. Endogenous constraints include a forager's knowledge,

technology, and cognitive capacity. Exogenous factors, such as resource distribution

patterns, nutritional content of resources, and competition with outside groups and

organisms, also limit a forager's ability to optimize. Cultural traditions and beliefs can

also impose constraints. For example, while a particular resource may be plentiful and of

high quality, a forager may have to avoid that resource due to personal or cultural taboos

(e.g., Sahlins 1976; Jochim 1981; c.f. Hawkes et al. 1982). In most cases, constraints are

fixed, at least in the short term. An example of an independent constraint is a forager's

encounter rate with a particular resource.

Finally, the decision set involves the range of behavioral options. For diet breadth

models this usually involves "the diet combinations (or breadth) achieved by stepwise additions of resources which have been ranked by their pursuit and handling profitability"

(Winterhalder and Smith 2000). Diet breadth posits a cost-benefit relationship between return rates (profitability) and net acquisition rates (foraging efficiency). 42

As described by Bayham (1979), this relationship can be usefully depicted

graphically in terms of prey body size, caloric value (nutritional value) of a food item,

and the costs associated with searching and processing that item (efficiency) (Figure 2.1).

There is generally a linear increase in the value (energy gain per individual) of a food

item with body size, such that large food items provide substantially higher returns than

do small items. At the same time, there is a curvilinear relationship between processing

costs (i.e., time) and body size, such that for both very small and very large food items

processing time is quite high and generally lower for all other body sizes (see Byers and

Ugan 2005). Based on this relationship it is assumed that if the food item is very small and/or the associated processing time very great, that item is unlikely to find its way into a forager’s optimal diet. For the intervening food size classes, prey rank should increase with size.

Empirically derived post-encounter return rates and handling cost estimates are also available (e.g., Byers and Ugan 2005; Hawkes et al. 1982; Kelly 1995; O'Connell and

Hawkes 1981; Simms 1988). Kuhn and Stiner's (2001) recent review of the net yields

(kJ/hr) for various food classes highlights that body size, because of its strong correlation with post-encounter return rate, is an appropriate proxy estimate of prey rank in most cases (see also Broughton and Grayson 1993; Simms 1987; Stephens and Krebs 1986;

Szuter and Bayham 1989; Ugan 2005b). These data suggest that although seeds, nuts, and tubers are highly nutritious—often exceeding the caloric value per mass (kJ/kg) of game animals—the work required to collect, process, and cook them translates into relatively 43 low return rates for the time invested. In contrast, returns from game animals are very high for relatively minimal effort (Table 2.2).

Figure 2.1. Graphic depiction of the relationships between body size and (A) energy gain/individual, (B) processing costs, and (C) food rank (adapted from Bayham 1979: Figure 1).

Table 2.2. Net and Gross Energy Yields of Food Classes Consumed by Modern Foragers (Data from Kuhn and Stiner 2006:Table 1) Cases Net Yield (kJ/hr) Gross Yield (kJ/kg) Resource Class Used Mean Range Mean Std. Dev. Large mammals 4 63,398 36,000 - 75,115 6,980 1,383 Small mammals 14 16,034 1,672 - 56,317 6,980 1,383 Reptiles 3 15,850 12,435 - 17,556 4,489 715 Birds 3 4,472 961 - 8,255 -- -- Roots and tubers 14 6,120 418 - 26,133 2,926 1,680 Roots and tubers 9 10,412 3,695 - 23,333 2,926 1,680 Roots and tubers 13 1,882 1,045 - 2,300 3,136 2,338 Seeds and nuts 34 3,520 380 - 18,538 13,188 9,334 Seeds and nuts 9 6,508 1,203 - 24,933 13,188 9,334 Seeds and nuts 6 -- -- 19,372 6,250 Foliage ------1,250 819 Foliage 3 -- -- 1,534 186 Fruits ------2,403 1,436

In a series of papers, Stiner (2001, 2005; Stiner and Munro 2002; Stiner et al. 2000)

shows that relying solely on body size as a measure of prey rank potentially overlooks

significant differences in prey handling costs when harvesting animals of the same size.

This is because a forager's harvesting strategies must respond to different locomotor

habits and predator avoidance tactics for prey of comparable value. Stiner (2001, 2005)

suggests that relatively immobile prey, such as tortoises and shellfish, which are easy to

harvest, will be of higher rank than comparably-sized prey that are quick, agile, and/or below-ground dwellers (e.g., hares, partridges, and fish).

Key Predictions of Prey-Choice Models

Based on the relationships outlined above, prey-choice models provide several predictions (Table 2.3). It is anticipated that high-rank resources will be pursued whenever they are encountered (Charnov 1976; Emlen 1966; Heller 1980; MacArthur and Pianka 1966). Thus, a forager's encounter rate with high-ranked prey will influence diet breadth. Byers et al. (2005:1630) suggest there is a significant positive linear relationship between population density and encounter rates. When high-ranked resources are encountered often, diet breadth narrows; conversely, when encounter rates decline

(due to resource overexploitation, environmental change, or forager population growth) diet breadth expands.

Dietary expansion is a response to the increased search costs associated with trying to find the ever-diminishing high-ranked resources (Winterhalder 2001). There is, however, a point of diminishing returns for expanding diets (i.e., a point where no additional resources will be added to the diet). This occurs when the profitability of the next ranked resource to be taken is so low that adding it to a forager’s diet would reduce the overall net return rate for the entire set of resources taken (Bettinger and Baumhoff 1982;

MacArthur and Pianka 1966) (Figure 2.2). It is then that no additional items are added to the diet. 46

Table 2.3. Key Predictions Generated by Diet Breadth models (adapted from Kelly 1995; Smith 1982; Winterhalder 2001) 1. High-ranked prey will be taken whenever encountered; therefore, expansion and contraction to diet breadth fluctuates with availability of high-ranked prey 2. A resource item is added to or dropped from a forager's harvest set in order of its handling efficiency; an item is added if its profitability is greater than the net acquisition rate 3. Inclusion of a prey type in a forager's diet is dependent on the availability of higher-ranked prey, not its own availability.

Figure 2.2. An estimate of the optimal number of food items in the Ache diet that compares net energetic return of resources i (E i/h i) to (E/t), where E i = calories available in a unit of resource i, h i = handling time per unit of resource i, E = total calories acquired foraging, and t = search time (adapted from Hawkes, Hill and O'Connell 1982:Figure 1).

Recent Innovations to Diet Breadth Models

In the last couple of decades, much research has focused on situations in which traditional diet breadth models do not fully explain forager behavior. Overall, recent work shows that reliance on simplistic notions about resource encounter rates and resultant energy gains have limited explanatory value. Much of this new thinking originates from various ethnographic investigations (e.g., Hawkes and Bliege Bird 2002; Hawkes et al.

1997; Winterhalder and Smith 2000), although archaeologists are also making significant empirical and theoretical contributions as well (e.g., Stiner 2001, 2005; Stiner et al.

2000). The following section briefly discusses these recent changes using two examples: sexual division of labor and differential provision strategies, and risk and prey choice.

Sexual Division of Labor and Differential Provisioning Strategies. Early diet breadth studies treated hunter-gatherer subsistence as if it originated from the actions of undifferentiated foragers. However, foraging agendas can vary based upon the age, gender, and status of the forager (Bettinger and Baumhoff 1982; Elston and Zeanah 2002;

Engen and Stenseth 1989; Jochim 1988; Kuhn and Stiner 2006). In addition, humans often forage collectively and share the byproducts of their efforts among different members of their group (e.g., Hawkes 1991; Winterhalder 1986, 1997; Winterhalder and

Smith 2000). The realization of the full complexity of provision strategies utilized by hunter-gatherers exposes the necessity for researchers to add new components to the 48

traditional foraging models, such as the roles gender and age play in labor organization

and food sharing among foragers (e.g., Beaver 2007).

A significant amount of recent work has gone into understanding how sexual division

of labor influences foraging decisions. Although the nature of sexual division of labor

varies among groups, current research suggests men seem to focus more on hunting large

game, while females spend more time procuring small game and plant resources (Hawkes

1996; Hawkes et al. 1982; Hawkes and O'Connell 1992; Kuhn and Stiner 2006;

Waguespack 2003, 2005). While this sexual division of labor may allow a single group to

efficiently exploit different resources and different niches, which would be otherwise

difficult for a single forager to utilize at the same time (Elston and Zeanah 2002; Hawkes

et al. 1997; Jochim 1988), it is clear that the traditional emphasis on energy return as a limited currency does not explain all situations of sexual division of labor.

Hawkes and O'Connell (1992; Hawkes et al. 1997) show that the emphasis males place on hunting large game exceeds levels that are expected based solely on anticipated return rates. In addition, when large game is taken, it is commonly shared among households (e.g., trade and exchange, kin provisioning, theft/scrounging) to a degree greater than would be expected if the goal was simply to maximize family or personal caloric gain (Hawkes et al. 1982).

These unexpected results suggest that a currency other than post-encounter return rates—perhaps prestige or social networking—drives male interest in large game hunting.

The individual may be attempting to increase personal status, mating opportunities, or

connections with outside groups by hunting large game (Broughton and Bayham 2003; 49

Dean 2001; Hawkes and Bliege Bird 2002; Hawkes et al. 1997; Hildebrandt and McGuire

2002; McGuire and Hildebrandt 2005; Winterhalder and Smith 2000).

An implication of these findings is that the foraging decisions of men may not be solely directed towards increasing the long-term economic survival of the family group.

This conclusion does not suggest that meat is unimportant to group survival. Rather, a male’s success in hunting does not necessarily or directly benefit other members of his immediate family (Hawkes and Bliege Bird 2002; Hawkes et al. 1997). Women and children can provision themselves with meat through various means, such as exchange, gifts, and theft. None of these processes are easily modeled using only traditional foraging models.

Human propensity for sharing meat after a kill and cooperative hunting also point out a number of other weaknesses with traditional optimal foraging models (Beaver 2007;

Hawkes 1991; Winterhalder 1986; 1997; Winterhalder and Smith 2000). Because meat acquired through sharing or capture communally does not come with the same energetic costs as it would for the solo forager, the currency in networks is usually the inherent value of the food item, not energy gained or expended in getting it. As a result, certain individuals may be consuming shared food disproportionately to amount of energy they are expending to acquire food.

Traditional diet breadth models are probably more useful for explaining the foraging habits of females. Female foraging is very important to the survival of the family group, especially young children. The resources procured by females are often low-ranked foods and located in close proximity to settlements (Hawkes and O'Connell 1992; Hawkes et al. 50

1997). O'Connell et al. (1999) even suggest that the foraging behavior by females in the extended family group (i.e., grandmothers) provides a number of benefits, including increased child survival and female fertility.

Waguespack (2003, 2005) highlights some of the challenges in modeling the internal partitioning of labor in foraging decisions. She notes that a group’s reliance on one aspect of foraging, for example male-based hunting, can influence foraging decisions in other dietary decisions, such as female-based plant gathering. Her findings also suggest that these decisions may not be based purely on the economic value of the resources being utilized, such as would be expected from traditional foraging models.

For example, Waguespack (2005; see also Keeley 1988) shows that groups with higher dependence on hunted resources use a narrower range of plant resources than do groups with lower reliance on hunting, based on her study of 71 modern hunter-gatherer groups. Not only are fewer types of plant resources used as the contribution of hunting to the diet increases, but the types of plants used are more commonly those that are easy to process such as fruits, vegetables, and roots. Intensive use of hard-to-process plant resources, such as nuts and seeds, is most common in groups that do relatively little hunting.

It is important to note that this relationship is not entirely driven by environmental availability of plant resources. Waguespack (2005) explains that preferential focus on easy-to-process plants remains when the contribution of hunting is high in response to the need for women to spend more time conducting non-subsistence-related activities. 51

The contributions of children to forager diets are yet another problematic area for applying traditional diet breadth models to studying subsistence labor division (Blurton

Jones et al. 1994; Hawkes et al. 1995). The problem is highlighted by comparing the difference between two well-studied African populations: !Kung and Hadza (see discussion in Winterhalder and Smith 2000). !Kung children make almost no contribution to provisioning their group, while Hadza children contribute very significantly to group survival at least seasonally. The traditional goal of energy optimization does not explain these differences.

If one accepts that the foraging choices and yields of children are different than those of adults, it may be easier to explain the differences between the !Kung and Hadza.

Blurton Jones et al. (1994), for example, show that the difference in the children’s behavior relates to differences in the organization of the different environments and the potential dangers confronting the children when they are away from camp. These kinds of study also highlight how limitations in children's upper body strength and foraging experience affect their foraging decisions.

Risk and Prey Choice. Traditional foraging models are limited, in part, because they assume a uniform environment and set aside questions about the predictability of resource availability in an environment. However, real-world foragers must constantly respond to risks presented by differences in resource abundance and availability. This study uses Winterhalder et al.'s (1997:302) definition, which suggests that "risk is unpredictable variation in the outcome of a behavior, with consequences for an organism's fitness or utility." 52

Winterhalder (1986; Winterhalder et al. 1997), building on earlier work by Stephens and others (Stephens and Charnov 1982; Stephens and Krebs 1986), developed a risk- minimizing model, called a Z-score model, that tried to predict how foraging decisions should respond to differing levels of risk. The underlying assumption of this model is that each foraging choice has a mean (expected) food reward, and due to fluctuations in prey abundance and other environmental factors, there is a variance about the mean value of rewards.

Winterhalder's (1986) simulation studies found that the optimal risk-minimization diet choice converges on the rate-maximizing choice. These simulations show that in a context where the availability of high-ranked prey is relatively constant, expanding one's diet actually increases the risk of a dietary shortfall, more so than reducing diet breadth.

The underlying cause for the increased risk is that the broader diet has increased costs in terms of foraging efficiency. Under good conditions these increased costs can be easily handled; however, in dire circumstances these increased costs can have severe consequences for foragers, such as food shortage. However, as discussed above, when the abundance of high-ranked foods decreases, then a forger takes on greater risk to rely predominantly on them.

According to Elston and Zeanah (2002:115) the implication of the Z-score model is that if the mean return rates for patch exceed a forager's survival threshold (i.e., starvation level), then the pursuit of low-variance resources (e.g., plants and small game) minimize a forager's risk of falling below the survival threshold. However, when mean return rates do fall below that threshold, a forager has a better chance of surviving by gambling on 53

high variance resources (e.g., large game). These results predict that in very marginal

(low return rate) environments prey choices should be limited to high-ranked resources;

in richer (high return rate) environments prey choice should be broader.

CENTRAL PLACE FORAGING PATCH-CHOICE MODEL

There are two problems with applying traditional prey models to real-world situations. First, in most cases the fine-grained search assumption of prey models is often violated. This assumption states that distribution of different resource types are even across the landscape in proportion to their density (e.g., Heller 1980; MacArthur and

Pianka 1966; Stephens and Krebs 1986). This is critical to the model because it allows a forager to simultaneously search for different resources; the chance of encountering any particular resource is independent of previous encounters with that resource or any other type of resources. Actual resource availability, however, is heterogeneous (or patchy) both in space and time, and foragers must move between these patches to meet their subsistence needs (Charnov 1976; MacArthur and Pianka 1966; McNamara 1982; Wien

1976). Foragers must, therefore, make decisions about which patch to use, how long to use it, and which resources to exploit while in the patch (Charnov 1976).

The second problem is that traditional prey models assume that foragers consume or collect important resources as they move through the landscape; transportation costs are assumed to be zero. However, human and many non-human foragers invest substantial efforts into travel across the landscape in search of resources and then transporting those 54

resources back to their residential base (Houston and McNamara 1985; Orians and

Pearson 1979).

To understand how foraging activities are organized in response to the differential

distribution patterns of resources, a number of researchers have used the central place

patch choice model (Charnov 1976; Elston and Zeanah 2002; MacArthur and Pianka

1966). This theory represents a combination of Orians and Pearson's (1979) central-place

foraging model and Charnov's (1976) marginal value theorem . The Orians and Pearson's models attempt to address (1) how a forager selects patches, prey, and load size for transport if he or she has a central place to which to return, and (2) where a forager locates his or her central patch. The model assumes a landscape composed of multiple homogeneous resource patches scattered at various distances from a forager's residential base (see discussion in Cannon 2000). The marginal value theorem (see also Oaten's

[1977] revised "equal rate" theorem) examines how long a forager will stay in a patch and the effects foraging activities will have on the availability of resources within the patch.

The core theorem of this model is that an optimal forager will leave a patch when the marginal net yield from that patch drops below the average rate of return for the habitat minus the average travel costs to a new patch (Charnov 1976) (Figure 2.3). Habitat is defined as regional relatively homogeneous environmental zones that are distinguishable from other similarly scaled zones in terms of the resources they offer a forager

(Winterhalder 2001:19). Expressed in slightly different terms, as long as the anticipated returns from staying in the same patch exceed the costs associated with traveling and exploiting a new patch, a forager can be expected to remain within the original patch. 55

Figure 2.3. Patch choice model for central place foragers (adapted from Kelly 1995: Figure 3–5). The gain function provided in both patches A and B is represented and maximized energetic return per total (roundtrip) time (Ta and T b) spent in the patch is given by the dashed line tangential to the gain function beginning at the origin of the graph. Richer patches have taller gain functions and the patch that provides the highest overall rate of energy is the one that produces the steepest slope of the line tangential to the gain function.

Two assumptions underlie this theorem. First, the relevant cost-benefit relationship is

a comparison between total costs (energy and time) of traveling to the patch plus the time

in the patch and the marginal energetic gains obtained from that patch. Second, over time

the rate of return a forager obtains from a patch rises from zero at a negatively

accelerated rate to an asymptote, also called a diminishing marginal return curve. That is,

upon entering a patch the energetic gains accumulated by a forager diminish the longer he

exploits the patch. 56

Although quite simple, the central place foraging theory has important implications for understanding (1) which type of patch a forager will choose to utilize, (2) how long a forager will stay in a patch, (3) what types of prey are likely to be exploited once a forager arrives at a patch, (4) what type, if any, of field processing will be done in the foraging patch, (5) whether the foraging party will consist of a logistic selection of personnel or the entire residential community, and (6) how foragers respond to the risk of food shortage (Table 2.4). The following sections explore examples of how central place foraging models have been applied to several of these issues.

Table 2.4. Predictions Generated by Central Place Foraging Patch Choice model (adapted from Orians and Pearson 1979; Smith 1982; Winterhalder 2001) 1. Forager will leave patch when its declining marginal rate of return equals the mean net acquisition rate for all other patches 2. Increased productivity of habitat reduces patch residence time 3. Foraging efficiency will increase with distance between patch and central place 4. If patch quality remains constant, forager will increase load size with distance between patch and central place 5. Location of central place should be one that minimizes travel time; however, other factors (e.g., access to key resources and protection) also influence the decision 6. Upon entering a patch, foragers will initially exploit the highest ranked resources, adding less profitable resources types as high-ranked resources become more scarce 7. Residential moves will occur more often when (1) relocation costs are low, (2) resource depression around central place is rapid, and/or (3) alternative sites provide high initial rates of returns.

Choosing a Patch and which Resource to Exploit in that Patch

One of the primary questions asked by central place foraging theory is which patch

will an autonomous forager use and what variables influence his or her decisions. In most

cases concerning patch choice, central place foraging theory is similar to traditional diet

breadth models in that both assume that variables such as post-encounter return rates,

travel costs, and handling costs strongly influence a forager's decisions. Accordingly, a

forager will consider a patch containing a high density of high-return resources more

profitable than one containing primarily low-return resources. Moreover, the model

suggests that upon entering a patch, regardless of its quality, foragers are expected to

initially exploit only the highest-ranked resources, adding less profitable resources types

as high-ranked resources become scarcer.

However, there are important implications that marginal value theorem brings to

understanding how long a forager will remain in a patch (Emlen 1973; Heller 1980).

First, duration of patch use will increase as habitat quality decreases. Second, foragers

will stay in each patch longer as the travel time between patches increases.

At some point, however, depletion of resources within a patch forces the forager to

leave and move elsewhere. At this point, a forager has three choices: (1) exploit a lower

quality patch nearer to the central place than the original patch; (2) exploit a lower

quality patch farther away from the central place; or (3) exploit a higher quality patch farther away (Cannon 2000:323). While the patch choice model does not predict which of 58 these the forager will choose, it does allow one to understand the implications of each choice.

While travel to the higher quality patch seems the most likely choice, because of the high energetic costs of making the round trip between the residential base and foraging area and the time spent in the foraging area, it may actually be more efficient for an autonomous forager to utilize a lower quality patch nearer the residential base. I will return to this issue later in this chapter.

The central place foraging model does, however, predict that if the forager chooses to travel to a patch far from a residential base in search of food or pursuit of prey over great distances, he or she will use different selection criteria concerning which resource to exploit in those circumstance than when traveling short distances for food. For example, the farther a forager travels from their residential base, the less he or she is interested in pursuing small game, and at some point the forager will exclusively pursue large game

(Schoener 1971). This decision is expected because the likely energetic return from small game would be insufficient to make up for the effort already expended. Or as Bettinger et al. (1997:888) summarizes the situation, “larger rewards are needed to justify longer trips.”

RESOURCE DEPRESSION AND FORAGING EFFICIENCY

The final topic considered in this chapter is how foragers' activities structure the natural environment in which they live. As discussed already, patch-choice models 59 predict that once all patches a forager is likely to exploit are in use, continued exploitation in those patches will eventually result in a decrease in the relative abundance of exploited resources (Cannon 2000:323). Such extensive depression will reduce overall foraging efficiency and force foragers to modify their exploitation strategies. This study follows Charnov et al.'s (1976:247) definition for resource depression as the lowering of capture rates for a prey within a given habitat. In many cases, the lowered capture rates result from a real decrease in prey abundance.

Most studies that consider this issue primarily focus on the degree to which prehistoric foragers' exploitation patterns result in resource depression. However, as recognized by Charnov et al. (1976:247), this issue actually involves several separate topics, including (but not limited to): (1) What processes might cause resource depression? (2) How do prey respond to resource depression? (3) Which characteristics of prey most affect their depressibility? (4) What are the implications of resource depression to a forager? Each topic is discussed in turn.

Processes that Depress Prey Populations

Understanding the factors that lead to prey population depression is one of the most important issues explored in this research agenda. In general, it is assumed that prey abundance responds to changes in environmental conditions and/or hunting pressure.

Hunting (Harvesting) pressure. Not surprisingly, many researchers emphasize the role that a forager's predation or collection activities have on depressing prey populations 60 or resource availability. Because I have previously discussed how the marginal value theorem posits that the longer a forager exploits resources within a patch the greater decrease in the availability of that resource, this section will discuss the other effects predation pressure has on resource depression.

Charnov et al. (1976) note that predation is not the only process by which resource depression can occur; for example, disease and competition can have strong effects on prey populations. They also acknowledge that behavioral depression can occur, whereby prey changes their behavioral or spatial positioning to make it harder for a predator to capture them. For example, animal populations may aggregate into large groups (e.g., herding in land mammals or flocking in birds) to reduce the chance a predator will successfully capture its prey. Alternatively, prey may abandon an area where predators are common, forcing predators to work increasingly harder to locate prey. Martin and

Szuter’s (2004) study of the faunal resource structure of western North America at the time of the Lewis and Clark expedition can be viewed as an example of the effects of behavioral depression on human hunters. In most cases, this behavioral depression is usually temporary and unlikely to result in long-term shifts in predator-prey relationships.

It is likely that predators will quickly respond to these changes in prey behavior (Charnov et al. 1976). Next, I will focus on factors that result in longer-term shifts in prey availability.

Environmental Conditions . Population size and density are linked to a complex network of density-dependent and density-independent influences, which include environmental conditions, behavioral responses, and physiological factors (Krausman et 61

al. 1999; Pianka 1983; Smith 1966). A complete literature review on the ecological

constraints imposed on animal populations is beyond this study, but I will briefly

summarize some key findings related to studies of North American artiodactyl

populations.

In general, researchers who have studied this issue have relied on data from wildlife ecology studies that examine factors affecting prey population size and mortality patterns

(e.g., Byers et al. 2005; Mysterud et al. 2007; Osborn 1983, 1993, 2003). Specifically, these studies examine the nature of the relationship between prey population size and climatic conditions through regression analyses of multiyear census data for various species using several climatic variables, such as annual snowfall, total precipitation, or winter severity.

Although the results of these studies are rather variable, they suggest that no single factor controls population size and density when it comes to North American artiodactyls.

For example, in a series of studies Osborn (1983, 1993, 2003) demonstrates weak inverse correlations between Great Plains whitetail deer, elk, horse, bighorn sheep, and pronghorn population counts and fertility and total snowfall as an index of winter severity. Osborn concludes that winter climatic conditions cause reduced access to browse and it is this factor that leads to decreased reproductive success and increased mortality. However, he recognizes that other factors, such as disease, supplemental feeding, and hunting pressure, also greatly influence population size.

In a separate study, Byers et al. (2005) report the results of several wildlife analyses that suggest mule deer, bison, pronghorn, elk, and mountain sheep populations from arid 62

portions of the Great Basin and American Southwest are sensitive to variation in

temperature and precipitation. Specifically, these studies suggest positive effects of cool

and moist weather and negative effects of hot dry conditions on artiodactyl reproductive

success. This study concludes that the key link between artiodactyl population growth and climate pattern stems from the effects temperature and precipitation have on forage quality and availability of drinking water (Byers et al. 2005:128). Finally, in a potentially relevant study, Mysterud et al. (2007) demonstrated that for European Bison bonasus

populations colder and snowier winters correlated with higher mortality rates than did

warmer and drier ones.

In summary, although the extant data is far from definitive, researchers should expect

that long-term shifts in seasonal extremes in precipitation, especially drought or snowfall,

and temperature will likely affect prey population and density levels.

Effects of Resource Depression on Prey Populations

Changes in the absolute and relative frequency of prey population are not the only

effect of resource depression. As chapter 4 provides a more detailed discussion of this

issue, this section only briefly considers how changes in the body size or ontogenetic age

structure of prey populations often accompanies decline in population size when a

population is depressed (e.g., Lyman 2004a; Mannino and Thomas 2002; Munro 2004;

Stiner et al. 2000; Ugan 2005). The underlying prediction is that high levels of predation

result in a significant temporary decrease in mean body size of a prey species with slow 63 rates of population growth. However, if this hunting pressure is maintained for long periods of time then the change to the population can potentially be permanent. While this body size response is not universal, evidence for this pattern comes from a series of studies that examine multiple vertebrate and invertebrate species (Coltman et al. 2003;

Edwards 1967; Mannino and Thomas 2003; Reznick and Bryga 1987; Reznick et al.

1997; Stiner 2005; Stiner et al. 2000).

This change in body size is thought to result from preferentially removing too many adult members from a population. The elimination of the older (and often larger) members of a population results in a decrease in the average age of that group, and a corresponding decrease in mean body size. From the perspective of the archaeologist the changes to living animal populations and to archaeological assemblages will be asynchronous. That is, if human hunters preferentially take the bigger animals, evidence for this size reduction will only show up in archaeological assemblages some time later.

A key aspect of this argument is that unless intensive hunting pressure is maintained for a very long period, this change in body size only affects means body size for the group, not the maximum potential size attainable by an individual (Stiner et al. 2000).

Edwards (1967) notes that extended hunting pressure could eventually lead to genetic selection for a smaller optimum body size, and recent genetic work by Coltman et al.

(2003) on the effects of trophy hunting on bighorn sheep population lends at least some support to this hypothesis.

Because a change in the mean age of a prey population is the underlying cause for the decline in body size in hunting pressure models, independent measures of age structure 64 may isolate human exploitation as a primary factor in body size change. Hunting simulation studies and analyses of modern non-human and human death assemblages

(Stiner 1994, 2005; Wolverton 2001) demonstrate that intense predation pressure on artiodactyls can produce mortality age profiles in which juvenile individuals are proportionally more abundant than in populations that experience no harvest pressure.

Prey Characteristics that Affect Resource Depression

Several recent studies have used foraging theory and population ecology models to investigate how foraging tactics and resource population structure make particular species susceptible to overexploitation. Because of the lack of a long-term record of foraging intensity and population dynamics, researchers commonly investigate this issue through computer simulation studies.

Stiner et al. (1999, 2005; Stiner et al. 2000) and others (Beaver 2007; Winterhalder and Lu 1997) have shown that reproductive rates of prey species greatly influence predation strategies as well as diet breadth sustainability. Relatively low levels of predation pressure (>7% mortality) on animals with low rates of reproduction, such as tortoises and certain species of shellfish, can negatively affect prey populations and therefore have significant impacts on the long-term exploitation of that resource (Stiner et al. 2005). Large land mammals also have problems maintaining stable populations at relatively low levels (2 –11% mortality) of human exploitation (e.g., Beaver 2007; see also Owen-Smith 1988; Winterhalder and Lu 1997). In contrast, prey with very high rates 65

of reproduction, such as hares, rabbits, and ground birds, can maintain viable populations

even when subjected to very high levels of predation (18–66% mortality) (Stiner et al.

2005).

Implications of Resource Depression for Foragers

A number of studies have suggested that increases to diet breadth over time is a common response to a reduction in the availability of high-ranked resources caused by overexploitation (e.g., Bayham 1979; Belovsky 1988; Broughton 2002; Broughton and

Grayson 1993; Cannon 2000; Grayson 2001; Hildebrandt and McGuire 2002; Mannino and Thomas 2002; McGuire and Hildebrandt 2005; Munro 2004; Nagaoka 2001; Stiner et al. 2000; Stiner 2001; Stiner and Munro 2002; Ugan 2005). Surprisingly, a number of studies, especially from the western United States, recognize an opposite trend, although there is little agreement on explanation for the patterns (Butler and Campbell 2004; Byers et al. 2005; Hildebrandt and McGuire 2002; McGuire and Hildebrandt 2005; Speth and

Scott 1985).

Broughton (2002:63; see also Cannon 2000) suggests that, based on prey and patch choice models, depression of high-ranked prey within local resource patches will result in a forager more intensively utilizing lower-ranked prey in those patches and/or devoting the effort to travel to far-removed less-depleted patches. In both strategies foragers sacrifice a substantial degree of foraging efficiency. Interestingly, an implication of

Broughton's prediction is that resource depletion will initially be marked by a decline in 66

high-ranked prey in patches near the residential base, but over time this trend may change

towards a gradual increase in high-ranked prey use as foragers increasingly utilize more

distant, less-depleted patches (Figure 2.4). However, if these distant patches do not

contain large game or are already being heavily exploited then there should be indication

of a temporal decline in large game abundance.

Figure 2.4. Distribution of Deer Index ( ∑NISP Deer/ ∑NISP Mammals) values by stratum for Emeryville Shellmound (adapted from Broughton 2002:70).

The key issue here is to confirm that the later increase in use of high-rank prey reflects exploitation of distant, but formerly ignored, patches. In his study of the 67

Emeryville fauna, Broughton (2002) argued that the increased abundance of deer toward

the end of the site's occupation was derived from distant patches based on his study of the economic utility of the skeletal part representation at the site. His work demonstrated that in the later years, foragers responded to increased transport costs associated with carrying deer remains from distant patches by leaving low-utility portions of the carcasses at the kill locality and only transporting the most economically valuable skeletal parts back to camp.

ARCHAEOLOGY AND FORAGING THEORY

Now that we have reviewed the major assumptions and predictions of prey choice, patch choice, and resource depression models, it is necessary to briefly consider the limitations of applying these models to actual archaeological problems. The problem with applying these models to archaeological data usually is a result of either the indirect nature of observations about the structure and organization of past environments and human behavior and/or the fact that the archaeologist's view on past behavior is simply not as robust as a biologist's or anthropologist’s view of a living group (Binford 1981;

Stiner 1994, 2005).

One important problem is that archaeological data is commonly the result of time- averaged behavior (Beaver 2007; Grayson and Delpeche 1998; Lyman 2003; Stiner

1994). That is, the archaeological record is often a palimpsest of the activities of multiple foragers over significant periods of time. Unlike a wildlife biologist or anthropologist, 68 who can observe the result of a single forager's actions, archaeologists can only view the result of a combination of decisions made by many foragers, possibly under very different environmental or social settings. In this case, for example, an archaeological study of diet breadth can at best expect to measure total or maximum diet breadth of the combined population rather than some average diet breadth for each occupation

(Broughton and Grayson 1993; Madsen 1993).

The interpretative problems for the archaeologist are compounded even further due to the complex taphonomic histories of many deposits. In studying prehistoric diet breadth, for example, the analyst must exclude species that are unintentionally deposited at the site, which is not always possible. Sample size problems can also create problems for researchers trying to study diet breadth and patch choice. Madsen (1993), for example, demonstrates that taxonomic diversity of prey species at archaeological sites is often a function of size of the zooarchaeological sample recovered from the site; therefore these data are not useful measures of diet breadth. Although Broughton and Grayson

(1993:334) agree that taxonomic richness measures do not measure diet breadth, they do suggest that these measures can successfully be used to "measure the difference in the number of taxa exploited through time."

Smith (1982:629) and others (Beaver 2007; Kelly 1995:94–95) also suggest that precisely and quantitatively applying foraging theory models to archaeological situations is difficult because archaeologists generally lack precise data on several key variables, such as foraging costs, resource encounter rates, mean travel time between patches, return rates for all potential resource patches, and harvest rates of prehistoric foragers. A bigger 69

problem with applying these types of models to human subjects is that human foragers do not encounter patches randomly. As Kelly (1995:92) explains, human foragers often make decisions about where to forage prior to leaving camp and then go directly to those patches, ignoring other patches along the way.

In addition, human provisioning behavior rarely involves an autonomous forager moving about the landscape directly collecting food only for himself or herself, as envisioned by traditional prey choice or central place forager models. Instead human hunter-gatherers often form task groups that work together to accomplish a particular task and then share the results with their entire social groups, which implies provisioning for non-foraging members. Furthermore, these foraging subgroups are often organized around age, sex, and/or skill level difference among the group members. This type of task group organization not only results in particular subgroups being responsible for certain aspects of the foraging regime, but also may result in those different subgroups holding different agendas for their foraging behavior.

This said, there are some advantages in applying foraging theory to archaeological data, at least as long as caution is used and some sacrifice in precision is accepted. First, the archaeological record represents an appropriately broad evolutionary and ecological time-scale to observe long-term changes or adaptations. In contrast, contemporary or ethnographic observations can be often be too short in duration or too fine-scale to observe significant change. Second, the time-average nature of the archaeological record will diminish the overall affect that situational or chance variation has on the patterns 70 being observed. However, time-averaging of the record also reduces inter-individual or intra-group variation in foraging behavior.

Gremillion (2002:144) recommends that researchers try to create models that accurately predict the general direction of change (i.e., increase or decrease in diet breadth) rather than the specifics associated with that change (see also Broughton and

Grayson 1993). The explanatory value of these models will likely improve as more detailed information is gained on past environmental conditions or site formational history (Madsen 1993).

SUMMARY AND PREDICTIONS

Outlined below, but discussed in more detail in the following chapters, is a suite of general predictions concerning the behavior of rational foragers that can inform study of

Paleoindian behavior. These predictions were outgrowths of the behavioral ecology models and theorems outlined above, and try to accurately predict how forager behavior would change in response to specific environmental conditions.

The specifics of each prediction were informed by work of several earlier researchers, who have used prey models developed from foraging theory to predict how prehistoric hunting decisions would respond to faunal resource variation (e.g., Broughton 1994;

Byers et al. 2005; Cannon 2000). As with earlier studies, the current predictions rely on the assumption that prey abundance, responding both to diachronic (temporal) and synchronic (regional) environmental change and to hunting pressure, greatly influences 71

Paleoindian hunting decisions on multiple temporal and spatial scales. In particular, this

study explores Paleoindian landscape use, with an explicit interest in determining

whether differences in regional environmental (i.e., habitat) structure influences diet

variation and change (e.g., Charnov 1976; MacArthur and Pianka 1966).

Based on prior work that considered how spatial and seasonal variation in resource

availability and predictability affects foraging decisions (see Elston and Zeanah 2002).,

the predictions below were constructed to consider how the movement of Paleoindian

foragers between different landscapes would affect the level of risk tolerance foragers

accept. These models also try to consider how sexual division of labor and strategies of resource sharing contribute to subsistence and land-use decisions, based on ethnographic observations concerning how age, sex, and status differences affect foraging decisions.

Extended discussions of the testable hypotheses concerning diet change and land-use variation are discussed more in chapters 5, 6, and 7. Chapter 3 summarizes specific predictions concerning variation in bison use. Below are summarized some of the key predictions of how foragers move about landscape and how diets change in response to shifts in habitat structure.

• Paleoindian residential mobility will change with resource richness and

predictability of the landscape being occupied.

o Foragers occupying habitats (e.g., alluvial valleys and foothill/mountains)

with a high density (closely spaced) of high and moderate net yield biotic

resources (plants and animals) and abundant abiotic resources (stone and 72

water) will exhibit low residential mobility and high site occupation density.

As a result, these habitats will contain high proportions of camp and workshop

localities, and lithic assemblages dominated by locally available chipped stone

resources.

o Foragers occupying habitats (e.g., plains/rolling hills) with a low density

(widely spaced) of high and moderate net yield biotic resources and a low

abundance of abiotic resources will exhibit high residential mobility and low

site occupation density. As a result, these habitats will contain high

proportions of short-term special function sites (e.g., kills) and lithic

assemblages dominated by exotic chipped stone resources.

• Paleoindian diets will reflect the landscape differences in distribution of food

resources and changing land-use strategies.

o Large game hunting will be conducted in all habitats and throughout the

Paleoindian period in proportion to availability of large game. Thus, when

large game is abundant it will play a prominent role in the Paleoindian diet;

alternatively, when large game is scarce lower-ranked resources will play a

more prominent role in the diet.

o Zooarchaeological assemblages in alluvial and foothill/mountain habitats will

exhibit the highest diversity in diet breadth, with mixed use of high rank/high

risk (i.e., large game) and low rank/low risk (i.e., plants and small game)

resources. 73

o Assemblages in plains/rolling hill habitats will exhibit the lowest diversity in

diet breadth with an emphasis on use of high rank/high risk resources.

• Because of their low population density, Paleoindians had relatively limited

impacts on the environments and prey populations.

o Low levels of hunting pressure and landscape modification suggest that

Paleoindians were not largely responsible for the extinction of Pleistocene

megafauna.

o The greatest evidence for resource depression will be seen in areas of high

occupational density (e.g., alluvial valleys and foothill/mountains) and

observed least in low-occupational-density habitats (e.g., plains/rolling hills). 74

CHAPTER 3: BISON PALEOBIOLOGY

The testable hypotheses outlined in the previous chapter detail general patterns

expected for Paleoindian subsistence and settlement strategies. These hypotheses were

not scaled to any time period or particular resource. In contrast, this chapter takes a more

detailed look at one particular resource: bison. This chapter will be a review of the

literature on bison ethology and biology in order to understand the parameters of the

relationship between environmental conditions and bison behavior and biology. This

literature review is then used to construct testable hypotheses of human-bison

interrelationship that are relevant to understanding Paleoindian bison exploitation

strategies.

Focusing on bison is appropriate for several reasons. As discussed more fully in

chapter 6, although the debate still rages about the degree to which prehistoric

populations on the Great Plains relied on bison to satisfy their basic subsistence and

technological needs, bison was an important prey species (Cannon and Meltzer 2004;

Frison 1991; Meltzer 1988, 1993, 1995; Waguespack and Surovell 2003).

Unquestionably, bison remains dominate many Plains and Rocky Mountain Paleoindian

zooarchaeological records, so archaeologists need to have at least a rudimentary

understanding of bison hunting and butchering behaviors (Hill 2007). An important topic

for archaeologists to understand is how prehistoric hunters responded to the natural

variation in the behavior and physiology of their prey (Frison 2004). Basic questions that hunters must answer to successfully hunt bison include: When are the best (and worst) 75 times to undertake a hunt? How will bison react to the hunters (will they flee or stay and fight)? How does the condition of bison vary throughout the year? Did bison migrate in and out of the hunters’ exploitation area?

This chapter considers four broad research topics of bison behavior and physiology: social organization, reproduction, mobility, and physiology. As seen in table 3.1, these general issues include a broad suite of behavioral and physiologic issues related to the study of modern bison. More importantly to archaeologists, these issues have direct implications for zooarchaeological studies in terms of (1) kill composition and size, (2) season of mortality and number of kill events, (3) carcass butchering strategies and intensity, (4) isotopic signals, (5) geographic distribution of prey, and (6) predation pressure.

APPROACHES TO STUDYING BISON ETHOLOGY AND PHYSIOLOGY

Archaeologists have traditionally taken two approaches to study bison behavior and biology (Bamforth 1987). They either utilize historical documents recording the location and characteristics of bison herds or rely on the findings of modern ecological studies of bison biology and behavior.

Historic records (e.g., Allen 1876, 1877; Arthur 1975; Dary 1974; Garretson 1938;

Hornaday 1889; Jones 1899; Seton 1910; Whitney et al. 1904) have the advantage of recording the actions of bison before their numbers were greatly reduced in the late 19th 76 century and their distribution was confined to private ranches and a few protected reserves in the United States and Canada.

Table 3.1. Modern bison research topics and their implications for archaeological interpretations Research Behavioral and Physiologic Archaeological Implications Topic Issue • kill size Social • type of social group • herd composition organization • group membership • procurement scheduling • isotopic studies • nature and timing of • season of mortality mating and partition • number of kill events • physical condition of bulls • encounter rates with prey and cows • carcass butchery intensity • agnostic behavior and the rut • herd composition and size Reproduction • gestation length • dangers associated with making kills • factors affecting birth synchronization • ability of hunters to direct herd to kill site • group size and replacement rates • effect predation has on herd composition • fecundity • encounter rates with prey • annual and seasonal • geographic distribution of movements Mobility prey • factors influencing • herd composition mobility • isotopic studies • seasonal variation in condition • factors influencing • procurement scheduling Physiology condition • carcass butchery intensity • age and sex variation in condition

At best, historic accounts represent qualitative short-term encounters with bison.

However, these observers were usually not formally trained in animal behavior or 77

environmental studies. Also, there was no systematic collection of this information

(Bamforth 1987, 1988; Lott 2002). Rarely are the accounts based on observations of the

same herd throughout a single year. In addition, these studies often do not provide

adequate information to allow the reader to evaluate whether observations were broadly

representative or to reconstruct the ecological context of those events. As a result,

conclusions drawn from historic accounts are at best impressionistic and at worst

fictional. That said, some researchers have found that, if used with care, historic records

have some value for understanding bison behavior (see Soper 1941; Speth 1983; Wheat

1972).

While today it is not possible to study bison in unmanaged habitats, modern

ecological studies can inform archaeologists on many aspects of the adaptations and

physiology of bison. Their greatest value is to identify the conditions to which bison are

adapted and describe how bison respond to changes to those conditions.

The value of modern ecological studies is limited by the fact that modern bison are

not entirely comparable to prehistoric bison (Lott 2002). Bison nearly went extinct in the late 1800s and early 1900s, and all modern bison are descendants from small founder herds (Berger and Cunningham 1994; McHugh 1958; Pucek et al. 2003; Soper 1941). As a result, modern bison are highly inbred and genetically homogeneous compared to their

prehistoric ancestors. In addition, all modern herds are managed to one degree or another,

either in terms of their feed, foraging areas, population size, sex ratio, and/or breeding

strategies. Lott (2002) estimates that 90% of bison in North America are domesticated or

being domesticated. This transformation in the nature of human-bison interaction is not 78

limited to private ranches, but includes public bison herds through enthusiastic

management, predator control, and genetic dilution. Although the full implication of

these changes is not fully understood, it is true that “ Bison bison is the only wild animal that is not allowed to live as a wild animal...anywhere in its original range” (Lott

2002:201).

Despite these limitations, ecological studies of modern bison provide the best avenue for understanding the causal link between bison behavior and environmental stimuli.

Modern ecological studies are not, however, simply heuristic devices or sources of analogies for understanding the actions of prehistoric bison. As Osborn (1993:155) states, ecological studies “are essential components of a more inclusive explanatory framework.

This framework serves to integrate higher-level optimality theory with low-level physiological, biogeographical, and biometerological theory. Such a general theory also generates hypotheses and assists in operationalizing critical variables.”

SOCIAL ORGANIZATION

Bison have two types of social groups during the non-breeding season: bull and cow/calf groups (McHugh 1958:14–16). Membership in bull groups is usually limited to mature males (4 years old and older) with a few three-year-olds and the very occasional two-year-old bull. McHugh (1958:14) states that the instances of cows attaching themselves to bull herds are only in rare cases when a very young barren cow briefly joins the herd. On the other hand, cow/calf groups are composed of mature cows, 79 yearlings, calves, and two-year-old bulls. Sometimes three-year-old bulls and very rarely older bulls will also join a cow/calf herd for a short period (McHugh 1958).

Group structure, however, is variable. Membership in both groups changes rapidly as new members join and others leave (Lott 1991; Lott and Minta 1983; Van Vuren 1983).

Group size adapts to local conditions and seasonal fluctuations (Figure 3.1). Bull groups are generally less gregarious than cow/calf groups (Berger and Cunningham 1994;

Hansen 1984; McHugh 1958; Melton et al. 1989; Van Vuren 1983). During the non- breeding season, bull herds range in size from one to a few dozen members. At the same time of year, cow groups range in size from two to several hundred animals.

Figure 3.1. Idealized annual bison herd size (data from Berger and Cunningham 1993).

During the mating season (the rut) the size of cow/calf groups increases dramatically with the incorporation of bulls into the groups. This social fusion forms a third group 80

type: mixed groups. McHugh (1958:14) notes that this change does not drastically alter the nature of the herd because these groups are still dominated by cows and calves and their nature is essentially matriarchal. Even during these periods there are still a few old isolated bulls wandering across the landscape. The overall age and sex ratio for modern bison populations varies throughout the year and across space. Estimates for overall annual sex and age structure suggest a range of between 40–55% of all bison are immature (less than 4 years old), 21–24% are adult male, and 20–39% are adult female

(Fuller 1960; Meaghers 1973).

Males leave the mixed groups at the end of the rut and reform into bull groups. Bull groups and cow/calf groups often intermix during the non-breeding season for a few hours or less, but these period of mixing, are temporary and have little influence on the composition of either herd (McHugh 1958:15–16).

Membership in a particular herd, either bull or cow/calf, is not based on clan (i.e., familial) affiliation. Instead, Van Vuren (1983:331) concludes that “bison social structure was very fluid: group composition changed frequently, usually whenever another group was encountered.” However, the degree to which group membership changes may be different between bull and cow/calf groups. McHugh (1958:16) notes that “bull groups

(with the exception of some pairs) showed a low degree of cohesiveness, and cow groups a high degree.” This suggests that members of bull groups are more likely to change group affiliation than members of cow/calf groups. However, frequent membership changes do occur in cow/calf groups. 81

Lott (1991:142) concludes that group stability was typically low for bison because the

Great Plains was a large uniform environment. Movement in all directions should provide a suitable habitat for an individual. Furthermore, strong year-to-year fluctuations in climatic conditions of the Great Plains make the amount and quality of forage highly variable over time. In good times, when there is adequate forage everywhere, herd membership is expected to be fluid and populations will be distributed across the landscape. In times of drought or other poor climatic conditions (e.g., severe cold, heavy snow) group composition will likely be stable and populations may cluster in areas of available forage and water.

REPRODUCTION

In this section I discuss the major stages of reproduction and specifically discuss the factors that influence the synchronization of births and behavior of bison.

Rut

The first reproductive stage is the rut. The rut is a period of frenzied activity for mature bulls. Historic and modern records show that the rut begins gradually in middle of

June, reaches a peak in July and August, and slowly ends in the later half of September

(Lott 1981; McHugh 1958, 1972; Roe 1951). There is, however, some variation in the

timing of the rut throughout the Plains. In Northern Plains bison the rut is delayed several 82

weeks, until the middle of August, about 2 to 3 weeks later than the rut in the Southern

Plains.

Bull behavior is noticeably different during this period. The normally reclusive

mature bulls join cow/calf herds for the purpose of mating. They take part in loud

bellowing, wallowing, and vicious battles with other bulls to establish dominance within the group, which determines which males will have preferential access to females for mating (Lott 2002; Roe 1951; McHugh 1958). Battles are often fierce and only occasionally deadly (Burgett 1990).

Although bulls are sexually mature by three or four years old, most very young bulls do not mate (Berger and Cunningham 1994; Pucket et al. 2003). Successful mating does not begin until about year five and high rates of successful mating generally only occur after about year 8 (Wilson et al. 2002). The long delay between sexual maturity and successful mating results from the fact that winning battles with other bulls is clearly important if a male wishes to mate.

After establishing their dominance, males generally search for estrus females (Berger and Cunningham 1994; Lott 1981; McHugh 1958). Females with the appropriate signs of being in estrus—i.e., low levels of progesterone and swollen vulvas—form a tending bond with a dominant male (Green and Berger 1990; Komer et. al. 1993; Lott 1974;

McHugh 1958). These tending bonds appear to last only a day or two (Berger 1989).

Behavior associated with these bonds consists of bulls staying closely alongside their selected females and rejecting any interest shown to that female from other males by bellowing, threatening, and charging (Berger and Cunningham 1994; McHugh 1958). 83

During this tending period mating occurs (see descriptions in Lott 1981, 2002). At

this point mated males break the tending bonds and move away, often searching for other

mates. Females, on the other hand, remain monogamous. Between 80 and 95% of the

mated females copulate only once a year; if they copulate more than once they usually

have only one partner (Lott 1981). After a female has mated, she generally does not elicit

interest from other males.

Gestation

After mating, female bison go through a fairly long gestation period that varies from

approximately 264 to 285 days (Berger 1992; Berger and Cunningham 1994; Haugen

1974; McHugh 1958; Pucek et al. 2003). The length of gestation appears to be correlated

with (1) the physical condition of the mother, (2) the previous reproductive status of the

mother, and (3) the timing of mating.

Physical conditioning of the mother and timing of mating are key factors in

explaining much of the variation in gestation length, because both appear to affect the ability of mothers to shorten gestation in order to increase birth synchronization with other members of the herd (Berger 1992; Berger and Cunningham 1994:121–125). Late- breeding females in good physical condition are able to shorten gestation by a couple of weeks. No similar adjustment occurs among poorly conditioned females (Berger and

Cunningham 1994:123). Of course, the adjustments in gestation length are relatively slight even for animals in good condition. Only females mating within two weeks of the 84 peak mating period can shorten their gestation. After this time no adjustments are possible.

As with mating timing, former reproductive status also influences gestation length.

Primiparous females (first-time mothers) exceed multiparous females in mean gestation length. First-time mothers have a gestation period that lasts on average 8 days longer than that of females which have already given birth (Berger 1992; Berger and Cunningham

1994).

Parturition

As gestation draws to a close, pregnant cows separate themselves to find a protected area in which to give birth (McHugh 1958, 1972; Berger and Cunningham 1994). The degree of separation from the herd seems to vary with herd size, the nature of the surrounding topography, and presence/absence of predators (Lott 2002). Berger and

Cunningham (1994:114) suggest that for the National Bison Range herd in Montana and the Badlands herd in South Dakota, when herd size is large and protective cover is not available, cows generally give birth in small herds of pregnant females. On the other hand, among bison living on Santa Catalina Island in California, where herds are small, topography is rugged, and predators are not present, isolated births are more common

(Lott 2002).

Timing of Parturition and Birth Synchronization

The timing and sequence of the birth period for bison is fairly predictable (U.S. Fish and Wildlife Service 1977). The first birth of the year usually occurs sometime between mid-April and mid-May. After a slow beginning, there is a relatively short period of a few weeks in which the majority of births occur and then the birth rates declines dramatically. Occasional births, however, have been recorded into the late fall and winter

(Haugen 1974; McHugh 1958, 1972; Roe 1951; Rutberg 1984; Young 1973).

The beginning of the birth season varies depending on which part of the Plains is considered (Berger and Cunningham 1994:118–119; Reynold, Glaholt, and Hamley

1982) (Table 3.2). For bison in northern latitudes or at high elevations, parturition begins slightly later than in southern latitudes or lower elevation areas. The earliest births at sites in the Northern Plains, such as Yellowstone in Wyoming, National Bison Range in

Montana, and Wood Buffalo Park in Canada, occur in early May (Carbyn and Trottier

1987, 1988; Meagers 1973). Bison at Wind Cave and the Badlands National Park, on the

Central Plains, have their first births during the first week of April (Berger and

Cunningham 1994). Halloron (1968) reports that bison in Oklahoma start to give birth between March 10 and April 7.

Despite the variation in date of onset of the birth period, there is a strong pattern for clustering in the majority of births during the birth season. Figure 3.2 presents the distribution of birth scheduling for modern bison herds from across the plains. These data show that although cohort births occur over a period of 97 to 220 days (mean of 86

Table 3.2. Distributions for the Timing of Annual Birth for Modern Bison Herds Study First 50% 75% 95% Herd, Location 25% Birth 80% Date Last Birth Citation Year(s) Birth Birth Birth Date Antelope Island State Wolf and Kimball 1989: 1987 1-Apr 15-Apr 22-Apr 13-May 29-May 24-Jun 30-Oct Park, UT Table 1 Badland National Park, Berger and Cunningham 1989 6-Apr 17-Apr 2-May 30-May 31-May 20-Aug SD 1994: Table 6.1 Badland National Park, Berger and Cunningham 1986 8-Apr 24-Apr 8-May 21-May 5-Jun 15-Sep SD 1994: Table 6.1 Badland National Park, Berger and Cunningham 1985 7-Apr 20-Apr 2-May 27-May 6-Jun 1-Nov SD 1994: Table 6.1 Badland National Park, Berger and Cunningham 1987 8-Apr 24-Apr 5-May 5-Jun 8-Jun 15-Oct SD 1994: Table 6.1 Badland National Park, Berger and Cunningham 1988 3-Apr 22-Apr 10-May 3-Jun 11-Jun 15-Oct SD 1994: Table 6.1 Burwash Crown Game 1958 3-Jan Young 1973 Preserve, Ontario Elk Island National end of Chapman and Feldhamer July Park, Alberta April 1982 Konza Prairie Research 1991- April May Towne 1999: 362 Natural Area, Kansas 1997 National Bison Range, 1982 20-Apr 29-Apr 4-May 8-May 11-May 31-May 26-Jul Rutberg 1984: Figure 1 MT Slave River Lowlands, 1971- Calef and Van Camp 1-Apr NWT 1978 1987 Egerton 1962 cited in Waterton Lakes April July Chapman and Feldhamer National Park, Alberta 1982 Wichita Mountains 1959- 30-Mar Halloran 1968 Wildlife Refuge, OK 1968 Wind Cave National Shult 1972 cited in 1969 21-Apr 1-May 10-May 2-Jun 10-Jun 15-Sep 1-Oct Park Haugen 1974: Figure 2 Wind Cave National Green and Rothstein 1984 1-Apr 3-May 3-May 10-May 10-May 28-Jun 7-Nov Park, SD 1993: Figure 1 Wind Cave National Green and Rothstein 1982 1-Apr 12-Apr 26-Apr 10-May 17-May 17-Aug 20-Sep Park, SD 1993: Figure 1 Wind Cave National Green and Rothstein 1983 1-Apr 19-Apr 26-Apr 24-May 7-Jun 20-Sep 27-Sep Park, SD 1993: Figure 1 Wood Buffalo National 1932- 10-May early June Soper 1941: 378 Park,NWT and Alberta 1934 Wood Buffalo National 1980 12-May Carbyn and Trottier 1987 Park,NWT and Alberta Yellowstone National 1902- end of 15-Apr October McHugh 1958: 30 Park-Lamar Valley, WY 1925 May Yellowstone National 1937- 15-Apr early June September Meagher 1973: 75 Park-Lamar Valley, WY 1950 Yellowstone National 1991- Berger and Cain 1999: Park-southern Greater 29-Apr 28-Jun 1992 Figure 3 Yellowstone Ecosystem

Figure 3.2. Distribution of bison birth frequency for modern bison herds (Sources: Antelope Island: Wolf and Kimball 1989: Table 1; Badland National Park (1988–1987): Berger and Cunningham 1994: Table 6.1; National Bison Range: Rutberg 1984: Figure 1; Wind Cave National Park: Green and Rothstein 1993: Figure 1; Yellowstone National Park (1902–1925): McHugh 1958: 30; Yellowstone National Park (1902–1925): Meagher 1973: 75).

175 days) after the start of birthing season, the majority of births (75 to 80%) occur within an average of 55 days (range = 22 to 95 days) from the season’s beginning.

Therefore, the extreme duration of the birthing season is really a function of how long it takes for the last few births of the year to occur. Stated another way, in a large sample of bison cohorts, there is a high probability that most were born within about a month and a half of each other.

Factors influencing Birth Synchronization

While geographic and other variation has been observed, all researchers report a strong pattern of birth synchrony in bison. It is clear that the synchronized births provide numerous adaptive advantages to the mothers and their offspring. These include reduced predation on neonates, decreased inter-litter competition, increased vigilance against threats, and increased fecundity for next year. While these and other factors are benefits of birth synchrony, they do not appear to control the variation in birth dates within and between populations (Berger 1992; Berger and Cunningham 1994:124). Instead, researchers generally focus on either resource availability or predation pressure as being the most important selective forces driving birth synchrony (Green and Rothstein 1993;

Rutberg 1984). The first explanation suggests that births are closely spaced to minimize environmental or energetic stress on the mother and offspring, while the latter proposes reduced loss of the vulnerable newborns either by satiating or confusing predators. Each factor is discussed below. 89

Resource availability is an expression of the intractable interrelationship between plant productivity, climatic conditions, and population density (Berger and Cunningham

1994:119). These factors have a direct influence on the nutrition of mother and offspring, which in turn affects body condition. As Berger and Cunningham (1994:119) state,

“animals in superior physical condition conceive earlier than those in poorer shape, and it is likely that the uniformity in better nutrition...help[s] to promote higher degree of birth synchrony...” This seems to explain why the level of synchrony varies from year to year within the same population living in the same area (Green and Rothestein 1993:923).

Annual variations in the availability of nutrients will cause variation in the level of birth synchrony.

This hypothesis also explains why parturition begins at different times and lasts for varying lengths in northern and southern populations. Northern bison give birth later than bison, in part to assure that birth occurs after the arrival of spring vegetation. Without this vegetation, mothers could not convert grass into milk and thus their calves would not survive (Guthrie 1980:69). In addition, by concentrating their births in the spring, young calves have enough time to develop before the onset of harsh winter conditions (Berger

1992). Bison born late in the season will likely not be able to survive the winter.

The alternative view is that predation has a strong influence on birth synchrony patterns. However, Green and Rothstein (1993:923) question the value of this explanation for birth synchronization. They note, for example, that bison herds exposed to intensive predation pressure by wolves (e.g., Wood Bison National Park) do not have a more synchronized birth pattern than bison facing lower levels of predation (e.g., Catalina 90

Island, Yellowstone National Park, Wind Cave). There is no question that wolves prey upon calves; however, this evolutionary force does not appear to influence calving patterns as much as resource availability does.

MOBILITY

Mobility patterns are one of the most intensively investigated aspects of bison behavior (Krasinska et al. 2000; Roe 1951; Seton 1910). All grazing animals deplete the available forage in the area in which they live and therefore must move when forage becomes scarce in their local area. The key questions that most interest researchers concerning bison mobility are: How often do they move? How far do they move? What factors influence this mobility? As discussed below, the answers to these questions largely depend on whether one is using ethnohistoric or modern ecological information.

Ethnohistoric Accounts of Bison Mobility

Early historic accounts (e.g., Allen 1876; Arthur 1975; Roe 1951; Seton 1910) argued that bison seasonally migrated hundreds or even thousands of miles. Allen (1876:59–61), for example, reports, “I have...been assured by former agents of the American Fur

Company that before the great overland emigrations to California...divided the buffalo into two bands, the buffaloes that were found in summer on the plains of Saskatchewan and the Red River of the North spent winter in Texas.” 91

Later historians, though discounting these very long migrations, nonetheless steadfastly argued that bison seasonally moved hundreds of miles. Seton (1910:266) states, “I conclude with Hornaday that the buffalo did migrate 300 to 400 miles northward in the spring, and as far southward again in the autumn.” Statements such as these are representative of most early ethnohistoric records (see Bamforth 1987;

Hornaday 1889; Roe 1951; Soper 1941).

Recently, researchers have questioned the veracity of these early ethnohistoric accounts. Roe (1951:78) concluded that the idea of long bison migrations is based on analogies to the cattle drives of the late 19th century in which domestic cattle were annually moved vast distances. As Roe (1951:77) states, “a careful study...reveals how flimsy is much of the evidence upon which various forms of the migration argument are based.”

Several researchers (Bamforth 1987; Hanson 1984; Roe 1951) have outlined problems with these early models of long-distance bison mobility. First, there is no definitive evidence that an individual bison could even walk the entire 1400 miles from

Texas to southern Canada in a single year. The only proof ever presented were references to albino bison in both far northern and southern locations of the plains, under the mistaken assumption that these represented the same animals. Second, none of these models explains the impetus for such a long-distance migration. It has been assumed that annual climatic changes forced bison to migrate from one end of the plains to the other, but this assumption is at odds with present knowledge of bison herds, which live year- round without supplemented diets in small confined national parks and private ranches 92 throughout northern areas of the United States and Canada. Third, the distances proposed for these continent-long migrations actually underestimate the actual mileages bison would have had to traverse. Due to geographic impediments such as rivers, mountains, and lakes, individual animals would in reality have to walk much farther than the straight-line distance assumed. In conclusion, long-distance migration models are simply unrealistic. More recently, the focus of research has switched from the distance bison may move to the ecological factors which cause bison mobility.

Ecological Studies of Bison Mobility

Based on the definitions outlined by Baker (1978), bison are neither nomadic (i.e., having no apparent fixed pattern of direction) nor long-distance movers (Lott 1991).

There are several reasons why bison move, including the search for food and water, search for other bison, and escape from danger or uncomfortable conditions (e.g., cold temperatures, snowfall, insects, and predators) (Hanson 1984; Krasinska et al. 2000; Lott

1991; Morgan 1980; Roe 1951; Senft et al. 1987). Random events, such as inclement weather or attacks by predators (including humans) also cause bison to take flight. But these moves usually occur over short distances, most less than 10 km (Bamforth 1988).

In addition to being tied to a relatively small area of land, bison also appear to show a surprising spatial fidelity. Krasinka et al’s. (2000) multi-year tracking study of European bison, for example, saw a high proportion of bison returning to the same summer and winter ranges over multiple consecutive years. Similarly, a number of accounts suggest 93

that Plains and wood bison repeatedly move back and forth between the same summer

and winter forager areas (Lott 1991; Meagher 1973; Soper 1941).

Baker (1978) identified this type of movement as “seasonal latitudinal return

migrations" (see Lott 1991). That is, modern bison living on refuges or nature reserves appear to have distinctive seasonal home ranges (e.g., warm-season and winter home ranges) (Krasinksa et al. 2000; Lott 1991). At Yellowstone National Park, McHugh

(1958:13) observes that with the advent of snow in the late fall, the Lamar Herd bison move approximately 25 miles from the meadows at 8000–9000 ft to the lower Lamar

Valley at 6000 ft. In May or June, with the disappearance of snow, the bison return to the higher elevations. Soper (1941:380) observed a similar pattern for the herds at Wood

Bison National Park, Canada. In the spring, bison move from the plains of the Western

Slave River to the uplands of the Alberta Plateau. As the weather deteriorates in the fall, bison abandon the uplands and returns to the plains. The bison at Henry Mountain, Utah also move into their winter range in September and return to their summer range

(approximately 15 km away) in May and June (Nelson 1965:24–29). Occasionally, these seasonal migrations result in bison moving across ecotones. Reynolds, Hansen, and

Peden (1978:587) suggest that bison in the Slave River Bison Refuge in Canada move from their summer range in the prairie to their winter range in wooded habitats.

Meagher (1973) and others (Fortin et al. 2003; Krasinkska et al. 2000; Reynolds,

Hansen, and Peden 1978) have attributed these moves to environmental heterogeneity.

Seasonal changes in snowfall, temperature, forage quantity, and resource structure create strong contrasts between these different seasonal ranges. 94

Telemetry studies and other direct observations of free-range bison indicate that these

seasonal moves are generally short (as little as 5 to 15 miles) and that bison restrict their

movements to within relatively circumscribed areas (Krasinka et al. 2000; Widga 2003,

2006a, 2006b). Table 3.3 presents the annual home range areas for bison from various

herds in North America and Europe. While there is significant variation in the data, it seems that an accurate low-end estimate for annual range for bison is between 50 and 125 km 2 a year (foraging radius of 4 to 6 km), and high-end estimates range between 400 and

800 km 2 (foraging radius of 11 to 16 km). These estimates, of course, are developed from

managed herds and underestimate the actual ranges of wild herds not restricted to game

preserves. Unfortunately, these types of data are all that is available and researchers can

either choose to learn from them or disregard them entirely. However, as recently

summarized by Widga (2006b:Table 4.12), the estimates from modern managed herds

fall within the expected range for an animal with the body size of a bison. Moreover,

recent isotopic analyses (e.g., strontium, stable carbon) of archaeological bison remains

suggest that prehistoric bison, like their modern counterparts, probably did not range over

vast territories, but instead were restricted to relatively small seasonal and annual patches

(Widga 2003, 2006a, 2006b).

Several factors appear to influence the variation within home range, including

composition of forage patches, herd size, degree of snow cover, and participation in

mating activities (Fortin et al. 2003; Hanson 1984; Meagher 1973). When forage is

abundant, bison herds move infrequently, and when they do move it is only for short 95

distances; however, when forage quality is poor, bison move frequently and cover a

greater distance (Fortin et al. 2003).

Table 3.3. Annual Home Range and estimated Foraging Radius for Modern Bison Populations Herd Annual Home Foraging Herd Environment Citation Type Range (km 2) Radius (km) a Limited range herds Catalina Island, California cow/calf grassland 56.1 4.2 Lott 1991 mountain Henry Mountain, Utah cow/calf 52.2 4.1 Van Vuren 1983 forest Free-range herds Badland National Park, Berger and cow/calf grassland 111.2 (est.) 5.9 South Dakota Cunningham 1994 Bialowieza Primeval young Forest, Poland and forest 44 3.7 Pucek et al. 2003 bulls Belarussia Bialowieza Primeval prime Forest, Poland and forest 84.3 5.2 Pucek et al. 2003 bulls Belarussia Bialowieza Primeval prime Forest, Poland and forest 100 5.6 Pucek et al. 2003 cows Belarussia Bialowieza Primeval isolated Forest, Poland and forest 136.5 - 151.6 6.5 - 6.9 Pucek et al. 2003 bulls Belarussia Mackenzie Bison Larter and Gates cow/calf forest 897 16.9 Sanctuary, Canada 1990 Mackenzie Bison Larter and Gates bull forest 433 11.7 Sanctuary, Canada 1990 Larter and Gates Mink Lake, Canada cow/calf grassland 335 10.3 1990 Yellowstone National Park cow/calf parkland 93 (est.) 5.4 McHugh 1958 (Hayden Herd), Wyoming a estimate calculated based on an assumed circular annual range

Herd size also affects this pattern, and therefore bull herds and cow/calf herds have

different mobility patterns (Van Vuren 1983; c.f., Krasinska et al. 2000). The home

ranges of bull herds are significantly smaller than the home ranges of cow/calf herds. 96

This difference results from differential forage needs of the two types of herds. Female-

dominated herds are usually much larger than bull herds. Consequently, a patch of a

particular size provides more available per-capita forage to a small bull group than it does

for a large cow/calf herd. Thus, the cow/calf groups have to visit more or larger patches

to find adequate forage, resulting in larger home ranges than males (Larter and Gates

1990, 1994).

Several studies (Berger and Cunningham 1994; Daleszcyk 2004; Krasinska et al.

2000; Larter and Gates 1994; Van Vuren 1983) showed that reproductive status

influences mobility patterns. Males of different ages have different home ranges

irrespective of forage availability during the rut. Mature males who win most of their

battles during the rut have smaller home ranges than younger males who are

reproductively less successful. These younger bulls have to move more than older bulls in

searching for cow groups with a low percentage of males. Age, per se, is not as important

as breeding success, because for all ages mated bulls move much less than those who

were unsuccessful. Female mobility, on the other hand, is not similarly affected by the

mating season. The mobility of females does not change dramatically during the rut

because they do not have to search for mates.

BISON PHYSIOLOGY

This section examines changes in the physiological condition of bison, paying special attention to variation in response to such factors as age, season, sex, reproductive status, 97 diet, and health (e.g., Brink 1997; Speth 1983). Several indexes have been used to measure the physical condition of animals (DelGuidice et al. 1994), but most biologists rely on the quantity and quality of skeletal fat as a measure of physical condition of wild animals (Anderson et al. 1972; Ballard and Whitman 1987; Cederlund et al. 1989; Davis et al. 1987; Depperschmidt et al. 1987; Marquez and Coblenz 1987; Ranson 1965: Torbit et al. 1988). Studying skeletal fat is important because in times of stress, when the animal cannot acquire adequate forage, it survives by metabolizing stored fat (see Mitchell 1962;

Speth 1983).

Dietary and Seasonal Variation

The timing and nature of seasonal stress to animals are generally fairly predictable. In temperate and northern latitudes, all animals face the greatest amount of stress during the late winter and early spring (Speth 1990:151). For bison these seasons are difficult because forage quality decreases, especially the crude protein and phosphorus levels in grasses. Protein is critical for sustaining body weight and microbial actions in the rumen.

Speth (1983) reported that the normal minimum maintenance level of crude protein for adult bison is 5% of their total diet. For pregnant and lactating cows this requirement is even higher, about 10%. Several researchers (Everitt et al. 1982; Meyer and Brown 1985;

Peden et al. 1974; Speth 1983) have shown that on the Southern Plains, at least, the crude protein levels in grass drop below this minimum level in late October/early November, 98

and remain below this level until mid-March. This means that bison must operate at or below maintenance levels for most of the winter and spring (see Speth 1983:121).

A similar problem is faced in maintaining phosphorus intake. Phosphorus helps maintain normal microbial activity in rumen and adequate milk production in lactating cows. The minimum maintenance requirements for bison have not been established, but domestic cattle require phosphorous to make up 0.15% of their total diet (Speth

1983:130). Unless supplemental sources can be found—for example, by chewing bones

(Johnson 1985; Guthrie 1984)—bison cannot naturally obtain an adequate quantity of phosphorus from mid-August to mid-March.

Other periods of the year are also stressful to bison. These periods are marked by either inadequate feeding or great energy expenditures. Reproductive activities represent one of the stressful periods. Males are often in poor condition following the rut, while females are most stressed when they are pregnant or lactating. Physical condition is often correlated with age; for example, very old and very young animals are usually much leaner than prime age adults. Other periods of stress, such as those caused by disease or parasitic infestations, may be much less predictable.

As stated previously, bison rely on their stores of fat to survive when they are stressed or unable to acquire adequate forage. This mobilization of fat occurs in a set sequence through the skeleton. Speth (1983) outlined that for modern bison fat mobilization begins in late winter from the fat deposits along the hump, around internal organs, and in subcutaneous fat. By early spring, these fat levels reach extremely low levels, as low as 1 to 2% of their normal levels. In fact, it is not uncommon, especially during very stressful 99

periods, for these fat deposits to completely disappear (Speth and Spielmann 1983). At

the same time, fat in the bone marrow may be greatly reduced. However, some fat usually

remains in the marrow cavity even after it is completely metabolized from other parts of

the body. Mobilization in the limb bones also occurs in sequence, starting in the proximal limbs and moving down the leg. Even in severely stressed animals, some fat usually remains in the marrow cavity of the mandible, distal limbs (especially phalanges and metapodials), and around the brain (Burger et al. 2005; Stiner 1990, 1994). There is also an important deposit of lipids that persists in the cancellous tissue of the vertebrae even late into the spring. While fat does remain in severely stressed animals it is little more than water and connective tissue (Vickery 1977). Also, the composition of fat changes.

Often all that is left in animals in poor physical condition are the low quality saturated fatty acids (Wood 1984).

The sequence of fat metabolization in the different body parts seems to be related to the proportion of oleic acid and unsaturated fatty acids with low melting points. This is

why the large deposits of subcutaneous fat, such as along the hump, are metabolized

before deposits of intermuscular, intramuscular, and within bone fat deposit. As Wood

(1984) shows for Bos taurus , there is a progressive decrease in the quantity of oleic acid as one moves from the peripheral fat deposits (i.e., subcutaneous fat) through intramuscular and intermuscular deposits to the deep body sites, such as fat of the abdomen and within bones.

For the appendicular skeleton, the proportion of low-melting-point fatty acids (such as oleic acid) decreases as one moves up the limb, reaching its lowest point in the 100

proximal femur and humerus (Binford 1978; Speth 1983:103). The sharpest decline

occurs in the proximal radius and tibia. Therefore, fat remains in lower limbs long after it

has been metabolized in the rest of the skeleton.

The distribution of unsaturated fatty acids, according to Byers and Schelling (1988)

and Young (1976), is inversely correlated with the internal body temperature at the

location of the fat deposit. Saturated fats are most commonly found near the core of the

animal, while low temperature, unsaturated fats are dominant in areas away from the

body core. Brink (1997:266) proposes “this is because unsaturated fatty acids, with lower

melting points, are required in the limb bones located farthest from the core of the body

in order to prevent them from becoming stiff in cold weather.”

The proportion of oleic acid and unsaturated fatty acids with low melting points also correlates with the flavor of meat. Several studies (Rhodes 1971; Westerling and Hedrick

1979) demonstrate that there is a positive correlation between proportion of oleic acid

and increased flavor, although Dryden and Marchello (1970) report low or variable

correlations between fatty acid composition and eating quality.

As the period of stress passes, the condition of the bison improves and they start to

replace their fat reserves. This improvement is related to increased forage quality,

especially forage with a greater percentage of protein and phosphorus. The replacement

sequence of fat is the reverse of the sequence of metabolism. Fat is replaced first in the

marrow of lower limbs, followed soon after by the upper limbs, and culminates with the

replacement of subcutaneous fat deposits (Speth 1990:152).

101

Variation by Sex

The above discussion addresses factors (e.g., availability and quality of forage) that

affect adult bison of both sexes equally. However, based on extensive study of

domesticated animals, it is known that males, pregnant females, and non-pregnant

females vary in their relative condition at different times of the year (Figure 3.3).

Figure 3.3. Idealized variation in bison physiology for bison cows and bulls.

Bulls. Adult males are in their best overall condition during the late spring or early summer, and in their poorest condition during the rut. Speth (1983:105; Speth and

Spielmann 1983:3) outlines the annual changes in the physical condition of male bison.

Nutritional shortages during the winter result in a low point in the physical condition of 102 bulls. After this low point, males begin to replenish their fat reserves, which results in rapid improvement in their physical condition during the late spring/early summer. Peak condition is reached just before the onset of the rut. However, the huge energy expenditures coupled with minimal forage intake during the rut results in a precipitous decline in the physical condition of males. This decline is so dramatic that males have been known to lose up to 15% of their pre-rut body weight.

There are, however, factors that cause variation in the effect of the rut. The most important is that not all bulls participate in the rut, and those which do participate may not do so for the same amount of time (McHugh 1958:30). This means that mature, but young, bulls will be in quite good shape at the end of the rut.

Young bulls are not the only members who do not fully participate in the rut. Older bulls, generally older than 13 years of age, often choose not to participate in the rut

(McHugh 1958:30). Members of both these age cohorts, because of limited involvement in the rut, do not face many of its detrimental physiological effects.

Following the rut, in late fall/early winter, males return to their normal non-mating activity patterns and their physical condition improves slightly. However, despite this improvement, males are simply unable to make up for the losses caused by the rut and generally enter the winter in comparatively poor condition.

Cows. The annual variation in the physical condition of pregnant females follows a comparable cycle, though offsets relative to males (see Figure 3.3) (Speth 1983:105–

106). Pregnant females are in their worst condition during the calving season and in their best condition in the fall and early winter. The rut is not so taxing on females because 103 they are able to graze instead of searching for mates, so they enter the winter in superior condition to males; consequently they are in better condition to males at the beginning of spring. However, later in the spring the condition of late-term pregnant cows and especially newly postpartum mothers deteriorates precipitously. The physical condition of cows without young will generally follow the cycle of bulls outlined in the previous section.

The nutritional burden of reproduction in bison, as with all mammals, falls almost entirely on females and relates not only to intrauterine nutritional demand but also consequent changes in the mother’s behavior and metabolism (Mitchell 1962:326). For most of the gestation period, pregnancy has relatively little detrimental effect on the mother. In fact, several controlled feeding studies of domesticated cattle (Slonaker 1925;

Wang 1935) suggest that the physical condition of pregnant females actually improves during pregnancy. Pregnant females have greater body weight—more than can be accounted for by the growing fetus and other physical changes associated with pregnancy—and larger deposits of fat, calcium, and other nutrients than do nonpregnant females eating identical diets. Hansard and Berry (1969:40) and others (Eckels 1916;

Mitchell 1962; Snapp and Bull 1944) attribute these changes to increased food intake, nutrient absorption, and decreased activity. These improvements in the mother’s condition likely prepare her for the greatly increased nutritional demands of lactation and may even be essential for the survival of the offspring (Mitchell 1962: 539).

Only toward the end of gestation does the condition of bison begin to decline. Part of this is due to the poor forage quality available to the mother during the early spring, 104

although much of this decline can be credited to a voluntary drop in food intake on the

part of the pregnant mother. Blaxter (1957) suggests that during the later stages of

gestation the uterus of most ruminants takes up so much space in the abdominal cavity

that it exerts pressure on the alimentary tract and consequently reduces appetite.

Following calving, the condition of females with young does not improve; in fact, it declines further. One reason for this rapid decline is that cows with young require three to four times more food than barren cows. Postpartum cows simply cannot eat enough forage to maintain their physical condition. Green and Rothstein (1991:524) report that females can lose up to 40% of their barren year weight in parturient years. The first few months of nursing, when the calf is most reliant on its mother, is particularly costly for the mother (Daleszczyk 2004; Green 1986; Moen 1973). Blaxter (1957) calculates that the caloric needs of only 1 or 2 days of lactation is equivalent to the entire needs of the fetus. Even if the mother’s diet is highly nutritious she continues to lose weight while nursing. It is common for bison cows to lactate for up to four months while subsisting on a sub-maintenance diet (Daleszczyk 2004; Speth 1983:130).

In the summer and early fall, as the demands of the calf on the mother decrease and diet improves, her condition improves gradually. By late fall or early winter females reach their peak condition.

105

Ontogenic Variation

The previous discussion has focused only on adult animals. However, this focus

misses some of the important changes in physiology of bison as they develop from young calves to fully mature adult animals. One important aspect in which adult and young animals differ is in the quantity of skeletal fat. All fetal and neonatal mammals are characterized by having very low fat deposits (Nobel 1984; Robbins 1983). This condition changes as the animal gets older (Wood 1984). As an animal ages and dietary energy is diverted away from skeletal and muscle growth and more into fat growth, two things occur. First more fat is deposited within the animal’s skeleton, and second, the composition of that fat changes. Wood (1984:410) outlines the processes of fat accumulation in young mammals: “As the animal gets older...[fat] cell size increases and those of connective tissue and water decrease... In subcutaneous fat the number of fat cells also increase during growth... This is mainly attributable to recent filling of cells which have been present from an early age.”

Concurrent with this build-up of fat is a change in the composition of fatty acid. The little fat that young animals do have is mostly composed of water and connective tissue.

The fatty acids found in very young animals are commonly low quality saturated fats rather than high quality fats such as oleic acid (Hansard and Berry 1969). As the animal matures this changes, and larger deposits of high quality fats are found throughout the body.

106

MODELING BISON PROCUREMENT AND PROCESSING

Based on the information presented above, I have developed a set of predictions concerning the actions of Paleoindian hunters behaving optimally with respect to efficiency. These predictions have implications for the composition of Paleoindian zooarchaeological records related to (1) kill size and age and sex composition, (2) season of mortality and number of kill events, (3) carcass butchering strategies and intensity, (4) geographic distribution of prey, and (5) predation pressure. Three of these prediction are highlighted below.

Kill Composition and Size

• Since for most of the year cow/calf herds are larger than bull herds, they therefore

represent higher density of food value, which makes them preferential targets for

mass kills. Therefore the sexual composition of most mass kill sites, irrespective

of season, should exhibit a high frequency of cows and calves.

• For kills of single or small number of animals, the larger size of bulls makes them

more attractive targets than cows or calves. Therefore, assemblages representing

hunting of individuals should be dominated by an abundance of males.

• Because the size of bison herds changes seasonally, the potential size of mass kills

should reflect the gross changes in the natural fluctuations in bison herd size.

Therefore, the largest possible mass kill during late winter/early spring is 107

probably relatively small because few bison are expected to herd together during

this period. In contrast, very large mass kills are possible during the late summer

and early fall, because this is a time when multiple bison herds aggregate

together, and therefore potentially many animals can be taken at one time. As

herd populations are expected to gradually decrease through the late fall and early

winter, the maximum size of kills during this period are expected to be smaller

than during the summer/early fall but still larger than during late winter/early

spring.

Season of Mortality

• With a large sample of teeth, analysts can be confident that the majority of bison

in an annual cohort were born during the birth pulse. Although there will be a

small proportion of out-of-sync births, animals born late in the year are not

expected to survive the winter.

• When assigning seasonality estimates to assemblages, researchers need to

consider the south-north cline in the timing of the birth pulse.

• Mass kills, especially of cow/calf herds, can occur during any time of year.

Therefore, all seasons can be represented. However, late summer through winter

will be the preferred time for mass kills. The spring and rut (July and August)

seasons are expected to be poorly represented in seasonality data. 108

• Because bull condition is superior to cows during the spring and early summer,

site assemblages dominated by males (i.e., predation of bull herds) will

preferentially date to these periods.

Distribution of Prey

• Bison can be expected to use similar summer and winter range. Therefore, the

spatial distribution of bison herds should be predictable to hunters with

knowledge of forage conditions, snowfall, and thermal conditions.

• The spatial distribution of bull and cow/calf herds may not overlap for most of the

year. Cow/calf herds will be distributed over larger forage patches than bull herds.

• Annual foraging ranges for bison are probably too small for a hunter-gatherer

group to rely exclusively on a single herd. Hunters must prey on multiple herds so

as not to deplete local population levels. 109

CHAPTER 4: BISON DIMINUTION: EVALUATING THE ROLE OF HUMAN

HUNTING VERSUS CLIMATE CHANGE FROM OSTEOMETRIC, AGE

STRUCTURE, AND STABLE CARBON ISOTOPES

Extinctions have fascinated the general public and scientific community for over 200 years (Grayson 1984). Scientific interest in the North American Late Pleistocene extinctions has largely been directed toward understanding the cause(s) behind the disappearance of 35 genera of large-bodied land mammals from the Rancholabrean fauna

(Grayson and Meltzer 2002). Among the competing hypotheses, climate-driven environmental change (Graham 1986; Lundelius 1988; Guthrie 1995, 2003) and human overhunting (Martin 1984, 2005; Owens-Smith 1988; Alroy 1999; Haynes 2002a, 2000b;

Surovell et al. 2005) are the most common explanations for the disappearance of the large game.

As recently noted by Surovell et al. (2005:6231), directly testing these hypotheses is difficult because we do not fully understand the relative timing of (1) the arrival of humans on the North American continent, (2) the disappearance of the Pleistocene fauna, and (3) the onset of periods of major climatic change that may be associated with the extinction events (e.g., Younger Dryas). To overcome these difficulties, Surovell et al.

(2005) propose using simplified predictions of how the overkill and climatic-change hypotheses should produce detectable patterns in the spatiotemporal patterning in archaeological sites showing overlap between hominid predators and extinct prey. A similar approach can be used with other datasets to clarify the nature and manner in 110 which overhunting and climate change affected the large-bodied fauna of North America.

In this case, I consider how alternative extinction hypotheses affect the physiology and population structure of the large mammal genus, Bison , that survived the end of the

Pleistocene.

This chapter draws on data concerning body size and mortality age profiles from

North American bison and published results from soil stable carbon isotope studies as proxy measures for understanding the relative effects of climate-driven environmental change and human hunting pressure on large terrestrial mammals of North America.

Body size and mortality age profiles are the primary focus of this study because they play key roles in many aspects of animal ecology and behavior and respond to shifts in hunting pressure and environmental conditions (Edwards 1967; Peters 1983; Guthrie

1984; McNab 1990; Brown and Nicoletto 1991; Stiner 1994, 2005; Blackburn et al.

1999; Ashton et al. 2000). Body size in particular, as Allen et al. (2006:631) suggest, is one of the most ecologically integrative attributes of a species and strongly related to a number of ecological attributes, including home range size and metabolic rates. Body size has been described as a "taxon-free" attribute because consistent patterns in the distribution of body mass are seen across ecological communities and species.

Of the large game that survived the end of the Pleistocene, bison is the most appropriate taxon to use in this study for several reasons. First, there is abundant archaeological and paleontological evidence for Late Quaternary bison from across the

Great Plains of North America (McDonald 1981), including evidence from deposits that predate the initial human colonization of North America (circa 14,500 calendar years BP 111

[calBP]). Second, human exploitation of bison as an important food resource predates extinction of other megafauna species (Baker 1998; Bement and Carter 2003; Ferring

2001; Hemming 1970; Hester 1972), and high levels of bison hunting continued (and probably increased) through the Holocene (Fawcett 1987; Frison 2004). Finally, Late

Quaternary bison systematics suggests a continuous, polyphyletic population rather than one interrupted by the introduction of new lineages or species (Wilson 1975; Guthrie

1980; Shapiro et al. 2004; cf., McDonald 1981).

Prior studies have already demonstrated a high degree of geographic and temporal plasticity in bison (Hillerud 1970; Hughes 1978; Wilson 1978; McDonald 1981; Hofman and Todd 2001; Lyman 2004b). Late Pleistocene bison were up to 20% larger than modern bison and underwent a dramatic size reduction in a relatively short period of time at the end of the Pleistocene. By 4000 calBP diminution had continued to the point where bison body size overlapped with the modern form. These earlier studies produced many valuable insights into bison paleoecology, although they usually (1) are based on data from a small number of assemblages, (2) are limited in temporal focus to just the last

12,000 years, and (3) emphasize data from the northern Great Plains to the exclusion of other regions. Furthermore, many studies limited their focus to demonstrating that size reduction took place and provided little in the way of explanation concerning the specific mechanism(s) responsible for the change in body size (cf., Wilson 1980; McDonald

1981; Guthrie 1984; Lyman 2004b).

This study expands the temporal and geographic database on bison body size by utilizing data from assemblages spanning the last 37,000 years, and compares the trends 112

in body size change between northern and southern bison populations. I evaluate the

factors driving this size reduction by empirically comparing the potential effects human

hunting and climate-driven changes in forage composition and quality had on body size.

EXPLAINING BODY SIZE CHANGE

Before examining changes in bison body size throughout the Quaternary, I review

how human hunting and environmental change, according to paleoecologists and wildlife

biologists, affect body size in Late Quaternary terrestrial mammals. The following

sections briefly describe each mechanism and outline testable hypotheses that will be

used later in this chapter.

Human Prey Selection as a Catalyst for Body Size Diminution

A number of researchers have noted a link between intense hunting pressure and a

decrease in body size of key prey species (Edwards 1967; Stiner et al. 2000; Coltman et

al. 2003; Mannino and Thomas 2002; Stiner 2005; Wolverton 2001). This decline in prey

body size is argued to result from a demographic shift in the prey population after

preferential removal of older members of a prey population. Because the older members

of a population are also larger members, this produces both a decrease in the average age of that group, and a corresponding decrease in mean body size. These effects are most pronounced for certain taxa such as fish and reptile, which continue to grow throughout 113

adulthood; although a similar pattern is argued to occur in mammals (Edwards 1967;

Wolverton 2001).

Because a change in the mean age of a prey population is the underlying cause for the decline in body size in these models, independent measures of the population age structure are needed. Hunting simulation studies and analyses of modern non-human and human death assemblages (Stiner 1994, 2005; Wolverton 2001) demonstrate that intense predation pressure produces mortality age profiles in which juvenile individuals are proportionally more abundant than in populations under no harvest pressure. As Beaver

(2007) explains, this occurs for two reasons. First, as increased predation results in fewer animals surviving from one age class to the next, the cumulative effect of this continual removal of individuals from the population is that older members have suffered greater membership decline because they have been subjected to more years of predation.

Second, by lowering overall population level, increased predation reduces intraspecific

competition, which reduces juvenile mortality more than it does adult mortality.

The Holocene survival of smaller, modern bison continues to be used as a line of

evidence supporting the detrimental impact of human groups on Late Pleistocene

megafauna (e.g., McDonald 1981; Martin and Steadman 1999; Martin 2005). While

many researchers have de-emphasized predator pressure as a driving force in the

diminution of Holocene bison (e.g., Guthrie 1980; Hughes 1978; Wilson 1975), the

impact of predation on body size trends has rarely been empirically addressed. I explicitly

address the impact of human predation on Late Quaternary bison populations by 114 examining the potential effects of different predation strategies on herd composition and body size of bison populations.

The relationship between human hunters and their prey is often very complex. With respect to large, herding ungulates such as bison, hunter-gatherers may acquire animals through the stalking and dispatch of single animals or the capture of small groups facilitated by the use of constructed traps or topographic features (e.g., arroyo traps, jumps, pounds) (Frison 2004). Hunters may be economic specialists, utilizing bison to the exclusion of other species, or they may use bison whenever they are available and an attractive food resource given immediate conditions. The North American archaeological record presents us with all of these scenarios as possible strategies for bison procurement.

Bison populations may have responded differently to different human predation strategies. Figure 4.1 depicts four possible outcomes in terms of bison population age structure and body size changes associated with differing frequencies of catastrophic

(mass kills) and selective hunting. It should be noted that no single hunting strategy characterizes hunter-gatherer groups throughout prehistory. Rather, the careful consideration of all of these models allows us to understand the range of effects, if any, human predation may have had on bison body size and demographic profiles.

Scenario #1: Regularly Occurring, Catastrophic Kills. Multi-animal bison kills are relatively common on the Great Plains of North America and have been documented for all known cultural periods (Frison 2004). Many researchers have suggested that multi- animal kills were among the most common strategies employed to acquire bison (e.g.,

Frison 1991; Reher 1977; Hofman and Todd 2001). If this was the case, one could expect 115

bison populations to be under constant predation for a long period of time. A long-term

culling of a broad range of age and sex classes from bison populations should have an

effect on the social dynamics of the larger metapopulation. A key factor here is the

duration of the hunting pressure. Regardless of how intensive the hunting is, unless it is

sustained over a long time it is unlikely to result in a strong age-structure effect (Beaver

2007). In contrast, long-term predation, which has already resulted in a significant

decrease in overall prey population levels, would be associated with decreased juvenile recruitment into prime and old-age age classes.

Figure 4.1. Expected responses of bison populations to different hunting scenarios.

A juvenile bias in the overall population would decrease the average size of bison on

the landscape. However, constant culling would also release nutritional limitations on 116

bison growth rates (i.e., few animals on the landscape = more available forage), resulting

in larger potential body size for adult animals. Conservationists commonly see these

density-dependent relationships in modern ungulate populations exposed to both human

and non-human predation (e.g., Messier 1994; Toigo et al. 2006; Wolverton et al. 2007).

Scenario #2: Sporadic Catastrophic Kills . Some researchers believe that multi-animal kill events were in fact only occasionally used by prehistoric hunters (e.g., Kornfeld

1996; Meltzer and Smith 1986). Instead, they argue that the perceived high frequency of large bison kill sites reflects a discovery bias for contexts where the exposed bones of large animals are more obvious than the materials in functionally different sites.

Within this scenario, multi-animal kill events, while part of the hunting repertoire, would have only occurred a few times in each generation of hunters. As such, their impact on prey populations would have been much less than under the regime outlined in scenario #1. As Beaver (2007) notes low intensity or short duration hunting results in limited—and even then only temporarily—changes to population structure and body size.

It is difficult to say whether these changes would be large enough to be visible within the time/space averaged fossil record.

Scenario #3: Regular Selective Hunting. While most of the samples used in this study are from multi-animal kill assemblages, this does not mean that selective hunting did not have an impact on the population dynamics of ungulates, particularly if scenario #2 applies. Most optimality treatments of human hunting behavior operate under the assumption that hunters encounter single prey animals at a particular rate (Winterhalder and Smith 1981; Winterhalder 2001). Although this does not always characterize how 117

humans utilized bison, in appropriate contexts, as discussed in chapter 7, bison were

taken singly and/or sequentially (Frison 1991, 2004).

This strategy, especially if maintained at high levels for a long period of time, would

alter bison populations differently than multi-animal hunting. In selective hunting it is

likely that the hunters would preferentially select adult animals, probably in their prime,

and possibly bulls more frequently than cows (Stiner 2004). Under long-term, heavy

selective predation, bison populations would over time show a demographic profile with

a strong juvenile bias (Beaver 2007). Such a long-term bias for juveniles would likely be associated with a smaller average body size; however, the concurrent reduction of density-dependent limitations on animal growth would increase the maximum potential adult body size.

Scenario #4: Sporadic Selective Hunting. If human groups underutilized bison resources for cultural or economic reasons, then the level of predation would be sustainable and no human-induced changes in body size or demographic characteristics would be expected.

Summary. As noted above, the response of animal populations to hunting pressure can be multi-facetted and complex (Beaver 2007). A number of factors influence the nature of the relationship, including (1) duration and intensity of hunting pressure, (2) degree of age and sex bias in prey selection, and (3) the overall size and demographics of the prey population. As discussed above, infrequent hunting or very low levels of continuous hunting are unlikely to produce archaeologically viable demographic or physiological shifts in prey populations. 118

Long-term, intensive predation is expected to show some degree of juvenile bias.

Unselective predation will result in a juvenile bias only after population levels have been significantly reduced. Prior to the overall decline in prey population levels, unselective hunting will be associated with little or no juvenile bias to populations. In contrast, intensive selective hunting will result in a greater degree of juvenile bias.

Finally, according to most models, unless intensive hunting pressure is maintained for a very long period and there is a strong juvenile bias, there will probably not be much of a change in prey body size. Even then intensive hunting pressure is expected to only affect mean body size for the group, not the maximum potential size attainable by an individual, which in fact should either remain the same or increase (Stiner et al. 2000).

While these expectations reflect our current understanding of the human impacts on landscape processes, there may be exceptions to these general patterns. Edwards (1967) suggests that extended periods of intense hunting pressure could eventually select for a smaller optimum body size, and there is some support for this hypothesis in the long-term study of guppies ( Poecilia reticulata ) (Reznick and Bryga 1987; Reznick et al. 1997).

These authors suggest that selective predation of large-bodied guppies results in genetic changes to descendant populations, who mature at an earlier age and have smaller body sizes compared to populations that did not undergo selective predation. In addition,

Coltman et al.'s (2003) genetic work on the effects of trophy hunting on male bighorn sheep population lends at least some support for this hypothesis in large game animals.

However, these empirical exceptions to our expectations remain limited to short-term changes in body size trends of a local population. Neither example reflects the sweeping 119 impact of climatic and cultural changes that occurred on the North American landscape during the Pleistocene-Holocene transition.

Environmental Factors

There is little agreement on the mechanisms underlying differences in body size as a result of environmental factors. Much of the controversy in the literature results from the fact that multiple processes operating at different scales affect body size. Some of the non-mutually-exclusive factors that have been proposed to explain the observed body size patterns include (1) heat conservation and/or dissipation, (2) habitat productivity and food quality and abundance, (3) differential ability to survive seasonal fasting, (4) predator avoidance or other behavioral adaptation (e.g., migration), (5) competition over food and/or mates, (6) life history characteristics (e.g., growth rates, reproductive history), (7) phylogenetic history, or (8) some combination of these factors (Dayan et al.

1991; Ashton et al. 2000; Allen et al. 2006; Rodríguez et al. 2006).

All these hypotheses probably adequately explain elements of observed patterning in body size; however, their explanatory power varies depending on the spatial and temporal scales at which they operate (Allen et al. 2006). For example, small- and large-bodied animals appear to respond differently in terms of the direction and degree of body size change as well as the underlying factors controlling those changes (Blackburn et al.

1999). In addition, climatic, energetic, and phylogenic factors generally operate at broad spatial and temporal scales, while competition and community interaction factors interact 120

within the shorter term and within intermediate and smaller geographic scales (Allen et

al. 2006). To simplify the discussion, the rest of this chapter will focus primarily on body

size changes in large-bodied land mammals.

Probably the best known (and likely the most controversial) environmental

explanation for body size change in large mammals is Bergmann's Rule, named for

nineteenth-century naturalist Carl Bergmann. Following its original definition (Bergmann

1847), this ecogeographic rule suggests that there is a tendency for larger body size in

homeothermic species in cooler climates than species from warmer climates. Mayr's

(1956) redefinition of this rule limited the rule to applying only to size clines within

single species, rather than communities of species.

The original explanation for this rule was that larger individuals had lower surface-to-

volume ratios than small animals. As a result, the larger animals had higher heat-

producing capacity (correlated with volume) and lower rates of heat loss (correlated to

surface area). Therefore, larger animals conserve heat more efficiently in colder climates

than do small animals (Mayr 1956). However, recent research with various birds and

mammals shows mixed results in terms of the degree to which species and multi-taxa

communities follow the predicted relationship between climate and body size, and

highlighting researchers' inability to identify a unifying explanation for these patterns

(Dayan et al. 1991; Ashton et al. 2000; Blackburn and Hawkins 2004; Meiri et al. 2004;

Blackenhorn et al. 2006; Rodríguez et al. 2006).

Meiri et al. (2004) even question whether prior studies, which generally conclude that the majority (~50 to 65%) of mammal species conform to Bergmann's rule, accurately 121 reflect the true tendency of mammal species to follow the rule. They (Meiri et al. 2004) and others (e.g., Lomolino et al. 2006) note that rare species often do not get studied and that certain species (bears, wolves, and other predators) located in mid-latitudes receive more attention than do other taxa, especially those living in the tropics. Moreover, because of the bias for studies with positive results to be submitted and published, the perceived percentage of taxa that comply with Bergmann's rule may be overestimated.

Moreover, even in situations where a taxon conforms to a Bergmannian size gradient, there is significant difficulty in evaluating validity of the thermal explanation for several reasons. First, there is a high degree of covariation among the potential environmental variables affecting body size (Yom-Tov and Yom-Tov 2005; Rodríguez et al. 2006).

Thus, it is difficult to distinguish the influences of temperature versus resource availability on the body size of North American fauna because temperature influences both body size and levels of plant productivity (Blackburn and Hawkins 2004:723).

Second, the traditional model of a linear relationship between body size and climate is probably overly simplistic. Rodríguez et al. (2006), for example, show that mean annual temperature correlates strongly with body size in northern portions of Europe, but in southern Europe, seasonality in plant productivity is a better predictor of body size. The conclusion drawn here is that if relationship between temperature and body size exists at all, it is certainly nonlinear and is influenced by other environmental variables, such as aridity, plant productivity, and seasonality (Dayan et al. 1991; Bugalho et al. 2001;

Blackburn and Hawkins 2004).

As a result of this complexity in factors influencing body size, some suggest that at 122 best, Bergman's rule should be treated as a generalized empirical finding or pattern (with many exceptions) and not as an explanation (Mayr 1956; Dayan et al. 1991; Blackburn et al. 1999; Lomolino et al. 2006).

Other Environmental Factors Affecting Body Size

If climatic factors, such as temperature, are not the sole influence on body size, or perhaps even a major influence, then other factors must play a role. A number of researchers emphasize the strong correlation between body size and the amount and duration of access to high-quality forage that a young animal has following weaning (e.g.,

Klein 1964; Purdue 1989; Guthrie 1984; Festa-Bianshet et al. 2004). Here we need to distinguish between nutrition-based and genetic limits on somatic growth. Maximum body sizes for all taxa have genetic limits and are likely fixed by evolutionary processes.

Adequate nutrition allows individuals to attain maximum size within the limits of a genetic heritage; however, individuals subjected to inadequate nutrition will not reach their potential maximum body size. However, over time nutrition-based reductions can be fixed by natural selection.

Dietary studies of domesticated ruminants suggest availability of calories alone does not result in body growth, however (Speth 1983; Byers and Schelling 1988; Owens-

Smith 1988). Limits to the ability of cells to process nutrients on a daily basis means that increasing nutrition above the level necessary for bone and muscle development creates added skeletal fat, not continued growth. Instead, maximizing somatic growth requires a 123 young animal to have access to the appropriate mixture of calories, protein

(approximately 5% or greater), and essential minerals (e.g., approximately 0.15% phosphorus) for an extended period of time between weaning and sexual maturity (Speth

1983; Guthrie 1984; Geist 1987).

A number of researchers believe the effects of limited access to high-quality forage are seen in populations of large-bodied mammals isolated on islands or mountain tops, which often undergo a considerable size reduction compared to their mainland or lowland relatives (Lundelius 1990; Purdue 1993; Kirby and MacFadden 2005; Lomolino 2005; c.f. Meiri et al. 2006). In these contexts, decreased body size of these newly insular populations is explained as a reduction of somatic growth due to reduced quality diets overall or limited seasonal access to high-quality foods, a situation that is often intensified under higher population densities faced by insular populations.

Guthrie's (1984, 1995, 2003) seminal work on body size diminution among

Pleistocene megafauna also names resource limitations and intra- and interspecific competition as major factors underlying the changes in body size. Guthrie suggests that the cooler and drier conditions and lower atmospheric CO 2 concentrations during the

Pleistocene resulted in greater access to more nutritious low-growth forage for grazing animals, for a greater portion of the year, than during the Holocene. These factors, combined with increased over-winter die-off among large-bodied grazing animals, allowed a high bioenergetic surplus beyond what was needed for basic metabolic maintenance, allowing hypersomatic growth. As conditions became warmer and wetter and atmospheric CO 2 concentrations rose during the Holocene, prior limits on plant 124

growth were lifted. As a result, forage became less nutritious. The amelioration of

climatic conditions during the Holocene also resulted in higher proportions of animals

surviving the winter, resulting in increased intraspecific competition for food. The net

effect is that Holocene animal populations had less energy to direct toward growth. Thus,

body size decreased compared to earlier Pleistocene populations.

Geist (1971, 1989) also recognizes a strong correlation between body size changes

and resource availability, especially protein content of the diet. In resource/protein rich

circumstances, animals with "dispersal" phenotypes, characterized by rapid growth and

realization of maximum body size, are expected to be more common. In contrast,

"maintenance" phenotypes, which include reduction in maximum body size, occur when

resource availability is limited and there is greater competition for local resources.

This is not to say that dietary factors alone explain all changes in body size. A key

component of many models of body size change is the role competition plays as an

evolutionary force (Guthrie 1984, 1995). Dayan et al. (1991), for example, have shown

that morphological size relationships exist among competing sympatric taxa. Changes in

body size or other features can accompany a species release, that is, the removal of

selective forces imposed by the presence of another species (i.e., when a competitor

disappears from the environment). For example, isolating large-bodied animal

populations on islands releases these populations from a need to maintain a large body

size as protection from predators, and favors reproduction at an earlier age (Lundelius

1990; Raia et al. 2003). Together, these factors promote dwarfism. Competition can also be a result of interspecific population changes. Increases in the population of a taxa can 125 cause higher levels of intraspecific competition, which in turn would select for smaller individuals (Wilson 1978).

Expected Effects of Environmental Changes

Environmentally driven shifts in carrying capacity and intraspecific competition affect birth and morality rates, and will be reflected in changes in maximum body size as well as in population age structures (e.g., Wolverton 2001). In terms of body size changes, climate-induced changes in available forage will reduce both the median body size of a population and the maximum size attainable by any individual. Moreover, because there are often geographic differences in the expression of climatic change, one would expect patterns in body size to vary by latitude.

Exposure to long-term environmental stress is also expected to result in significant changes in mortality patterns, with higher mortality among old adults and juveniles. As a result, there should be overabundance of prime age animals as compared to a stable population, as they are most able to withstand severe stress (Stiner 1994).

MATERIALS AND METHODS

Evaluating the differential effects of hunting pressure and environmental change requires reliable estimates of body size and population age structure. Osteometric measurements from adult calcanei ( os calcis ) and humeri are used here as a proxy for 126 body size (Todd 1987; Hill 1996), while mortality age profiles are constructed using known sequences in the eruption and wear of bison teeth from a series of kill and processing localities (Todd et al. 1996).

Limb Element Measurements as Proxy for Body Size

Fossil records of true body size are rarely preserved, and there has been much debate about which variable most accurately represents "body size" (e.g., Damuth and

MacFadden 1990; Meiri et al. 2006). Inferences about body size must be drawn from skeletal features that are well preserved in the archaeological and paleontological record, and which accurately reflect body size for the taxon under consideration (Todd 1987). I use linear measurements of greatest length from 849 calcanei and distal articular breadth from 1011 humeri as a proxy for body size, following previous studies of bison body size change and sexual dimorphism (Duffield 1973; Speth 1983; Todd 1987; Morlan 1991;

Hill 1996). These data come from measurements taken by myself, unpublished data collected by Matthew G. Hill and Chris Widga, or published sources (Hillerud 1970;

Speth 1983; Todd 1987; McKee 1988; Todd et al. 1992; Kornfeld et al. 1999; Borreson

2001; Hofman and Todd 2001). These calcanei and humeri come from 49 archaeological components and 10 paleontological localities dating between approximately 37,000 and

250 calBP (Figure 4.2; Table 4.1). For comparison, identical measurements were taken on 27 calcanei and 35 humeri from modern ranch-raised bison. 127

Figure 4.2. Map of the continental United States and southern Canada with archaeological sites and paleontological localities mentioned in the text and tables. Site numbers: 1 Heron-Eden, 2 Rustad, 3 Itasca, 4 Interstate Park, 5 Mill Iron, 6 Big Goose Creek, 7 Mavrakis-Bentzen-Roberts, 8 Rourke, 9 Carter-Kerr/McGee, 10 Hawken, 11 Vore, 12 Horner II, 13 Bugas-Holding, 14 River Bend, 15 Glenrock, 16 Casper, 17 Finely, 18 Wardell, 19 Scoggin, 20 Agate Basin, 21 Lodgepole Creek, 22 Hudson-Meng, 23 Scottsbluff, 24 Clary Ranch, 25 Milburn, 26 Spring Creek, 27 Red Smoke, 28 Logan Creek, 29 Simonsen, 30 Lindenmeier, 31 Jurgens, 32 Frazier, 33 Lamb Springs, 34 Frasca, 35 Olsen-Chubbuck, 36 Burntwood Creek, 37, 12 Mile Creek, 38 Scott County Pueblo, 39 Cooper, 40 Certain, 41 Lipscomb, 42 Rex Rodgers, 43 Twilla, 44 Plainview, 45 Lubbock Lake, 46 Bonfire Shelter, 47 Folsom, 48 Blackwater Draw, 49 Milnesand, 50 Garnsey, 51 Rancho La Brea, 52 Big Bone Lick. 128

Table 4.1. Summary Information for Sites Used in this Analysis Site Site Location a 14 C yrs BP Citation Function 10,504, average 12 Mile Creek S kill Hill 2002b of three dates Agate Basin-Agate Basin (Area 2) N camp/kill 10,430 Hill 2001 10445, average Agate Basin-Folsom (Area 2& 3) N kill Hill 2001 of two dates Agate Basin-Hell Gap (Area 3) N camp/kill 10,445 Hill 2001 Big Bone Lick-Locality 1, Zone B S kill? 530 Widga 2006a Big Goose Creek N kill 530 Frison 1991 Blackwater Draw-Brown Sand S kill? 11,220 estimate Haynes 1995 Wedge (Strat D2/C) Blackwater Draw-Gray Sand S kill 12,250 estimate Haynes 1995 (Strat B1) processing 10,081, average Dibble and Bonfire Shelter-Bone bed 2 S locale of four dates Lorrain 1967 2632, average of Dibble and Bonfire Shelter-Bone bed 3 S kill two dates Lorrain 1967 Bugas-Holding N camp 500 Rapson 1990 Burntwood Creek S kill? 9085 Hill et al. 1992 kill/processing Carter-Kerr/McGee-Cody N 9000 estimate Frison 1984 locale 10,016, average Casper N kill Frison 1991 of two dates 1400, average of Certain S kill Buehler 1997 three dates processing Clary Ranch N 9040 Hill 2001 locale Cooper S kill 10500, estimate Bement 1999 9009, average of Finley N camp/kill Frison 1991 two dates 10,891, average Folsom S kill Meltzer 2006 of seven dates Frasca N kill 8910 Frison 1991 kill/processing 9600, average of Wormington Frazier N locale two dates 1988 375, average of Garnsey S kill Speth 1983 17 dates Glenrock N kill 280 Frison 1991 6326, average of Hawken N kill Frison 1991 two dates Heron-Eden N kill 9037, average of Corbeil 1995 129

Table 4.1. Summary Information for Sites Used in this Analysis Site Site Location a 14 C yrs BP Citation Function three dates Horner II N kill 10060 Frison 1991 natural 9418, average of Hudson-Meng N Agenbroad 1978 accumulation three dates natural 6639, average of Interstate Park N this study accumulation four dates kill/processing 7167, average of Itasca N Widga 2006b locale three dates kill/processing Jones-Miller N 10020 Stanford 1984 locale Jurgens N camp/kill 9070 Wheat 1979 14310 to 15300, natural Marcus and La Brea-Pit 13 S based on four accumulation Berger 1984 dates 19300 to 21400, natural Marcus and La Brea-Pit 3 Lower S based on four accumulation Berger 1984 dates 12650 to 1443 0, natural Marcus and La Brea-Pit 3 Upper S based on eight accumulation Berger 1984 dates 11130 to 12200, natural Marcus and La Brea-Pit 61/67 S based on five accumulation Berger 1984 dates 28200 to 33100, natural Marcus and La Brea-Pit 77 S based on three accumulation Berger 1984 dates natural Marcus and La Brea-Pit 81 S 10,940 accumulation Berger 1984 25100 to 33000, natural Marcus and La Brea-Pit 91 S based on six accumulation Berger 1984 dates Lamb Spring S kill 9000 estimate McCartney 1990 10,569, average Lindenmeier N camp/kill Haynes 1995 of three dates Lipscomb S kill 10,820 Hofman 1995 Lodgepole Creek N kill 8700 Frison 1991 processing Logan Creek-Zone B N 6340 Widga 2006b locale processing Lubbock Lake-FA5-6 S 400 estimate Johnson 1987 locale Lubbock Lake-TTM Station E, kill/processing Johnson and S 5650 estimate Strat 3 (Strat 3/4A) locale Holliday 1986 130

Table 4.1. Summary Information for Sites Used in this Analysis Site Site Location a 14 C yrs BP Citation Function 2506, average of Mavrakis-Bentzen-Roberts N kill Frison 1991 two dates natural Milburn N 8300 Hillerud 1970 accumulation 11,076, average Mill Iron N camp/kill Frison 1996 of nine dates Milnesand S kill 10000, estimate Hill 2002a 9395, average of Holliday et al. Olsen-Chubbuck S kill eight dates 1999 Holliday et al. Plainview S kill 10000, estimate 1999 8810, average of Red Smoke N camp Knudson 2002 seven dates Rex Rodgers S kill 9391 Spear 1978 River Bend N camp 250 estimate McKee 1988 Niven and Hill Rourke N kill 2100 estimate 1998 7437, average of Rustad N camp Widga 2006b five dates Scoggin N kill 4540 Frison 1991 Scott County Pueblo S camp 250 this study 9905, average of Scottsbluff N kill this study two dates kill/processing 6956, average of Simonsen N this study locale three dates Spring Creek camp 6145 Widga 2006b Twilla S kill 1120 Buehler 1997 10392, average Kornfeld et al. Upper Twin Mountain N kill of two dates 1999 Vore N kill 370 Frison 1991 1226, average of Wardell N kill Frison 1991 three dates aN = North; S = South

Wildlife biology studies of bison have shown that length and breadth measurements

for various body parts are highly correlated with body size (e.g., Berger and Cunningham 131

1994; Krasinska and Kasinski 2002). Scott's (1990:327) extensive study of living African ungulates shows that while all postcranial skeletal measures are significantly correlated with body mass, length and breadth measures for proximal limb elements are especially good predictors of body mass.

Most prior studies have not evaluated the reliability of using the length of the calcaneus as a proxy for body size. The length of the calcaneus is at least in part a reflection of the mechanical advantage of the lever arm formed by the distal hind limb

(Christiansen 2002). Figure 4.3 shows that for the 57 ungulate species from around the world examined by Christiansen (2002:Table 1), the log transform of length of the calcaneal tuber, a measure strongly correlated with greatest length of calcaneus, correlates significantly (r 2 = .82) with the log transform of body mass. Calcaneal tuber measurements were taken from specimens curated in the Zoological Museum in

Copenhagen, and body mass was determined either by direct measurement of the living animals or taken from published sources.

The calcanei used in the current study had complete epiphyseal union for the calcaneal tuber, which in modern Bison bonasus occurs by early in the fifth year of life

(Koch 1935; Empel and Roskosz 1963). All humeri used in this study had fused distal epiphyses, which occurs early in the fourth year (Koch 1935; Empel and Roskosz 1963).

Bement (1999) suggests the possibility that Early Holocene bison fusion rates occurred at slightly different times than modern bison. He suggests that the calcaneal tuber fused during the sixth year and distal end of the humerus fused during the second year.

Unfortunately, the accuracy of Bement's estimates have yet to be evaluated in modern 132

Bison bison of known age. In any case, bison reach sexual maturity by their fourth year of life, and growth ceases (females) or slows significantly (males) (Berger and Peacock

1988; Berger and Cunningham 1994; Bement 1999; Krasinska and Krasinski 2002).

Figure 4.3. Relationship between body mass and calcaneal tuber length for 57 ungulate species from across the globe. Data based on Christiansen (2002:Table 1).

The sex was assigned to specimens using bivariate scatter plots of multiple measurements. This technique has been shown to successfully separate bulls from cows in known sex samples (Todd 1986, 1987; Morlan 1991; Hill 1996). Separate studies by

Walde (2004) and Lewis et al. (2005) of large (n > 30) assemblages of modern bison limb bones from known sex individuals indicate that a sexual misidentification rate of between 133

2 and 4% can be expected. While our procedure may assign a small number of specimens to the wrong sex, any mistake should be relatively infrequent and should not significantly influence the outcome of this study.

Two procedures are used to depict body size changes in bison through time. First, a

"ratio diagram" is used to display the overall distribution of the measurements used in this study through time regardless of sex (e.g., Simpson 1941; Lyman 2004b). In this diagram, the x-axis represents the difference between the log 10 of the mean measurement for a comparative sample and the log 10 of that measurement for each prehistoric specimen, while the y-axis represents the age of the deposit from which the specimen originates. By setting the comparative standard on the x-axis to 0, a negative ratio value indicates a prehistoric specimen smaller than the comparative standard, and a positive value indicates a specimen larger than the modern standard. The mean measures of samples from the Glenrock site are used as the comparative sample, as they represent a large sample of very Late Holocene (approximately 300 years BP) bison, which predate historic bison ranching activities.

The other procedure depicts mean body size of bison by using the mean values for calcaneal and humeral measurements along with their associated 95% confidence limits

(CI) (Sokal and Rohlf 1995) calculated for presumed bulls and cows from each locality.

Statistically, assemblages with non-overlapping 95% CI plots have significantly different mean values.

The age of each deposit represents a calibrated calendar age estimate based on dating either the bison remains themselves or charcoal collected from the associated geologic 134

stratum. When multiple overlapping radiocarbon ages are available for a deposit, an

average was calculated using the Calib 5.01 software (Stuiver and Reimer 1993). All

radiocarbon ages were calibrated using IntCal04 curve with the online CalPal software

(Jöris and Weninger 1998).

Mortality Patterns as Measure of Predation Pressure and Climatic Stress

The dental ages for animals in all assemblages were determined using the same

recording system, developed from the analysis of a large assemblage of known age bison

raised on commercial bison farms in Wyoming (Frison and Reher 1970; Niven et al.

2004). Each profile was constructed based on the frequencies of individuals (tooth-based

MNI) in 1 of 12 dental age groups (i.e., 1 to 12+). The youngest cohort (Age Class 1)

consists of newborn specimens, while the oldest age cohort (Age Class 12+) represents

animals near the maximum potential life expectancy for bison. As Todd et al. (1996:147)

note, dental age groups, especially in young animals ( ≤ year 5) equate with corresponding chronologic age based on comparisons with eruption and wear schedules for modern bison of known age. For older specimens, the cumulative effects of diet-related variation in molar attrition rates means that the older dental age groups do not precisely reflect annual intervals. The discord between dental age and chronological age most strongly affects the older cohorts (> Age Class 9).

To simplify presentation of these data, the frequencies of each of the 12 age groups were combined into one of three age categories (juvenile, prime adults, and old adults) 135 following the system outlined by Stiner (1990, 1994, 2005). The juvenile category includes the three youngest dental age cohorts (Group 1–3), and corresponds to the period in the animal's life between birth and when the third molar (M 3) begins to show occlusal wear (year 3). The prime adult cohort begins at the appearance of wear on the M 3 and continues until half the height of the crown has been worn away (e.g., Group 9) (based on crown height data from Todd et al., 1996). Animals older than prime adults were assigned to the old adult category.

RESULTS

The results of this study support prior observations that Late Pleistocene bison were significantly larger than modern bison, and that diminution occurred toward the end of the Pleistocene and continued through much of the Holocene (Hughes 1978; Wilson

1978; Hofman and Todd 2001). Mortality age data indicate that the decrease in body size occurred in a context where there was no evidence for a juvenile bias through time. In fact, over time the proportional representation of older animals generally increased and younger animals decreased in kill samples. These results suggest that bison diminution is not associated with a concurrent decrease in the average age of bison populations. Thus, the changes in body size likely were a reaction to environmental conditions rather than human predation pressure.

136

Body Size: General Trends

Diminution in bison body size over the last 37,000 years represents a dramatic change in both the median body size as well as the maximum and minimum size attained by any individuals (Figures 4.4 and 4.5). The measurements depicted in these figures represent all measurable calcanei but only a sample (n=462; 46%) of the humeri. Using only a sample of the humeri was necessary because data on individual humeri were not available for all sites. Later analyses, which use assemblage mean values, will involve the total sample of measurable elements.

Both calcanei and humeri indicate that Late Pleistocene bison are very large both in terms of mean size as well as maximum size. Prior to 14,000 calBP ago, only the smallest calcanei and humeri of Late Pleistocene bison overlap in size with the largest of Late

Holocene specimens. In fact, it is not until the terminal Pleistocene (around 13,000 calBP for humeri and 11,000 calBP for calcanei) that any individual specimens fall below the size of comparative Late Holocene standard. By about 8000 calBP both datasets suggest that most specimens fall within the range of the Late Holocene standard. Diminution, however, appears to have continued until approximately 3000 years ago, when some specimens fall below the lower range for the standard.

137

Figure 4.4. Ratio diagram of dimensions of bison calcanei plotted against the mean value for the modern standard (0.0 vertical line). Points with positive values are larger than the modern bison standard; points with negative values are smaller than the Late Holocene (Glenrock) bison standard. Gray bar represents range of values for Late Holocene comparative standard.

138

Figure 4.5. Ratio diagram of dimensions of bison humeri plotted against the mean value for the modern standard (0.0 vertical line). Points with positive values are larger than the modern bison standard; points with negative values are smaller than the Late Holocene (Glenrock) bison standard. Gray bar represents range of values for Late Holocene comparative standard. 139

To better track changes in body size through time, Figures 4.6 and 4.7 depict the

smoothed trend for median calcanei and humeri measurements grouped into 1000-year

intervals. Both data sets generally show comparable trends in body size through time;

however, there are a few differences. While there was an overall trend toward diminution

through time, this trend is neither unidirectional nor continuous. Throughout most of Late

Pleistocene, bison body size appears to remain relatively stable. Only after 13,000 or

12,000 years ago, depending on which element is examined, does bison begin to show

significant decreases in body size. Although it is difficult to say precisely when this

period of diminution began, it is clear that it post-dates the earliest documented

appearance of human hunters in the American mid-continent (Meltzer 1995, 2004;

Waguespack 2007).

By the Early to Middle Holocene, bison body size appears to stabilize at a level below

that of earlier bison but is still larger than Late Holocene bison. This body-size plateau

lasts for approximately three millennia. During the mid-Holocene bison undergo another

period of decline in body size. Here the calcanei and humeri datasets depict slightly

contradictory trends in this second round of body size decline. According to the larger

calcanei dataset, around 8000 calBP bison underwent a 2000-year-long period of significant diminution. At the end of this period the median size of bison was significantly smaller than the modern standard. A reversal of the trend for diminution has

occurred within the last 2000 years, when once again bison body size approached that of

the modern standard. The lower resolution humeri datasets suggest that sometime after 140

about 6000 years ago this decline occurred, although a lack of data for much of this time period makes it difficult to track this change. Data from published sources can be used to fill in this gap, as discussed further below.

Figure 4.6. Smoothed trend line of ratio diagram of dimensions of bison calcanei plotted against the mean value for the modern standard (0.0 vertical line), with relationships to major Late Quaternary climatic periods and colonization of North America by human hunters.

141

Figure 4.7. Smoothed trend line of ratio diagram of dimensions of bison humeri plotted against the mean value for the modern standard (0.0 vertical line), with relationships to major Late Quaternary climatic periods and colonization of North America by human hunters.

Body Size: Regional Trends

No distinction has yet been made between bulls and cows or between regional populations. In this section, all measurable calcanei and humeri are used, but the sample is divided into latitudinal and sex groups and body size trends are compared for these subgroups. To track these changes, mean calcaneal and humeral measurements were 142 depicted separately for bulls and cows along with 95% confidence limits for each assemblage used in the study (Tables 4.2 and 4.3). To determine whether there are latitudinal trends in the data, sites located north and south of 40° N latitude are plotted separately. Because Rancho La Brea was the only locality dating to before about 15,000 calBP, the figures presented here depict only data from the last 15,000 years. Calcanei data are shown in Figures 4.8 and 4.9, and humeri data are shown in Figures 4.10 and

4.11.

Although bulls and cows follow the same general trend in size diminution, these figures show that there were important size differences between southern and northern populations. Prior to about 12,500 calBP southern bison populations were in general the same size or slightly larger than northern populations, although any difference is not statistically significant. These Late Pleistocene bison were so much larger than Holocene bison that at this time cows were about the size or slightly larger than Late Holocene bulls.

As already noted, at the end of the Pleistocene bison underwent a dramatic size reduction. However, it appears that southern bison underwent greater degrees of diminution than did northern bison. This diminution during the Holocene was so pronounced that between approximately 6000 and 2000 years ago, the body size of southern bulls and cows fell below the range of Late Holocene bison, reaching their smallest size at anytime during the last 37,000 years. Within the last 500 years body size increased slightly to within the range of the comparative standard. Unfortunately, humeri measurements from comparable-aged sites on the Southern Plains are not available, so 143

Table 4.2. Summary Statistics for Calcanei Greatest Length for Sites Used in this Study. Female Std Male Std Site N Min Max N Min Max Reference Mean Dev. Mean Dev. 12 Mile Creek 182.8 10.8 4 170 196 Hill 2002b Agate Basin-Agate Basin (Area 158.7 5.6 29 147 167 181.7 3.8 3 179 186 Hill 2001 2) Agate Basin-Folsom (Area 2& 156 1.4 2 155 157 180 1 Hill 2001 3) Agate Basin-Hell Gap (Area 3) 160.5 2.1 2 159 162 Hill 2001 Big Bone Lick-Locality 1, Zone 139.3 4.3 21 129 146 159.7 2.5 3 157 162 this study B Big Goose Creek 142 4.1 8 134 146 156.5 3.9 6 150 160 Hill 1996 Blackwater Draw-Brown Sand 163 1 179 1 this study Wedge (Strat D2/C) Blackwater Draw-Gray Sand 174.3 5.1 3 170 180 this study (Strat B1) Bonfire Shelter-Bone bed 2 169.5 4.9 2 166 173 this study Bonfire Shelter-Bone bed 3 138 8.3 21 120 149 154.8 2.9 5 151 158 this study Burntwood Creek 150.7 3.1 3 148 154 this study Casper 157.2 4.5 35 150 165 171.5 4.4 4 166 176 Hill 2001 Clary Ranch 154.3 6.9 9 140 163 174 5.7 2 170 178 Hill 1996 Finley 158.9 6 15 145 166 172.3 3.5 23 167 182 this study Folsom 164.8 2.4 5 162 168 182.3 5.3 7 175 189 this study Frasca 158.8 5.6 19 150 167 176.5 3.6 11 171 183 Hill 1996 Borreson Frazier 155.5 4.2 11 152 164 184 1 2001 Garnsey 143.7 2.4 4 140.5 146.7 159.8 0.5 2 159.5 160.2 Speth 1983 Glenrock 142.3 6.4 108 128 155 159.8 2.6 23 157 166 this study Hawken 149.6 5.2 18 140 155 166.4 2.4 5 164 169 this study Hudson-Meng 151.6 5.6 36 139 163 176.2 8.5 5 167 185 Hill 1996 Interstate Park 145.3 5 7 139 152 163.8 1.9 4 161 165 this study Itasca 149.6 4.7 5 143 155 163.2 3.7 8 161 170 this study Jurgens 152.6 2.5 9 148 155 163.8 7.7 5 159 177 this study La Brea-Pit 3 Upper 168.5 4.6 21 155 175 180.9 3 7 178 185 this study La Brea-Pit 3 Lower 176.5 2.6 4 174 180 this study La Brea-Pit 13 167.9 2.9 8 165 173 189 1 this study La Brea-Pit 61/67 168.7 5.9 3 162 173 186 5.7 2 182 190 this study La Brea-Pit 77 167 4.1 4 161 170 185.5 0.7 2 185 186 this study La Brea-Pit 81 161 1 this study La Brea-Pit 91 170.5 0.7 2 170 171 177.5 3.5 2 175 180 this study Lindenmeier 164 2.9 5 160 167 184.4 4 10 179 192 Hill 1996 Lipscomb 160.4 5.5 17 151 170 184 5.7 2 180 188 Hill 1996 Lodgepole Creek 155 3.6 3 152 159 168.7 3.2 3 165 171 Hill 1996 Widga et Logan Creek-Zone B 143.5 3.6 12 136.5 147.5 167 1 al. 2000 Lubbock Lake-FA5-6 141.3 4.9 3 138 147 156 1 this study 144

Table 4.2. Summary Statistics for Calcanei Greatest Length for Sites Used in this Study. Female Std Male Std Site N Min Max N Min Max Reference Mean Dev. Mean Dev. Lubbock Lake-TTM Station E, 140 0.9 6 139 141 152.2 1.6 5 151 155 this study Strat 3 (Strat 3/4A) Mavrakis-Bentzen-Roberts 146.6 5.3 37 132 154 159.3 3.8 10 156 169 this study Milburn 146.6 3.1 12 143 152 165 2.8 2 163 167 this study Mill Iron 159.5 3.5 4 156 163 Hill 2001 Milnesand 150 1 Hill 2002a Modern Bison 145.5 5.5 19 135 155 155.9 3.6 8 152 160 Hill 1996 Olsen-Chubbuck 154.3 4.1 57 144 162 168.9 4 23 165 180 this study Plainview this study Red Smoke 153.5 4.9 4 147 159 170.5 2.1 2 169 172 this study Rex Rodgers 173.7 3.2 3 170 176 this study Rourke 141.6 2.7 7 138 145 152 4.2 2 149 155 this study Rustad 150 1 163 2 3 161 165 this study Scottsbluff 156.7 3.7 12 152 163 180 1 this study Simonsen 145.5 0.6 4 145 146 159.2 4.3 16 154 166 this study

145

Table 4.3. Summary Statistics for Humeri Distal Articular Breadth for Sites Used in this Study. Female Std Male Std Site N Min Max N Min Max Reference Mean Dev. Mean Dev. 12 Mile Creek 106.5 2.4 4 105.0 110.0 Hill 2002b Agate Basin-Folsom 95.9 1.0 5 94.8 97.3 116.1 1 Hill 2001 (Area 2& 3) Hofman and Bugas-Holding 78.6 0.1 2 78.5 78.7 93.0 1.7 8 90.2 95.7 Todd 2001: Table 1 Burntwood Creek 85.8 1 97.0 8.5 2 91.0 103.0 Hill et al. 1992 Hofman and Casper 90.2 3.0 56 84.2 90.2 105.6 1.7 6 103.5 107.7 Todd 2001: Table 1 Clary Ranch 88.3 3.1 22 82.0 95.0 101.5 4.6 6 96.7 109.0 Hill 2001 Finley 89.4 4.6 38 81.6 98.4 100.8 1.9 28 97.7 104.8 this study Hofman and Folsom 95.4 3.4 9 91.5 100.9 106.8 2.9 8 103.7 113.3 Todd 2001: Table 1 Frasca 86.6 2.9 15 83.2 93.0 101.3 3.4 17 96.4 109.8 Todd 1987 Hofman and Frazier 91.2 3.4 3 87.4 95.6 103.0 0.0 2 103.0 103.0 Todd 2001: Table 1 Hofman and Glenrock 79.0 1.9 25 75.3 82.9 89.8 3.1 51 83.4 98.7 Todd 2001: Table 1 Hofman and Hawken 85.1 3.3 11 80.5 90.7 97.1 3.1 13 91.3 102.4 Todd 2001: Table 1 Horner II 86.6 4.9 23 80.0 95.0 100.6 2.0 16 97.0 103.0 Todd 1987 Hofman and Hudson-Meng 84.5 4.0 32 79.1 93.6 105.0 1 Todd 2001: Table 1 Interstate Park 79.1 1 92.6 2.9 4 89.4 96.4 this study Itasca 97.1 1 101.6 1 this study Jones-Miller 89.8 3.1 179 81.2 98.1 105.3 2.5 15 101.5 109.3 Todd 1987 Jurgens 86.8 1.8 11 83.3 90.2 98.7 5.0 3 95.4 104.4 this study La Brea-Pit 13 95.8 1 99.3 1.9 3 97.8 101.4 this study La Brea-Pit 3 Lower 94.2 1.0 2 93.5 94.9 104.3 0.6 3 103.7 104.7 this study La Brea-Pit 3 Upper 97.2 4.1 21 91.0 103.9 112.0 1.1 4 111.0 113.5 this study La Brea-Pit 61/67 97.0 1.3 7 95.5 98.8 109.0 2.5 3 107.2 118.8 this study La Brea-Pit 77 97.7 2.3 7 93.9 101.6 108.0 6.4 2 103.5 112.5 this study La Brea-Pit 91 98.2 3.5 4 95.7 103.2 this study Lamb Spring 82.4 3.5 3 78.5 85.2 100.6 5.0 10 92.1 106.9 Todd 1987 Lindenmeier 91.7 2.5 2 89.9 93.5 105.1 2.4 8 101.0 108.3 Hofman and 146

Table 4.3. Summary Statistics for Humeri Distal Articular Breadth for Sites Used in this Study. Female Std Male Std Site N Min Max N Min Max Reference Mean Dev. Mean Dev. Todd 2001: Table 1 Todd et al. Lipscomb 92.7 3.2 17 87.2 99.4 107.8 3.4 8 104.2 113.2 1992 Logan Creek-Zone B 83.9 6.5 11 76.3 96.4 this study Lubbock Lake-TTM Station E, Strat 3 (Strat 74.6 1 91.8 0.4 4 90.7 92.5 this study 3/4A) Milnesand 90.0 1.4 2 89.0 91.0 103.0 1 Hill 2002a Modern Bison 76.4 2.9 27 72.0 81.0 91.0 2.2 8 88.0 94.0 Todd 1987 Plainview 92.0 2.0 2 90.6 93.4 108.6 3.6 4 103.9 112.7 this study Rex Rodgers 86.5 1 91.5 1 this study River Bend 76.6 3.7 26 70.5 82.2 92.4 3.1 14 87.0 99.5 McKee 1988 Hofman and Scoggin 84.2 5.9 5 74.8 89.5 97.2 1 Todd 2001: Table 1 Scott County Pueblo 78.1 1.6 10 76.0 81.0 91.0 4.2 16 84.0 97.0 this study Simonsen 91.3 4.9 2 87.8 94.8 94.8 3.3 7 90.6 99.7 this study Spring Creek 82.2 4.6 10 75.7 91.0 94.9 2.6 3 91.9 96.7 this study Kornfeld et al. Upper Twin Mountain 102.5 1 1999 Hofman and Vore 79.3 2.5 62 74.4 84.6 91.9 3.0 67 86.2 99.9 Todd 2001: Table 1

147

Figure 4.8. Cross plot of calcaneal maximum length with 95% confidence interval over time for cows from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds). 148

Figure 4.9. Cross plot of calcaneal maximum length with 95% confidence interval over time for bulls from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds).

149

Figure 4.10. Cross plot of humeral distal articular breadth with 95% confidence interval over time for cows from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds).

150

Figure 4.11. Cross plot of humeral distal articular breadth with 95% confidence interval over time for bulls from Southern Plains (latitude below 40° N) (open squares) and Northern Plains (latitude above 40° N) (closed diamonds).

151 without more data it is difficult to confirm that bison body size reached a minimum during this period.

Northern Plains bison appear to have followed a slightly different trajectory of body- size reduction during the Holocene than southern populations. The paucity of data for northern sites between approximately 7300–2600 calBP makes it difficult to determine whether northern bison ever reached the minimum sizes that the southern population did.

While calcanei measurements from the 2100-year-old Rourke site do fall significantly below the range of the Late Holocene standard, cow measurements from this site fall within the range of the comparative standard. In contrast, humeri data from the 5200- year-old Scoggin site and calcanei data from 2500-year-old Mavrakis-Bentzen-Roberts site suggest northern bison at this time were at least the size of the Late Holocene standard, if not significantly larger. While it is difficult to know for certain whether northern bison underwent a significant decline in body size during the Mid- to Late-

Holocene, the best available evidence suggests that they likely did not.

It is possible that northern populations underwent short periods of reversals of the overall diminution trend between approximately 8300 and 7300 calBP. Measurements from Hawken, Itasca, and Rustad assemblages all suggest short-term increase in body size during this time. Without more data, however, the validity of this reversal is hard to establish.

152

Mortality Patterns

Published mortality profiles from 23 archaeological faunal assemblages spanning the last 13,000 years were used to approximate the age structure of bison populations

(Table 4.4). Mortality data used in this study were drawn only from archaeological deposits. This was necessary because paleontological deposits that predate the appearance of humans in the Americas generally represent time averaged accumulations formed over long periods of time, and have age and sex biases that poorly reflect living populations (e.g., Marcus and Berger 1984).

Based on tooth eruption and wear studies, site formational analyses, and dating of sites it appears that the assemblages used in this analysis result from either catastrophic kills or a series of closely spaced multi-animal kill events. In addition, assemblage formation appears to be largely unselective in terms of prey age or sex. As a result, it is believed that these assemblages were not the product of selective hunting.

Figure 4.12 presents the mortality age profiles of 23 archaeological assemblages considered in this study. Most assemblages cluster within the living-structure pattern area, albeit a number of them occur low on the old age axis. Four outliers (Bonfire

Shelter, Finley, Lamb Springs, and Mill Iron) fall within the prime-dominated pattern area. Overall this distribution generally fits with prior expectations for human hunters practicing a multi-animal ambush strategy (Stiner 1990; Frison 2004); however, when mean values for each cultural period are compared it is clear that the age distribution patterns are not consistent over time, nor do they reflect an idealized living population. 153

Table 4.4. Three-Cohort System Results for Bison Kill and Processing Localities, based on Mandibular Tooth Eruption and Wear Sequence. Percent of N Site Na Period b Season Source Juvenile Prime Old Dibble and Lorrain Bonfire Bone Bed 3 20 80 0 20 AR fall/winter 1968; this study Carter- 56 34 9 32 LP fall Reher 1974 Kerr/McGee-Cody Casper 31 58 11 74 LP summer/fall Buehler 1997 Certain 26 48 26 23 PRE/PH fall/winter Frison 1984 Clary Ranch 45 45 10 20 LP summer/fall Hill 2001 Cooper 47 53 0 47 EP summer Bement 1999 Finley 8 76 16 25 LP fall Todd and Hofman 1987 Garnsey 23 53 23 30 PRE/PH spring Speth 1983 Glenrock 22 54 24 120 PRE/PH fall Reher 1970 Hawken 25 64 11 96 AR winter Frison 1984 Heron Eden 34 58 8 38 LP fall/winter Corbeil 1995 Horner 39 43 18 44 LP fall Todd and Hofman 1987 Idealized Living 42 46 12 100 Reher 1974 structure Lamb Springs 30 70 0 23 LP winter-summer McCartney 1990 Lipscomb 41 56 3 32 EP summer/fall Todd et al. 1990 Mill Iron 14 74 11 35 EP spring/summer Todd et al. 1996 Rourke 30 57 13 23 AR spring/summer Niven and Hill 1998 Scoggins 35 59 6 34 AR multiple Niven and Hill 1998 Twilla 25 45 30 20 PRE/PH spring Hughes 1989 Vore-Level 1 31 33 36 58 PRE/PH multiple Reher and Frison 1980 Vore-Level 2 27 47 26 118 PRE/PH multiple Reher and Frison 1980 Vore-Level 3 34 51 15 168 PRE/PH multiple Reher and Frison 1980 Vore-Level 5 42 27 30 211 PRE/PH multiple Reher and Frison 1980 Wardell 21 54 25 121 PRE/PH fall Reher 1974 a N refers to tooth-based Minimum Number of Individuals b AR: Archaic (2000-9000 calBP); EP: Early Paleoindian (>12,000 calBP); LP: Late Paleoindian (9-12,000 calBP), PRE/PH: Late Prehistoric/Protohistoric (<2000 calBP)

154

Figure 4.12. Group means and observed mortality patterns in bison assemblages from Early Paleoindian, Late Paleoindian, Archaic, and Late Prehistoric/Protohistoric periods.

The age distributions presented in this figure do not support the hypothesis that bison population have been subject to trends for overhunting or intensive selective predation.

Had this been the case, later assemblages would show greater proportional representation of juveniles compared to earlier assemblages. This would result in later assemblages depicted in Figure 4.12 to be preferentially distributed toward the lower left corner of the graph, and earlier sites clustered more toward the upper right portion of the graph. In 155 addition, a box plot of the proportion of young in these assemblages shows that, while there is a high degree of variation and none of the differences are statistically significant, the median proportions of young animals in Early and Late Paleoindian assemblages are greater than the portion of young in either Archaic or Late Prehistoric/Protohistoric assemblage (Figure 4.13).

Instead, this analysis suggests a distinct pattern for the underrepresentation of juvenile animals in most assemblages compared to a living structure, and proposes that there is a trend for increasing representation of old adults through time. In fact, many Late

Prehistoric/Protohistoric period assemblages have higher proportions of old adults than expected in living populations. Overall, one conclusion is clear: the decline in Late

Quaternary bison body size is not correlated with a bias for young animals in those populations.

This paucity of young members in bison kill assemblage has previously been noted by Reher (1974). Although he suggests the preferential butchery of young animal carcasses may have caused under representation of these age groups, he acknowledges that there was no empirical evidence for the differential transport of young animals carcasses away from kill localities (Reher 1974:121–122).

In addition, because there is little difference in the mineral density of deciduous versus permanent teeth, differential preservation of remains of older animals compared to younger animals probably does not explain the biases observed in these data (Stiner

2005; see also Reher 1974). As a result, lacking any evidence to the contrary, it is believed that the archaeological age profiles accurately reflect the age structure of the 156

hunted animals. Intensive predation by wolves or other predators on bison populations may explain the apparent high mortality of juvenile animals (Carbyn and Trottier 1986).

The times when size reduction is most apparent (i.e., Early Paleoindian: 13,000–

11,000 calBP and Archaic: between 8000/6000–2000 calBP) are associated with mortality age distributions most biased toward prime adult individuals. While the 157

proportion of juveniles during these periods are slightly lower than in later times, there is

also a significant decrease in the frequency of old adults. In fact, in Archaic and Early

Paleoindian assemblages it is rare to see animals older than age class 11. In contrast, it is

not uncommon to have very old individuals (>age class 14) represented in Late

Prehistoric and Protohistoric assemblages, a period when body size may actually have

increased slightly.

This pattern fits with expectations of increased environmental stress in the Archaic

and Early Paleoindian periods, which results in fewer total age classes represented,

especially the older classes (Wolverton 2001). In fact, the bias for prime-age adults may

partly explain why in northern herds during the Early Archaic period there appears to be

a slight reversal in body size diminution.

DISCUSSION

That this study found that Great Plains bison underwent dramatic body size reduction

since the end of the Pleistocene is not surprising (MacDonald 1981; Wilson 1975, 1980).

Similar chronoclinal diminution has been reported for bison from Eurasia (Vereschchagin

1959), Beringia (Sher 1971), and elsewhere in North America (Butler et al. 1971; Lyman

2004b). Bison was just one of a number of taxa that experienced rapid size change

through the Late Quaternary. Several large-bodied herbivores, such as Alces , Equus ,

Cervus , Odocoileus , Mammuthus , and Rangifer , and a suite of carnivores underwent body size reduction during this period (Kurtén 1973; Kurtén and Anderson 1980; Guthrie 158

1984, 2003; Purdue 1989; Martin and Barnosky 1993; Geist 1999). While Alces , Bison ,

Cervus , Odocoileus , Mammuthus , and Rangifer were subject to at least modest human predation, Guthrie's (2003) study of Beringian Equus demonstrates that significant body size reduction can occur in the absence of human hunting. Likewise, there is little reason to suspect that the diminution in carnivores was related to direct human exploitation.

The overhunting hypothesis does not seem to fit the available evidence for a number of reasons (Purdue 1989; Guthrie 2003; Lyman 2004b). First, bison diminution was not simply a shift in the mean size of animals, but also represents a reduction in the maximum size bison could attain. This size reduction was so significant that by the Late

Holocene bull bison were about the same size as cow had been during the Late

Pleistocene. In addition, the largest Late Holocene specimen overlaps in body size with the smallest Late Pleistocene specimens.

There also appears to be significant temporal disjunction between the earliest colonization of human hunters in the Americas and periods of major body size change.

Finally, mortality profiles from multi-animal kill localities do not show the anticipated systematic bias for young animals expected with high levels of predation pressure. In fact, if anything, the opposite of the expected relation is found: periods of rapid reduction in body size are generally associated with populations exhibiting an overrepresentation of prime age animals.

The evidence considered here does not address how size reduction relates to shifts in life strategies, such as developmental rate, onset of sexual maturity, or birth synchronization (e.g., Raia et al. 2003). However, Bement's (1999) cross-element study 159 of bison bone size during sequential somatic growth periods from the 12,000-year-old

Cooper site suggests that the timing of growth was generally the same for Late

Pleistocene as it is for Late Holocene bison. If this conclusion is accurate, then during the entire Late Quaternary it is anticipated that most growth for bison occurs within the first three to four years of life, after which the growth curves flatten significantly (Berger and

Peacock 1988; Green and Rothstein 1991; Krasinska and Krasinski 2002). Accepting the assumption of a negative relationship between post-pubertal growth and reproductive performance for bison (Green and Rothstein 1991; Berger and Cunningham 1994),

Bement's data would suggest Late Pleistocene bison likely reached sexual maturity at about the same developmental stage as do modern bison. Therefore, it seems unlikely that the observed body size diminution resulted from a shortening in the developmental cycle.

Similarly, tooth eruption studies for bison suggest that patterns in birth synchronization do not appear to have changed significantly over the last 14,000 years (Todd et al. 1996).

If hunting pressure does not seem to have affected bison body size, then one must look for evidence of an environmentally induced change in bison morphology.

Unfortunately, as previously discussed, there are a number of non-mutually-exclusive environmental factors that could contribute to the changes in body size observed in this study (Dayan et al. 1991; Ashton et al. 2000; Allen et al. 2006; Rodríguez et al. 2006). A single factor alone probably cannot explain the pattern for Late Quaternary bison diminution. Most likely the decrease in body size was a multifunctional response to changing thermal conditions, decreases in plant productivity and food quality, and compositional changes to the predator guild preying on bison. Each of these factors will 160

be considered in turn, although discussions will only briefly touch on diminution as a

response to temperature change or predator release and instead focus more fully on how

changes in plant productivity and diet quality affect body size.

Temperature Change

Since I have already discussed the relationship body size and temperature at length,

this section will only briefly consider the topic. As the Late Quaternary diminution in

bison occurred within the context of a northern latitude transition from cool Pleistocene climate to a warm Holocene climate, it would make intuitive sense to attribute at least some of the changes observed here as reflecting a Bergmannian chronocline. While it is true bison must have adapted to the new thermal conditions faced at the end of the

Pleistocene, as discussed previously it is not a simple task to isolate the influence of temperature on body size without also considering covariant shifts in resource availability and plant productivity (Blackburn and Hawkins 2004; Yom-Tov and Yom-Tov 2005;

Rodríguez et al. 2006). As noted by other researchers, whatever effect climate had on bison body size it was certainly nonlinear. For example, a simple linear thermal relationship would predict that maximum bison body size occurred during the Last

Glacial Maximum and minimum body size during the Early Holocene. Neither prediction is supported with the available data. Furthermore, the 1000-year-long Younger Dryas, a period of global temperature decline, should be a time of body size increase and is not.

Although the available data on body size changes is somewhat contradictory for this 161 period, body size either remains stable or more likely decreases during this period.

In summary, while I do not disregard the effect thermal conditions have on long-term shifts in body size, evidence suggest that that role needs to be considered within a suite of other (related) environmental factors.

Predator and Competition Release

Throughout most of this chapter I have only considered the role human predation had on bison. However, there were a host of non-human predators preying on bison as well.

The North American landscape was populated by a guild of very large predators during the Late Pleistocene, including Smilodon sp. , Homotherium sp. , Miracinomyx trumani ,

Canis dirus , and Panthera atrox (Kurtén and Anderson 1980; Anderson 1984) . Large size, effective weaponry (teeth and claws), and even communal hunting strategies of some, made these species effective hunters of bison and other large herbivores

(Christiansen and Harris 2005; Coltrain et al. 2004; Feranec 2004). Despite poor chronologic resolution of the disappearance of these large carnivores, there appears to be at least a rough correspondence between the time when bison and other large herbivores undergo diminution and the disappearance of these large predators (Kurtén and Anderson

1980; Anderson 1984; Mead and Meltzer 1984; Grayson 1991). As a result, it is necessary to consider the potential effects of the disappearance of these large predators on bison morphology. 162

Geist (1989; c.f. Matheus 2001) suggests the large body size of Late Pleistocene

herbivores was in part a response to an intense predatory regime and interspecific

competition with other large-bodied herbivores. Within this model, the extinction of large

predators would release bison from a need to maximize body size to compete with these

super-predators. Therefore, the advantage of a large body size for bison vanished when there were no longer large predators to escape from or other large herbivores with which to compete for food. This change would give selective advantage to smaller individuals and those able to reproduce at higher rates.

Although Geist's model is insightful and intriguing, unfortunately it is also difficult to evaluate due to our poor understanding of the timing of predator extinction (Mead and

Meltzer 1984; Matheus 2001). With the available data it is not possible to precisely say whether the disappearance of these large predators preceded or followed the onset of diminution in bison. Moreover, although the loss of predator pressure could potentially explain the onset of the diminution of bison at the end of the Pleistocene, it is of less value for understanding the continued decline in body size during the middle Holocene.

During the Holocene the non-human predatory guild that preyed upon bison was primarily bears and wolves, and there does not appear to have been significant changes in either of these populations that would explain the post-Pleistocene body size shifts outlined above.

While changes to the structure of predator guild at the end of the Pleistocene likely played a role in bison diminution, other factors related to intraspecific competition also probably played a role. The major extinction event at the end of the Pleistocene likely 163

heralded dramatic changes in the level of intraspecific competition and population

densities in bison. The disappearance of mammoths, horses, and camels from the

landscape would have left a foraging void to be filled by bison. This filling of the open

niche by bison was likely associated with a dramatic increase in the bison populations

(McDonald 1981). Higher levels of intraspecific competition would have selected for smaller individuals (Wilson 1978).

Stable Carbon Isotopes and Grassland Ecosystems

As discussed above, Guthrie (1984, 1995) and others (Geist 1987; Purdue 1989,

1993; Lundelius 1990) suggest that size reduction in Late Quaternary fauna was likely a response to changes in the quality and duration of availability of spring and summer forage. While this model is quite valuable, as originally presented it provides no direct

evidence for a change in forage quality. To confirm a decrease in forage quality

correlated with the decline in bison body size, one needs an unrelated dataset that can be

used to estimate forage quality.

Stable-carbon isotope ratios from soil organic matter and bioapatite from animal teeth

and bones may provide a mechanism for evaluating whether changes in grassland

composition contributed to the pattern for Late Quaternary diminution in bison. Changes

in the isotopic composition of grasslands would have a great influence on a grazer's body

size, which highlights the important phenological and nutritional differences between C 3 and C 4 plants. 164

Under ideal precipitation, the C 3 plants (grasses, shrubs, and trees) sprout twice a

year: once during spring rains and again with fall precipitation. In contrast, C 4 grasses

grow best when rain falls during the hot months of the year, so their life cycle is limited

to the middle of summer (Waller and Lewis 1979; Paruelo et al. 1996; Tieszen et al.

1997). This pattern means that there will be important differences in the seasonal timing

and duration of availability of high-quality young grasses to large land mammals such as

bison (Dickinson and Dodd 1976). Also, climatic conditions that select for C 3 plants

result in a significantly longer growing season than do climatic conditions that select for

C4 grasses. Just as important as the duration of primary growth, there are inherent

nutritional differences between C 3 and C 4 plants that mean they have different nutritional value to herbivores (Bamforth 1988). Most important among these is that C 3 plants are more digestible and have higher protein levels than do C 4 plants, both factors that greatly influence growth in young animals.

The total effect of these phenological and nutritional differences between C3 and C 4

grasses suggests that to maximize growth, bison should eat a diet composed of a high

content of C 3 plants with a smaller contribution of C 4 plants. Such a diet provides a steady flow of high-quality forage for an extended part of the year.

Paleoenvironmental studies of the isotopic composition of Great Plains grasslands suggest there may have been a non-linear relationship between the grassland composition and body size reduction in bison. The rest of this section will focus on two periods

(13,000–11,000 calBP and 8000/6000–2000 calBP) when bison underwent their most 165 significant diminution, which correspond to times of significant increase in C 4 biomass across the Great Plains.

Limited isotopic data from deposits older than about 14,000 years make it difficult to determine what most Late Pleistocene grasslands were like, but it appears that prior to about 12,000 calBP C 3 plants dominated vegetative communities across much of the

Great Plains. For example, Coltrain et al. (2004) and others (Akersten et al. 1988) demonstrate that the very large Late Pleistocene Rancho La Brea bison had a diet consisting almost entirely of C 3 plants. Paleoclimatic studies from Oklahoma and Texas indicate that the Southern Plains also supported low proportions of C 4 grasses during this period (Bement et al. 2007; Boutton et al. 1998; Humphrey and Ferring 1994; Nordt et al.

1994, 2002; Koch et al. 2004). The mixed grasslands of the Central Plains during this period appear to be dominated by C 3 plants (Hajic et al. 1998; Johnson and Willey 2000;

Olson and Porter 2002). Far northern grasslands likely had very low proportions of C 4 grasses at this time (Boutton et al. 1998).

The first significant decline in bison body size corresponds to a period of increased C 4 production between about 14,000 and 11,000 calBP. A series of studies from across the southern Great Plains (Holliday 1995, 1997; Nordt et al. 2002; Meltzer 2006) demonstrates this period of climatic change brought a significant increase in the proportion of C 4 grasses due to increased drying and shifts in the timing of precipitation.

Although the Northern Plains at this same time appear to have been significantly wetter than the south, both Lovvorn et al.'s (2001) study of the isotopic composition of archaeological bison bone from Wyoming and Boutton et al.'s (1998) analysis of soil 166

organic carbon from a loess sequence in North Dakota suggest that there was an increase in the relative proportion of C 4 plants in northern grasslands at least by the end of this

period.

The next period of dramatic bison diminution occurs began around 8000–6000 calBP,

which correlates with an extended period during the middle Holocene (8000–4000 calBP)

when very warm dry conditions dominated much of the Great Plains. During this period

grasslands, especially in the south, attain some of their highest proportions of C 4 grasses

for any time during the Late Quaternary (Humphrey and Ferring 1994; Nordt et al. 1994,

2002; Holliday 1995; Kelly et al. 1998; Nordt 1998; Lovvorn et al. 2001). This change

was so dramatic that in parts of the Southern Plains there was a near local extirpation of

C3 plants (Humphrey and Ferring 1994). At the same time C 4 grasses dominated the

mixed grasslands vegetation of the Central Plains (Kelly et al. 1998; Olson and Porter

2002). However, such dramatic transformation in grasslands composition never occurred

in northern latitudes. At no time in the past did C 4 grasses dominate vegetation

communities anywhere in the Northern Plains. This may be an important reason why

bison in the north do not appear to have reached the extremely small sizes of the southern

bison.

The last 4000 years witnessed a general amelioration of Plains climates, with slightly

decreased temperatures and increased precipitation. Despite these changes, the isotopic

record suggests the content of C 4 grasses stabilized at relatively high levels, but these

were lower than the very high levels attained during the mid-Holocene (Kelly et al. 1993; 167

Nordt et al. 1994, 2002; Humphrey and Ferring 1994; Kelly et al. 1998; Lovvorn et al.

2001).

The apparent underrepresentation of very old and very young animals in assemblages during periods of dramatic body size diminution (Early Paleoindian and Archaic) possibly also related to the shift in grassland composition, especially considering these changes relate to a decrease in food quality or digestibility. Very young and very old animals generally have lower fat stores than prime age animals, and therefore probably cannot survive as long in times of food shortage (Speth 1983). Also, these groups are more susceptible than prime adults to the ravages of disease and natural predation brought on during periods of environmental stress. If these periods of environmental stress result in animals having to survive on a lower nutritional plane, which would likely have occurred in a diet consisting of a high C 4 content, then it is not surprising that there is a higher rate of mortality among the most vulnerable components of a population.

CONCLUSION

This research has important implications for the role humans played in the extinction of the Pleistocene megafauna, not because it speaks directly to whether human predation caused the extinction of large land mammals, but because it provides a context for understanding the nature of baseline ecological conditions at the time of colonization and the level of human hunting pressure at the time of the Pleistocene extinctions. Bison are an ideal case study for this type of analysis, because even though they did not become 168

extinct at the end of the Pleistocene, they were important in the economy of Paleoindian

(and later) hunters on the Great Plains. As a result, if human hunting was intense enough to drive a large number of land mammals extinct, it should also have had an effect on the body size of bison.

The hunting hypothesis does not explain the changes in Late Quaternary body size for several reasons. First, the shift in bison body size represented a decrease in the maximum potential size of bison, not just a change in mean size. Second, the timing of diminution contradicts the hunting hypothesis prediction, specifically that the onset of diminution post-dates the appearance of human hunters in North America and that later periods of body size decline appears unrelated to human demographic changes. Finally, the kind of age structure distortion observed in the data (i.e., old adult bias through time) seems to relate to environmental condition not hunting pressure.

In contrast, evidence suggests that overall diminution in bison body size was not the result of a slow steady change through time but instead occurred in brief periods of dramatic size reduction within long periods of relative stability. A number of environmental factors likely contributed to the decline in bison body size. The periods of most rapid size reduction are correlated with times of reorganization of grassland composition, changes in predator pressure and intraspecific competition, and climate change over the last 14,000 years. Specifically, bison body size declined during periods when C 4 grasses made a proportionally greater contribution to their diets. Because C 4

plants are less digestible, have lower protein levels, and are more seasonal in their

availability than are C 3 plants, somatic growth in bison would have been limited. 169

Changes in forage quality were probably not the only factors selecting for a smaller body size. The extinction of the Pleistocene megafauna likely relaxed natural predation pressure, and that would have reduced the selective advantage that a large body size provides in protection from predators. The Early Holocene also probably witnessed a dramatic increase in bison populations. As bison populations became more gregarious, greater intraspecific competition for food and perhaps changed social dynamics would have resulted in greater selection for small-bodied animals.

Bison provide a window on the nature of the Pleistocene landscape and can potentially help clarify the reasons that so many of the other large-bodied mammals did disappear. If hunters had minimal effect on a species for which there is unquestionable evidence for predation, one wonders how much influence humans could have had on the

33 genera of extinct Pleistocene fauna for which there is no direct evidence of predation

(Grayson and Meltzer 2002).

Also, if bison body size changes are instead correlated with changes in grassland conditions over the last 14,000 years, which resulted in greater seasonality and decreased forage quality and availability, these same processes may have contributed greatly to the extinction of other large prey that were less plastic in physiology and diet than bison. The nutritional limits imposed by changes in grassland conditions would have strongly selected for species that could rapidly reduce their absolute body size as well as those that could take advantage of lower quality, more seasonal forage.

With the onset of full Holocene conditions, shifting climatic conditions led to a decrease in overall forage quality. First, increased aridity associated with atmospheric 170 warming probably resulted in an overall decrease in grassland productivity. As a result, the beginning of the decline in bison size may have been directed more by decreased forage productivity as a result of overall aridity than an increase in proportion of poor- quality C 4 grasses in the bison diet. Guthrie (1984) has already noted that ruminants, such as bison, likely had an advantage over monogastric species, such as camels, horses, and elephants, because they could survive on the lower quality, higher fiber, and more toxic diet that characterized the short-lived, high C 4 quantity summer grasslands of the Great

Plains at the end of the Pleistocene. In addition, the abbreviated growing season that marked the end of the Pleistocene placed another selective disadvantage against the monogastric species, which generally had much longer gestation lengths than those species that survived (Guthrie 1984). In summary, it appears that the Pleistocene-

Holocene transition represents a significant habitat change, to which bison could adapt but to which others were unable to make the necessary changes.

The results from this chapter on bison osteometrics and mortality and grassland isotopes lays a foundation for the following chapters of this dissertation, which consider

Paleoindian population densities and predator-prey relations with large and small game.

171

CHAPTER 5: REGIONAL DIFFERENCES IN PALEOINDIAN LITHIC

ASSEMBLAGES, OCCUPATIONAL INTENSITY, AND DURATION

Faunal remains are only one aspect of archaeologically visible land-use patterns. In this chapter, I review how chipped stone assemblages are used to characterize Great

Plains Paleoindian land-use strategies. Given my emphasis on regional variation in

Paleoindian behavior, I explore how the compositions of Paleoindian lithic assemblages reflect differences in the constraints and opportunities in different settings. A landscape approach allows for systematic evaluations of spatial covariance between ecological parameters and human behavior (Potts et al. 1999). Similarly, long-term remodeling of regional vegetation and animal communities should also result in dramatic changes in a forager's behavior. By comparing records from a number of sites distributed across different ecological settings, there is the potential to define alternative strategies of human land use and to study how they altered through time (Potts et al. 1999: 751).

This study, like others addressing the Paleoindian period, must confront problems related to limited sample size, incomplete excavations, unequal regional coverage, and poor published records. However, because this study uses information from a number of sites, a regional approach reduces sampling error that can often be a problem when the focus is on a single site. Brantingham (2003:493) explains, "individual archaeological assemblages may be treated as repeated random samples of different sizes from the mobile toolkit. In the aggregate, multiple archaeological assemblages should provide a reasonable complete picture of stone raw material procurement dynamics." 172

For my exploratory analysis of large-scale patterns of Paleoindian land use, I use simple quantifiable measures of site function, artifact density, and percent non-local lithic raw material drawn from published data from 258 Paleoindian archaeological components from sites across the Great Plains and adjacent mountain environments

(Table 5.1). To facilitate regional comparisons, sites were grouped into one of the major physiographic regions of the project area (Fenneman 1928, 1931). I focused on the potential correspondence between land use and environmental settings by looking at inter- and intra-group variation for sites within different physiographic regions. Using published age estimates for the study sites, I examined how site function, artifact density, and use of local and exotic chipped stone varied over time and by region.

This chapter begins with a brief review of how prior hunter-gatherer studies have linked forager mobility and land use with environmental constraints. A discussion of the recent debate on Paleoindian life ways follows, with specific attention paid to how these differing views conceptualize adaptation to environmental conditions and how this is expressed in human mobility. I then describe the expected material correlates to hunter- gatherer mobility, and review the extant Paleoindian archaeological record for the Great

Plains and surrounding areas. This chapter concludes with my analysis and interpretations of these results.

173

Table 5.1. Sites used in this chapter Pitblado and Camp 2003 Anderson and Semken Site Reference Cherokee-Sewer 12 Mile Creek Hill 2002b 1980 Hovde and Blakeslee Clary Ranch Hill 2001 14SG516 Dick and Mountain 1977 Claypool Owsley and 1960 14WY9003 Brusweldheide 1977 Colby Frison and Todd 1986 48JO202 Frison 1976 Cooper Bement 1999 Custer County 48UT375 Smith et al. 2003 Schultz 1932 4Moore-Neal Litinionek et al. 2003 Bison Adair-Steadman Tunnel 1975 CW Cache Holen and Muniz 2005 Hofman and Ingbar Deadman Wash MacKey et al. 1982 Adobe Figgins 1933; Haynes et 1988 Dent Frison and Stanford al. 1988 Agate Basin 1982; Hill 2001 Devil's Mouth Johnson 1961 Bamforth 2002b, Domebo Leonhardy 1966 Allen Bamforth et al. 2005 Drake Stanford and Jodry 1988 Anadarko Cache Hammatt 1970 Dunn Ebell 1988 Anzick Owsley and Hunt 2001 Elida Hester 1962 Arenosa Shelter Collins 1976 Finley Haspel and Frison 1987 Aubrey Ferring 2001 Fletcher Forbis 1968 Meltzer et al. 2002; Baker Cave Collins 1976 Folsom Surovell et al. 2003, Meltzer 2006 Barger Gulch Agogino and Parrish 2005 Fowler-Parrish Barton Gulch Davis et al. 1989 1971 Fulgham and Stanford Beidleman Ranch Suhm 1960 Frasca Benz Root 1997 1982 Big Black Williams 2000 Frazier Borresen 2002 Big Lake Turpin et al. 1997 Gault Collins 2002 Billy Ross Galm and Hofman 1984 Gordon Creek Muniz 2004 Black Mountain Jodry 1999b Hanson Ingbar 1994 Hester 1972; Johnson Helen Lookingbill Kornfeld et al. 2001 and Holliday 1997; Irwin-Williams et al. Blackwater Draw Saunders and Daeschler Hell Gap 1973; Byers 2002; Knell 1994 et al. 2002 Blue Point Smith et al. 2003 Heron Eden Corbeil 1993, 1995 Mallouf and Mandel Bobtail Wolf Root 2000 Horace River Byerly et al. 2005; 1997 Bonfire Shelter Dibble and Lorrain 1968 Horner Frison and Todd 1987 Burntwood Creek Hill et al 1992 Horn Shelter Redder 1985 Busse Hofman 1995 Indian Creek Davis and Grieser 1992 Caribou Lake Pitblado 2000 Jakes Bluff Bement and Carter 2003 Carter/Kerr-McGee Frison 1984 JC Putnam Roberts 1945 Kornfeld and Frison Casper Frison 1974 Jerry Craig Cattleguard Jodry 1999a 2000 Chance Gulch Pitblado et al. 2001; Jim Pitts Sellet 2001 174

Table 5.1. Sites used in this Chapter OV Clary Hill et al 2006 Site Reference Packard Wyckoff 1989 Jones-Miller Stanford 1984 Pavo Real Collins et al. 2003 Jurgens Wheat 1979 Perry Ranch Hofman and Todd 1997 Kanorado Mandel et al. 2005 Pine Spring Kelly 2002 Krmpotich Peterson 2001 Plainview Sellards et al. 1947 Hofman and Blackmar Stanford and Patten Laird R-6 1997 1984 Harrison and Killen Rattlesnake Pass Smith and McNees 1990 Lake Theo 1978 Ray Long Wheeler 1995 Lamb Spring McCartney 1983 Red Smoke Knudson 2002 Lange-Ferguson Hannus 1990 Rex Rodgers Willey et al. 1978 Levi Rockshelter Alexander 1963, 1982 Richard Beane Thoms et al. 1996 Lewisville Crook and Harris 1957 Ryan Hartwell 1995 Lime Creek Bamforth 2002b Sailor Helton Mallouf 1994 Wilmsen and Roberts San Jon Hill et al. 1995 Lindenmeier 1978 Barbour and Schultz Scottsbluff Linger Jodry 1999b 1932b Lipscomb Hofman and Todd 2001 Seminole-Rose Collins et al. 1997 Lone-Wolf Creek Cook 1925 Frison and Stanford Sheaman Lubbock Lake Johnson 1987 1982 MacHaffie Forbis and Sperry 1952 Shifting Sands Hofman et al. 1990 Mammoth Agogino and Galloway Bonnichesen et al. 1992 Sister's Hill Meadows 1965 Holliday 1997; LaBelle Slim Arrow LaBelle 2001 Mark's Beach and Meltzer 1996 Sutter Katz 1971 McClean Ray and Bryan 1938 Tim Adrian O'Brien 1984 Medicine Lodge Upper Twin Kornfeld and Frison Frison 1976, 1991 Creek Mountain 2000 Barbour and Schultz Walsh Stanford 1997 Merserve 1932a Waugh Hill and Hofman 1997 Cannon and Cannon Wetzel Cassells 1983 Meyer-Hindman 2004 Wilson Leonard Collins 1998 Miami Holliday et al. 1994 Holliday and Meltzer Midland 1996 Mill Iron Frison 1996 Sellards 1955; Warnica Milnesand and Williamson 1965 Mona Lisa Wilson 1974 Mummy Cave Husted and Edgar 2002 Nall LaBelle et al. 2003 Nelson Cassells 1983 Niska Meyer 1985 Norton Hofman et al. 1995 Olsen Chubbuck Wheat 1972 175

MOBILITY AND RESOURCE DISTRIBUTION

A number of researchers have shown that resource availability and distribution strongly affects hunter-gatherer land use (Bamforth 1988; Binford 1980; Kelly 1983,

1995). As discussed in chapter 2, traditional behavioral ecology models assume the spatial distribution of prey is random, and therefore, the probability of a forager encountering prey or other key resources is uniform across space (e.g., Simms 1987;

Stephens and Krebs 1986). However, in the real world resource availability is patchy both in space and through time (Gremillion 2002; Hawkes et al. 1982; MacArthur and

Pianka 1966; Wien 1976). This variation relates to the forager's ability to find key resources (such as lithic, food, wood, water). Spatially, resources can vary in terms of patch location, size, and congruency. Thus, the key measures of temporal variability are both the constancy (continuous availability) and contingency (availability based on seasonality or certain natural events) of resource availability.

How we use the nature and structure of resource distribution to understand forager mobility has been an ongoing concern. Over 25 years ago, Lewis Binford (1980) introduced the concept of effective temperature (ET: a simultaneous measure of the intensity of solar radiation and its annual distribution) to the anthropological community as a concept to help understand resource distribution in different environmental settings as well as a framework for interpreting human settlement patterns in those settings. In general, environments with low ET (ET = 8–10) measures (such as tundras) have widely spaced patches of low primary plant productivity, while high ET (ET = 22–25) 176 environments (tropical rainforests) are associated with tightly spaced high primary plant productivity. For human foragers living in relatively low (ET = 11–16) ET grassland settings, such as the Great Plains of North America, much of the energy stored in plants is inaccessible because it is held within the inedible parts of plants, such as leaves and stems. In these cases, humans can only subsist on the portions of plants that are edible

(often with intensive processing) or exploit animals that can directly exploit the plants

(Kelly 1983: Table 3). The animal resources in these environments are often large and gregarious and easily monitored by humans (Kelly 1983).

It is fairly obvious how differential amounts of solar radiation affect plant growth, but how does human land use respond to the nature and structure of available resources?

Binford (1980) suggests a correspondence between environmental structure and forager settlement strategies. He showed that groups living in equatorial zones and high Arctic environments have high residential (movement of the entire local social group) mobility.

In more temperate settings, residential mobility can often be constrained by seasonal food scarcity and water availability when reliance on stored resources is important for group survival (see discussion in Kelly 1995: 117).

Robert Kelly (1983) took up this issue and made empirical correlations between the structure of hunter-gatherer mobility and environmental settings. Specifically, Kelly

(1983: Table 4.1, 1995) examined how five separate measures of mobility (number of residential moves per year, average distance moved, total distance moved per year, total area used, and average length of logistical foray) varied with gross abundance and distribution of food. 177

This work suggests high residential mobility is expected for foragers living in

environments where food choices are limited and/or critical resources are widely spaced.

These conditions happen in (1) cold environments with low ET and low primary

productivity, (2) tropical forests with high ET where primary productivity is high but

inedible to humans, and (3) temperate grasslands where secondary biomass (animals) is

highly accessible and mobile. In these types of environments, it is often more efficient for

foragers to move their residence than to invest great effort into making long, often

unsuccessful, logistical forays after resources that cannot be easily monitored from their

residence. While increased mobility may be a common response for groups living in

these different environments, the nature of that mobility does differ.

In temperate and cold environments with low primary productivity, groups move less frequently, but over greater distances, than groups in tropical settings. The increase in distance moved and size of territory used by these groups relates to their increased dependence on hunting. There are several factors influencing this relationship. First, foraging groups dependent on plants in areas where primary productivity is high and accessible to humans need to intensively cover the territory they use to efficiently locate and acquire the plant resources (Kelly 1983). Therefore, while they move more frequently, any specific move is over a short distance.

In contrast, large game hunters often need to cover much larger territories but do not need to be as thorough. This occurs because the large game these groups exploit are themselves highly mobile, since they live in environments with relatively low primary

(plant) biomass and to get adequate subsistence they need to exploit a large number of 178

patches. Thus, to find their prey the hunter too must cover a lot of territory (Kelly 1983).

Fortunately for the hunters, these large animals primarily live in social groups, so they are

more easily monitored by foraging groups moving through the landscape (e.g., Frison

2004; Haynes 2006). Kelly (1995: 125) and other (Byers et al. 2005; Ugan 2005a) note

that for temperate environments, as primary productivity increases and/or temperature

decreases, large game becomes less abundant and more dispersed.

Residential stability is anticipated for groups that live in environments with high

spatial congruence between important resources. Also foragers often are residentially less

mobile in areas where they can reliably expect to encounter important bulk resources,

such as caribou migrations or salmon runs, and can put up seasonal stores of food

(Binford 1980). Kelly (1983: 292), for example, shows that hunter-gatherer groups reliant

on aquatic resources have lower residential mobility than would be anticipated by

measuring environmental primary productivity alone. Finally, in very dry environments

human groups may be tethered to critical water resources and therefore depend on

logistic forays to acquire other resources (Kelly 1983; Taylor 1964).

These prior studies also make it clear that the distribution of non-food resources,

critical to the survival of foraging groups, also influences forager mobility and settlement.

Fixed resources, such as potable water and firewood, are obviously highly predictable in

terms of location and availability, but vary in quality and patch size. The shape and

nature of these non-food resources influence how foragers move through the environment

(Kelly 1983; MacDonell and Wandsnider 2003). In an environment where key resources are distributed as points, such as a small spring, playa, or small herd of bison, 179 exploitation is focused just on that point, because the other resources cannot be found in that region around the point resource. Once a forager harvests the resource(s) around that point, they often have few other options in the area and must leave the area to search for other resources. This is because in some environments, such as temperate grasslands, the long commuting time searching for a particular resource does not guarantee a forager will encounter other resources. In contrast, resources distributed in a linear form or over a very large region, as in streams or rivers, allow foragers to move over great distance exploiting other resources, while still utilizing the original resource.

PRIOR VIEWS OF PALEOINDIAN LIFE WAYS

As described in chapter 1, in the Great Plains and Rocky Mountains Paleoindian land use is often portrayed in terms of the degree and type of mobility practiced and the lithic technology utilized linked to spatial and seasonal variation in resource availability. Two end-member models describing Paleoindian land use and lithic use strategies have been presented.

Mobile, High-Tech Foragers

In this view, Paleoindians were constantly on the move, and carried a flexible and highly reliable biface-based tool kit (Goodyear 1989; Hofman 1992, 1994; Kelly 1988).

This extreme mobility is linked to their status as specialized hunters, preying on 180

mammoth and bison on a year-round basis. This model was initially developed to

characterize the Early Paleoindian period, but Plains archaeologists have applied this

model to the entire Paleoindian period (Bamforth 2002a: 57).

Kelly and Todd (1988) have characterized this strategy as technology-focused,

whereby foragers move through the landscape practicing similar behaviors regardless of

environmental setting or season. According to Hofman (1992:198), this required hunters

to move continuously from kill to kill. These hunters had an apparently limited

commitment to storage of food, and had they not immediately sought new prey they

would have found themselves in an area of depleted resources with little time to locate

alternative food resources.

There are at least four artifactual implications of this model (see discussions in

Bamforth 2002a). First, most sites, regardless of location, should show fairly low

occupational intensity. Second, as the same suite of behaviors occurred at most sites

(kill/carcass processing), artifact assemblages at different sites should be fairly uniform in

terms of assemblage size and diversity. Sites that are associated with large artifact

assemblages reflect either a palimpsest of multiple occupations or lithic workshops

(Hofman 1999). Third, most Paleoindian assemblages are expected to contain at least some stone from distant raw material sources, often from sources 100–300 km away

(Amick 2000; Hofman 1999, 2003; MacDonald 1998). Finally, lithic assemblages at these sites should be derived from highly curated-biface reduction strategies; that is, bifaces are used as core and tools, and many of the other tools should be made on biface- 181 struck blanks. Because these tools are carried for extended periods of time, tools should show evidence of extensive resharpening and reworking.

Low-Mobility, Low-Tech Foragers

In this view, Paleoindian foragers intensively and repeatedly occupied regions of high ecological diversity, often constructing features and structures (e.g., Irwin-Williams et al.

1973). Their technology was not based on high-quality bifacial tools, but instead these groups used informal reduction of locally available raw materials. While criticisms of the mobile, high-tech forager model have been presented by a number of authors (Cannon and Meltzer 2004; Gero 1993; Hill 2001), Bamforth (1991, 2002a, Bamforth et al. 2005) has been the most vocal critic, and has presented a very different view of Paleoindian life ways. He argues that Paleoindians practiced a low residential mobility place-focused strategy, in which foragers' behavior responded to seasonal environmental changes on a local scale.

He also stresses that there is little evidence that Paleoindians predominantly used bifacial cores. Instead, most tools, including finished points and knives, were made from flakes from non-bifacial cores (Bamforth and Becker 2000). Furthermore, most

Paleoindian assemblages contain almost no evidence for long-distance transport of raw material; most raw material was instead derived from sources immediately adjacent to or a relatively short distance from the site. The only exception to this pattern is in areas, such as the Southern High Plains, where access to raw material was very limited. 182

Otherwise, in areas where stone is locally available, evidence for exotic raw material is usually extremely low. In fact, Bamforth (2002a:85) even suggests inter-group trade was responsible for some cases where raw materials have been transported hundreds of kilometers from their sources. In addition, there is limited evidence for extensive resharpening in most Paleoindian assemblages. Except for hafted tools, such as points and end scrapers, most tools appear to have been discarded with little or no resharpening.

ARCHAEOLOGICAL EVIDENCE OF MOBILITY

The aforementioned models provide a framework for developing testable hypotheses.

Each has explicit expectations concerning occupational intensity, relative abundance of non-local raw material, and reliance on bifacial technology. I discuss below how these variables can be operationalized in studying regional land use. I assume that opposing models are not necessarily mutually exclusive. A single group of foragers may shift between these end-member strategies depending on the particular environment they find themselves in or the task they hope to accomplish.

Occupation duration or intensity reflects differing levels of residential mobility, whereby a decrease in overall residential mobility results in increased occupation length and intensity at sites. The hypothesis of continuous mobility suggests that all Paleoindian occupations should reflect very low site occupation intensity level, reflected in small artifact assemblages (often just a few projectile points and butchery tools). The alternative view is that land use tracked the distribution of key resources, and areas with a 183

greater density of key resource (e.g., water, wood, food, stone) were more intensively

occupied.

The key to testing this hypothesis is identifying a measure of site use intensity.

Surovell (2003), drawing on work by Schiffer (1987), argues that artifact density

(artifacts per m 2) and relative abundance of locally available lithic raw material to non- local materials (ratio of artifacts of local lithics to artifacts of nonlocal lithics) can be used as rough measures for occupation span (c.f. LaBelle 2005). His argument is based on two assumptions: (1) artifact density is related to occupation span and rate of discard, and (2) as occupation span is lengthened, artifact assemblages should become increasingly dominated by locally available lithic raw materials (especially debitage) (see also Andrefsky 1994; Gould 1980; Kuhn 1991; O'Connell 1977). Archaeologists should be able to use the simple ratios of local to nonlocal artifacts and number of pieces of debitage to the number of nonlocal tools to monitor per capita occupation span (Surovell

2003). Although this test cannot be easily applied to sites with very few artifacts (e.g., kill sites) or those lacking either local or nonlocal raw material, it provides an easy comparative measure of occupation intensity for residential camps, although it is based on a very gross scale view of past behavior.

The second component of this study is to determine if site intensity correlates with resource availability. If land use is non-random, then sites with high occupational intensity, which might represent residential camps, should be preferentially situated in protected settings characterized by high ecological diversity (Hofman 1999), or associated with key resources (e.g., lithic raw material, water, wood, shelter, large and 184 small game). Similarly, diachronic decrease in group mobility or regionalization should be reflected with the appearance of increasing proportion of sites with high occupational intensity through time.

PALEOINDIAN LAND USE RECORD

The data here come from published site reports of 258 Paleoindian archaeological components from sites across the Great Plains and adjacent mountain environments

(Figure 5.1). Each component in this study was assigned to one of four possible temporal groups that roughly corresponds to major artifact based chronologies for the Paleoindian period: >12,700 calBP (Clovis, early Goshen, early Folsom); 12,700–11,250 cal BP (late

Folsom, Agate Basin, Hell Gap, Plainview); 11,250–10,200 calBP (late Agate Basin, late

Hell Gap, Cody, Plainview); and <10,000 calBP (Allen/Frederick, late Cody, Angostura,

Golondrina) (based on Holliday 2000).

Following the approach of previous studies, sites were assigned to major physiographic regions as proxy measures for megahabitat zones (e.g., Amick 1995;

Blackmar 2001; Holliday 1997; Pitblado 2003). These regions include the High Plains,

Edward Plateau, Rolling Hills, Colorado Piedmont/Raton Volcanics, Rocky Mountains, and Northern Plains (Fenneman 1928, 1931). These physiographic regions are very coarse-grain ecological zones and certainly do not represent uniform landscape features; however, these regions do represent large environments of generally comparable geologic, fluvial, and vegetative communities. 185

Figure 5.1. Map of the Great Plains and Rocky Mountains with sites mentioned in this chapter. 186

Although these regions are not perfected for making landscape comparisons, their size and inherent differences in resources composition make them appropriate analytical units for comparing inter- and intra-regional variation in hunter-gatherer land-use strategies.

Site Function

This analysis builds on the work of earlier researchers (Judge 1973; Hofman 1994;

Hofman and Ingbar 1988; LaBelle 2005; Root 2000) by classifying sites based on the presence of particular tools, stage of tool manufacturing, and amount of debitage recovered. As discussed by LaBelle (2005:Table 5.5) more than 12 site types have been described in the Paleoindian literature. Defining these site types is often problematic, because it is not always clear (1) that different site type names actually reflect different behaviors which occurred at these localities (e.g., camps, base camps, or aggregation sites), (2) that all archaeologists use the same traits to define the same site types, and (3) to what degree natural site formation processes (e.g., time averaging, fluvial action, artifact relocation) or archaeological recovery techniques are responsible for aggregating/dispersing the artifact assemblages used to define site types.

However, given how the activities occurring at the site have a strong influence on the composition of zooarchaeological assemblage and may reflect important differences in land-use strategies, I feel that some attempt needs to be made to control for functional differences in site use. As a result, each assemblage in this study was classified into one 187 of three generalized functional categories: multi-function camps, kills/processing, and special function. These are very broad categories that I hope define gross scale differences in site use.

As mentioned in chapter 1, site function categories are based on interpretations of the lithic assemblages. A component was defined as a kill/processing locale if it was associated with a small chipped stone assemblage with a low diversity of tool types.

Weaponry, butchery, and hide-working tools dominate kill/processing assemblages, and chipped stone debitage, if present, consists mostly of tool resharpening flakes and utilized flakes.

Multi-function camps usually have large lithic assemblages with numerous non- weaponry tools, debitage, and preforms in various stages of preparation. Camp assemblages usually exhibit a low proportion of projectile points. When projectile points are present at camps, they are often broken or present only in the form of a point base.

Kill/processing localities were sometimes associated with some physical feature, such as a steep-sided arroyo, that presumably was used to trap animals during a kill. Camping localities often lacked such features, but instead were associated with cooking/heating, storage, or habitation features. Included in the camp designation are localities that have been defined as lithic workshops or quarries. Workshops have been previously described as localities with evidence of intensive tool or biface production and extensive evidence of lithic debris, while quarries are where lithic extraction occurs on site. Like workshops, quarries have abundant evidence for the initial stages of lithic reduction with large quantities of chipped stone debris. 188

Special function localities included caches and burials. Caches are clusters of multiple

tools (often bifaces or points) found together in a small area with no other artifacts in the

vicinity, while burials are any locality with human remains, and are often associated with

various grave goods.

Comparison of Artifact Contents . In order to evaluate whether there are differences in assemblage content among these different site types, I conducted a series of simple comparisons of the proportionality of key artifact types, assemblage size, and diversity for the 111 assemblages used in this study for which I have detailed artifact content data.

My primary goal in this analysis is to evaluate to what degree the characteristics (e.g., assemblage size, assemblage diversity, abundance of particular tool types) traditionally used to define site types in fact differ among the uses in this analysis.

This analysis compared the representation of projectile points to all lithic artifacts

(ppt/total lithic), the number of non-projectile point tools to projectile points

(nonppt/ppt), and the number of pieces of debitage to number of tools (debitage/tools) for the camps, kill/processing, and special function localities. As defined above, the expectation is that kill/processing sites should have a high proportion of weaponry and butchery tools to all other artifact classes, should have high ratio for ppt/total lithic, and very low values for the other ratios. The high proportion of debitage and non-butchery tools expected for camp localities should result in these assemblages having high debitage/tools and nonppt/ppt ratios and very low ppt/lithic ratios. Finally, because the

specially function sites used in this analysis are primarily caches, it is expected that they

will have relatively high ppt/total lithic ratios and generally low values for the other two 189 ratios. As seen in Figures 5.2–5.4, the different site type are statistically different from one another in terms of ppt/total lithics (Kruskal-Wallis H = 45.9, df = 2, p <.001), nonppt/ppt (Kruskal-Wallis H = 42, df = 2, p < .001), and debitage/tools (Kruskal-Wallis H=

25.7, df = 2, p <.001).

Pairwise comparisons show that the ratio of number of projectile points to total number of artifacts was statistically higher at kill/processing sites than either camp

(Mann-Whitney U = 254; Z = -6.62; p = < .001) or special function localities (Mann-

Whitney U = 254; Z = -6.62; p = < .001). Camps, on the other hand, had extremely low ppt/total lithic values, all less than .2, and special function sites had generally moderate to low values. However, the differences between camps and special function sites were not statistically significant.

For both the ratio of non-projectile point tools to projectile point and the ratio of debitage to all tools, pairwise comparisons show that camp assemblage have statistically higher values than either kill/processing or special function assemblages. However, in neither case were the differences between special function and kill/processing sites shown to be statistically significant.

These results suggest that at least in a few key areas there are important differences in the composition of the majority of lithic assemblages assigned different functional designations. This does not imply that a few kill/processing assemblages did not have high proportion of non-ppt tools or an abundance of debitage. These few exceptions were clearly outliers. The kill/processing localities with unusually high nonppt tool/ppt ratios were either small, partially excavated bison kills with very small lithic assemblages 190

Figure 5.2. Box plot comparison of the proportion of number of projectile point to number of lithic artifacts from 42 camp, 56 kills/processing, and 10 special function assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H = 45.9, df = 2, p <.001) rejects the null hypothesis for equality of population median among groups.

191

Figure 5.3. Box plot comparison of the proportion of number of non-projectile point tools to number of projectile points from 36 camp, 49 kills/processing, and 4 special function assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H = 42, df = 2, p < .001) rejects the null hypothesis for equality of population median among groups.

192

Figure 5.4. Box plot comparison of the proportion of number of pieces of debitage to number of tools from 37 camp, 39 kills/processing, and 4 special function assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H= 25.7, df = 2, p <.001) rejects the null hypothesis for equality of population median among groups.

(i.e., Laird, Rattlesnake Pass) or assemblages associated with bison processing localities with particularly large numbers of scrapers and utilized flakes (Cherokee-Sewer, Agate

Basin-Agate Basin, Frazier). Similarly, the kill/processing assemblages with outlier values for the debitage/tools ratios were large bison processing localities from which a high number of retouch flakes were recovered (i.e., Clary Ranch, Agate-Basin-Area 3

Hell Gap, and Cherokee-Sewer).

However, because throughout this dissertation comparisons are made between kill/processing localities and camps, some additional analysis was conducted on the 193

relative frequency of different tool types from these site types (Table 5.2). As these data

show, except for the number of projectile points, camp assemblages have significantly

high numbers of total lithics, non-projectile point tools, bifaces and preforms, scrapers,

cores and core tools, utilized flakes and flake tools, knives, denticulates, gravers, and

drills, and debitage than do kill sites.

Overall, these data suggest that camp assemblages are in general larger and more

diverse than are kill assemblages. Therefore, I believe it is reasonable to use site function

as a reliable variable for distinguishing among assemblages that are the result of either (1)

multiple (subsistence and non-subsistence) activities in the case of camps, or (2) activities

primarily directly related to making a kill or processing the carcasses (kill, butchery, hide

working) in the case of kill/processing localities.

Patterning in Distribution of Site Function. Due to limited sample size it is not possible to statistically evaluate whether there are trends in site function through time by physiographic region. However, as Table 5.3 shows, there is a significant non-temporal relationship between site function and physiographic region ( χ2=51.79, df=3, p<.001).

The analysis of standardized residuals indicates this relationship is largely driven by an overrepresentation of kill/processing locales in the High Plains while camp/workshops are abundant in both the Edwards Plateau and Rocky Mountains. In all other regions the frequency of camps/workshops and kill/processing locales are more similar.

When changes in the relative frequency of site function are broken down by temporal period, there is just a moderate trend through time that has little explanatory power ( χ2 =

6.21, df = 3, p = .10) (Table 5.4). Standardized residuals suggest that this trend is driven 194 largely by a slight overrepresentation in the number of kill/processing localities during the earliest time periods (>12,700 calBP; Clovis) and a slight underrepresentation for kill/processing localities for the period between 11,250–10,000 calBP.

Table 5.2. Comparison of frequency of the major tool categories for Kill/Processing and Multi-function Camp localities Site Mann- Artifact N Mean Std. Dev. Prob. type a Whitney U C 36 21824.9 79260 Total lithics 255 <.001 K 47 667 2303 Non-Projectile C 34 253.5 511.4 242 <.001 Point tools K 45 18.6 37.1 C 34 34.6 90.2 Points 981 .45 K 47 17.9 26 Biface and C 24 42.9 112.6 56.5 .003 preforms K 10 4.4 4.8 C 31 72 220.2 Scrapers 303.5 .004 K 26 8.8 9.5 Cores and C 24 40.7 145.8 34.5 .001 Core tools K 9 1.6 1.4 Utilized flakes C 26 93.1 145.4 150 <.001 and flake tools K 22 16.2 30.6 C 15 12.3 13.6 Knives 76 .05 K 14 3.1 4.4 Denticulates, C 31 21.2 35.9 Drills, and 123 .02 K 14 4.4 8.6 Gravers C 33 23480.5 82312.9 Debitage 206.5 <.001 K 39 760.5 2495.4 A C = camps; K = kill/processing

195

Table 5.3. Frequency of Archaeological Components by Site Function and Physiographic Region Kill/ Special Physiographic Region Camps Processing Function Colorado Piedmont/Raton Volcanic 5 6 2 Edwards Plateau 12 2 0 High Plains 30 49 6 Rocky Mountains/Intermountain 71 9 1 Basin Missouri Plateau 8 9 1 Rolling Hills 9 11 3 Note: χ2=51.79, df=5, p<.0001; due to small cell frequency it was necessary to drop special function sites from statistical analysis

Table 5.4. Frequency of Archaeological Components by Site Function and Time Period Camping/ Kill/ Special Time Period Workshops Processing Function >12,700 calBP 11 14 6 12,700–11,250 calBP 39 28 2 11,250–10,000 calBP 43 17 2 <10,000 calBP 42 25 1 Note: χ2=6.21, df=3, p=.10; due to small cell frequency it was necessary to drop special function sites from statistical analysis

Occupational Intensity

As already discussed, models of high residential mobility for Paleoindians mean never staying in one place very long (e.g., Hofman and Todd 2001; Kelly and Todd

1988). This hypothesis would predict that the majority of Paleoindian occupations should reflect low levels of site occupation intensity regardless of location. The alternative view is that land use tracked the distribution of key resources (e.g., water, wood, food, and stone), and in these areas there should be greater occupation density. This section looks at two variables as rough measures for occupation span (Surovell 2003): artifact density 196

(artifacts per m 2) and relative abundance of locally available lithic raw material to nonlocal materials.

Artifact Density . When data from all sites are compared there appears to be significant difference in artifact density at sites in different physiographic regions (F =

2.64; df = 5,83; p = .03) (Table 5.5), which contradicts one of the basic assumptions underlying the high mobility model for Paleoindian land use. Pairwise comparisons of site artifact density from the different physiographic regions suggest that sites within the

High Plains and Rolling Hills have lower artifact density than do sites in other regions.

The most significant differences are observed between High Plains and Rolling Hills sites and those in the Northern Plains and Rocky Mountains.

Table 5.5. Summary Statistics for Site Artifact Density (# Artifact/ m 2) by Physiographic Region Std. Physiographic Region N Mean Range Dev. Colorado Piedmont/Raton Volcanic 9 17.3 31.1 .2–99.1 Edwards Plateau 8 30.8 41.2 .3–124.0 High Plains 33 14.2 36.9 .2–201.6 Rocky Mountains/Intermountain 16 44.4 104.0 .6–385.5 Basin Missouri Plateau 12 177.1 429.5 .1–1448.8 Rolling Hills 10 5.0 9.8 .3–9.5 Note: F = 2.64, df = 5,83, p = .03; pair-wise comparisons suggest that High Plains sites have statistically lower density values than Rocky Mountain and Northern Plains sites. In addition, Rolling Hills sites have lower density values than Rocky Mountain sites.

The simple comparison of artifact density for sites from large physiographic regions

likely masks key aspects of hunter-gatherer lithic use and mobility. That is, foragers in 197 areas of abundant lithic raw materials have less reason to be very efficient in their use of that material than they do when they are in areas distant from raw material sources. This suggests that sites close to raw material sources should have evidence of greater degrees of discard of potentially useable debitage and broken tools and preforms than would occur at sites located far from good sources of lithic raw material (Henry 1989, 1992,

Kuhn 1991, 2005).

Much of the spatial variation in artifact density appears to relate to distance from raw material source and the quality of that lithic source. In general, sites located closer to a lithic source (<50 km) have higher artifact density values, and likely occupation intensity, than do sites located far away (>50 km) from a lithic source (Mann-Whitney U = 370; Z

= -5.03; p = <.001) (Figure 5.5). A significant difference is seen between camps (Mann-

Whitney U = 37; Z = -3.09; p = .002) located at differing distances from lithic sources, and a similar relationship is seen at kill/processing localities, although the statistical significance of this relationship is not as strong (Mann-Whitney U = 125; Z = -1.83; p =

<.07). Perhaps surprisingly, kill/processing localities near lithic sources have comparable artifact densities as camp/workshops located far from lithic sources.

Therefore it appears that the group mean differences observed in Table 5.5 are obscuring a high degree of intra-group variability within each environmental setting

(Figure 5.6). For example, the abundance of generally low-artifact-density kill/processing sites that characterize use of High Plains and Rolling Hills occur alongside relatively high-density camp sites located within alluvial valleys or very near major lithic raw material sources, such as Allen, Clary Ranch, Aubrey, and Hell Gap. 198

Figure 5.5. Box plot comparison of the artifact density (# of artifacts/m 2) of multifunction camps and kill/processing assemblages from sites <50 km and >50 km from lithic sources. Kruskal-Wallis one-way analysis of variance by ranks test (H = 42, df = 2, p < .001) rejects the null hypothesis for equality of population median among groups.

199

Figure 5.6. Map of the Great Plains and Rocky Mountains with site artifact density distribution.

200

Likewise, there are a number of low-density kill/processing localities, such as Bonfire

Shelter, Colby, Horner, Jerry Craig, Niska, and Rattlesnake Pass, located in the usually high-density Northern Plains, Edwards Plateau, and Rocky Mountain regions.

If we look at these data based on smaller-scale habitat differences (alluvial valley, plains/rolling hills, and foothill/mountains) and with consideration of site distance from lithic sources and of site function, there appears to be at least some limited ability to observe important patterning. For example, differences in artifact density for campsites located near lithic sources highlight this regional variation (Figure 5.7). For camps/workshops located within 50 km of lithic sources, sites located in plains/rolling hills settings generally have lower artifacts density values than those in either alluvial valleys or foothill/mountains. While the non-parametric Kruskal-Wallis test (Kruskal

Wallis H = 2.67; df = 2; p = .26) cannot show significant differences across habitat settings, pairwise comparisons suggest that artifact density is statistically lower in plains/rolling hills settings than in alluvial valleys (Mann-Whitney U = 7; Z = -1.84; p =

.08), at least at the .10 level. Similar patterns are seen for camp/workshops located farther

(>50 km) from lithic sources; however, that data is only available from a small number of alluvial valley (n = 1) and foothill/mountains sites (n = 2) which does not allow statistical comparison of these differences.

Density trends through time . Artifact density values do not appear to significantly change through time. When all sites are compared there do not appear to be any significant trends through time (r 2 <.001) (Figure 5.8), despite the expectation that residential mobility decreased during the early Holocene (e.g., Bamforth 2002a, 2002b; 201

Hill 2001). No statistically significant trends are seen when comparisons are made

separately for site function or environmental setting. Nor are any long-term trends

observed when artifact density trends are examined controlling for site function and

distance from a lithic source (less than and greater than 50 km). Overall, it appears clear

that factors that are controlling for artifact density are not changing through the

Paleoindian period.

Figure 5.7. Box plot comparison of the artifact density (# of artifacts/m 2) for 3 plains/rolling hills, 15 alluvial valleys, and 9 foothill/mountain multifunction camp sites located <50 km lithic sources. Kruskal-Wallis one-way analysis of variance by ranks test (H = 2.67; df = 2; p = .26) does not rejects the null hypothesis for equality of population median among groups. 202

Figure 5.8. Scatter plot of log (artifact density) versus site age (calBP). Best-fit regression line drawn though 59 camp/workshop, kill/processing, and special function sites (r 2 < .001; F = .004; df = 1, 57; p = .70).

Percent Local

Relative frequencies of local (sources <50 km away) versus non-local (sources > 50 km) raw material at sites were not useful at this scale comparison. As discussed above,

%Local is most influenced by distance a site is from a lithic raw material source. Sites associated near major lithic raw material sources generally have higher percentages of local raw material than sites far from major lithic sources. Larger scale patterning in these data is not apparent (Mann-Whitney U = 58.5, Z = -4.95; p = <.001). No statistically significant difference could be observed based on comparisons by physiographic region 203

(F = .92; df = 5,25; p = .48) or site function (F = .34; df = 2,28; p = .72) (Tables 5.6 and

5.7). Likewise, no significant trend through time was observed in the relative abundance of local versus non-local raw material. The lack of significant results from this study is at least partly a function of small sample size and high within-group variability. It is also possible that the large physiographic regions and/or distance scales used here are too coarse to show significant patterns.

Table 5.6. Summary Statistics for Raw Material (%Local) by Physiographic Region Std. Physiographic Region N Mean Range Dev. Colorado Piedmont/Raton Volcanic 8 43.9 41.8 0-100 Edwards Plateau 4 100 -- -- High Plains 19 31 44.4 0-100 Rocky Mountains/Intermountain 9 69.7 44.8 0-100 Basin Missouri Plateau 9 50.3 40.5 0-95 Rolling Hills 8 6.4 16.8 0-44.6 Note: Analysis conducted with Log 10 transformation of %Local values. No significant difference; F = 0.92, df = 5,25, p = .48

Table 5.7. Summary Statistics for Raw Material (%Local) by Site Type Std. Site Type N Mean Range Dev. Camping/Workshops 26 66 41.3 0-100 Kill/Processing 25 24 38.0 0-100 Special Function 6 14.3 19.3 .7-44.5 Note: Analysis conducted with Log 10 transformation of %Local values. No significant difference; F = 0.34, df = 2,28, p = .72

204

DISCUSSION

This section briefly summarizes the results of the analyses presented in this chapter and discusses important regional and temporal trends identified in this study.

Regional Trends

The results of this chapter suggest there are regional differences in how Paleoindians used different major physiographic regions. In addition, both the high-tech forager and low-mobility, low-tech forager Paleoindian land-use models describe how these foragers used different regions. For example, High Plains settings are dominated by a high proportion of kill/processing localities, which are characterized by low artifact densities and high proportions of non-local lithic artifacts. These characteristics largely match the expectations for the mobile, high-tech forager model. On the other hand, Edward Plateau and Rocky Mountain settings have a high proportion of camp/workshop localities with high artifact densities and a generally high proportion of local artifacts. These characteristics are largely what the low-mobility, low-tech forager model predicted.

As presented here, these results suggest a co-variance between megahabitat patch and land-use strategies. This is not to say that the huge physiographic regions used in this study represent uniform environments that prehistoric foragers used in a static manner.

Quite to the contrary, Paleoindian behavior was directed as a response to the local environmental conditions. Even large regions of similar habitats were composed of 205

diverse ecological patches. For example, environmentally diverse alluvial valleys and

canyons crossed the huge, highly uniform grasslands of the High Plains and Edwards

Plateau. Large intermountain valleys, dominated by grass and sage parklands, represented

ecologically uniform patches within the otherwise diverse Rocky Mountains.

These conclusions do not mean, as Kelly (1983:301) warns, that we should suppose a

one-to-one correspondence between environmental setting and specific land-use strategy

exists. Environmental changes that occur at the scale of year-to-year or even longer do

affect hunter-gatherer decision-making. We should expect mixed strategies for coping

with resource changes. Nevertheless, it does appear that Paleoindian foragers keyed onto

a few important resources within each major region and structured their use of the

landscape accordingly. When studying prehistoric land use it is important to understand

the nature and distribution of the important resource patches in different environments, and then analyze how Paleoindian foragers moved through those environments.

The upland portions of the High Plains, Rolling Hills, and Missouri Plateau were productive grasslands environments, but much of this biomass is not easily accessible to humans except through hunting large game (Kuhn and Stiner 2001). However, if

Paleoindian foragers wanted to prey on high-rank food sources they would have to be highly mobile in order to find their prey. Hunter-gatherers must practice high mobility (at least logistic mobility, if not residential mobility) in order to exploit large game in grasslands environments, because (1) these taxa have large foraging areas compared to smaller game, (2) the movement of game in these grassland environments are not confined by many natural boundaries, and (3) Paleoindian foragers had at least occasional 206 need to move out of grassland environments in search of other resources, such as lithic raw materials, wood, water, and other prey (Hofman 1999; Knell 2007).

In these grassland environments, even the locationally predictable resources, such as lithic raw material, wood, and water, were distributed in small, highly dispersed patches of variable quality and quantity. Faced with such an environment, it is not surprising that foragers would be highly mobile themselves and focus primarily upon the highest quality resource in this environment: large game. Such a strategy should cause foragers to (1) not stay in any one place very long or return to one particular patch very often (low occupation intensity), (2) rely on a highly portable technology made of raw material collected from somewhere else (low frequency of local raw material), and (3) have sites mostly focused on exploitation of large game (high frequency of kill sites). The available data from the grassland settings of the High Plains and Rolling Hills support these expectations.

Grassland environments within Colorado Piedmont, Edwards Plateau, Rocky

Mountains, and Northern Plains have comparable primary productivity as the High Plains and Rolling Hills, and thus likely contained numerous large game. However, these environments are located near patches of very high-quality lithic raw material often located in environmentally diverse settings where other valuable resources such as water and wood could be obtained in either nearby alluvial settings or adjacent upland environments. A similar dense packing of multiple important resources into small habitats could also be found in the major alluvial valleys, such as Platte, Canadian, Red, and Trinity rivers, that cross portions of the High Plains and Rolling Hills. 207

The greater ecological diversity and close spacing of key resources in these settings

meant that foragers would not have to travel very far to find valuable patches, but they

would have to be more thorough in their movement through these patches. The more

clustered nature of resources in these settings meant that the time spent searching for one

particular resource did not strand the forager in areas far from water, wood, or stone.

Such a resource structure would explain why these settings show a higher diversity of

activities (such as hunting large game, gathering plants and small game, and extracting

tool stone) than do grassland settings. In addition, foragers would be expected to have

higher occupational intensity in these environments, both in terms of spending more time at a particular patch and returning to a patch repeatedly. The pattern of greater evidence for habitation sites, greater artifact density, and potentially increased reliance on local raw material would be the expected outcome in this context.

Temporal Trends

This study was less successful in identifying robust temporal trends in the Paleoindian data. In fact, it may be that for the Paleoindian period the high degree of spatial variation in land-use strategies obscures our ability to identify long-term trends in the data.

The most significant temporal trend observed in this study is that the relative abundance of different site types changes slightly through time. A high proportion of

kill/processing localities characterize the oldest Paleoindian time period, while

camp/workshops are more common later in time. However, the differences in site 208

function through time are not related to long-term trends in site artifact density and local

raw material. Although statistically insignificant, site artifact density generally decreases

through time across most physiographic regions. In addition, no temporal trends were

observed in use of local versus exotic raw materials.

CONCLUSIONS

The results from this chapter point to the presence of regional differences in

Paleoindian land use, although it was not possible to identify clear temporal changes.

This supports prior observations that hunter-gatherer behavior responded to the

differences in resource availability and quality (e.g., Binford 1980, 1983; Kelly 1995;

MacDonell and Wandsnider 2003). However, due to the nature of this study, both in terms of the variables examined and how data grouped together, the patterns observed here only reflect very coarse-grained differences in Paleoindian land use.

I continue the landscape approach introduced here in the next two chapters. Chapters

6 and 7 use zooarchaeological data to evaluate spatial covariance between ecological parameters and use of fauna. It is hoped that by comparing results from different datasets and showing flexibility in how data is organized it will be possible to identify finer- grained temporal and spatial structure in Paleoindian land use and subsistence strategies. 209

CHAPTER 6: VARIATION IN PALEOINDIAN FAUNAL USE

The use of animal resources by Paleoindian hunter-gatherers has been a key research issue for archaeologists for nearly 80 years. In the last few decades, the debate over this issue has revolved around what I propose is a false dichotomy: specialized big game hunting versus broad-spectrum foraging (e.g., Amick, 1994; Waguespack and Surovell

2003; Cannon and Meltzer 2004). The hunting of large-bodied animals is argued to play a primary role in Paleoindian subsistence (Todd 1991; Bement 2003; Waguespack 2003,

2007; Waguespack and Surovell 2003), technology (Frison 1991), mobility and land-use strategies (Kelly and Todd 1988), and the success of the initial settlement of the

Americas (Surovell 2000; Waguespack and Surovell 2003). Some even suggest that human hunting was a primary cause for the extinction of 35 genera of Rancholabrean fauna (Alroy 1998, 2001; Martin 1984, 2005; Martin and Steadman 1999; Haynes 2002a,

2002b; Surovell et al. 2005).

This chapter does not wade into the murky pool that is the specialist-generalist debate. Instead, I hope to draw attention to important epistemological issues related to

Paleoindian faunal use, as well as the need to quantitatively analyze zooarchaeological patterning to obtain a more complete understanding of prehistoric life ways and land use.

Also, drawing on data from 69 Paleoindian assemblages in the Great Plains and Rocky

Mountains (Figure 6.1; Table 6.1), I weigh in on several debates in the literature, including the bias of kill sites, long-term changes in diet breadth, and exploitation of small game. 210

Figure 6.1. Sites with Paleoindian faunal remains used in this study. 211

Bement and Carter Jake Bluff Table 6.1. Sites with Paleoindian Faunal 2003 Remains. Kornfeld and Jerry Craig Site Reference Frison 2000 Jurgens Wheat 1979 12 Mile Creek Hill 2002b Hofman and Laird 48SW8842 Byers et al. 2005 Backmar 1997 48SW13156 Byers et al. 2005 Lake Theo Baxevanis 1997 48UT375 Smith et al. 2003 Lamb Spring McCartney 1983 Agate Basin Hill 2001 Lewisville Winkler 1982 Bamforth et al. Allen Lime Creek Jones 1999 2005 Linger Jodry 1999b Aubrey Ferring 2001 Lipscomb Todd et al. 1992 Big Black Williams 2000 Medicine Lodge Walker 1975 Bonfire Shelter Byerly et al. 2005 Creek Casper Frison 1974 Miami Sellards 1938 Cattle Guard Jodry 1999a Mill Iron Frison 1996 Cherokee Sewer Whittaker 1997 Mummy Cave Hughes 2003 Cannon and Clary Ranch Hill 2006 Myers-Hindman Frison and Todd Cannon 2004 Colby 1986 Olsen-Chubbuck Wheat 1972 Cooper Bement 1999 OV Clary Site Hill et al. 2006 Smith and McNees Domebo Leonhardy 1966 Rattlesnake Pass 1990 Finley Hill 2007 Red Smoke Hill 2007 Folsom Meltzer 2006 Rex Rodgers Hill 2006 Fulgham and Frasca Stanford 1982 Upper Twin Mountain Kornfeld et al. 1999 Hill and Hofman Frazier Borresen 2002 Waugh Byers 2002; Knell 1997 Hell Gap et al. 2002 Wilson Leonard Baker 1998 Heron Eden Corbeil 1995 Frison and Todd Horner 1987 212

First, some archaeologists argue that kill assemblages bias our inferences about

subsistence practices (Amick 1994; Meltzer 1988, 1993, 1995; Gero 1995; Cannon and

Meltzer 2004). These authors claim that by including data from kill sites, the contribution

of big game to the Paleoindian diet is overestimated. However, this claim is rarely

quantitatively evaluated so it is not often clear what effect this "bias" actually has. For

example, Paleoindian archaeologists do not often systematically compare the composition

of the zooarchaeological record from kill sites to those from camps. Furthermore, even if

there is a significant difference in the composition of these two site types, how does that

difference affect our understanding of regional variation in faunal use?

Second, others suggest Paleoindians may have altered their diet based on differences in the environmental setting being occupied and/or over time due to environmental change or human-induced resource depression (e.g., Greiser 1985; Bonnichsen et al.

1987; Hill 2001; Bamforth 2002a, 2002b; Byers 2002; Bousman et al. 2004; Cannon and

Meltzer 2004). These arguments presuppose that studies have adequately evaluated regional variation in Paleoindian faunal use, a fact that is probably debatable in most cases (c.f. Cannon and Meltzer 2004).

Finally, although small game is touted as an important resource for Paleoindian foragers, zooarchaeological analyses that explore this issue in ways other than species identification or presence/absence of these species are still comparatively rare (e.g.,

Amick 1994; Davis 1962; Johnson 1977, 1987; Waguespack 2003; Waguespack and

Surovell 2003). For several years, zooarchaeologists working in other time periods and 213 regions have analyzed small game remains to explore important topics, such as how prey behavior and reproductive characteristics or habitat differences influence human predator-prey interactions (e.g., Broughton 1994, 2002; Cannon 2000; Stiner 2001, 2005;

Stiner et al. 2000; Speth and Tchernov 2002; Butler and Campbell 2004; Munro 2004).

Applying these types of analytical strategies to data from Paleoindian faunal assemblages will likely provide insights into prehistoric diet breadth and subsistence not previously gained.

HISTORY OF RESEARCH

Almost immediately following the discoveries at the Folsom type-site, archaeologists began asking questions about what prey Paleoindian hunter-gatherers did or did not exploit (e.g., Barbour and Schultz 1932a; Roberts 1935). For most of the twentieth century Paleoindians were portrayed as specialized hunters of bison and mammoth, based on the numerous occasions where remains of these species were recovered in association with well-made fluted or lanceolate projectile points (Sellards 1952; Wormington 1957).

When evidence for small game was recovered at Paleoindian sites, it was often assumed that these resources were simply supplements to a diet otherwise rich in big game (Holder and Wilke 1949; Davis 1953, 1962).

Recent excavations, which use very careful excavation techniques and fine screening of back dirt, largely uphold the traditional view of the primacy of bison and mammoth remains in zooarchaeological assemblages. This is true for both kill sites (e.g., Hofman et 214

al. 1995; Collins et al. 1997; Hill and Hofman 1997; Hofman and Blackmar 1997;

Hofman and Todd 2001; Bement 1999; Kornfeld et al. 1999; Bement and Carter 2003;

Mandel and Hofman 2003) and camp localities (e.g., Jodry 1999a; Williams 2000;

Ferring 2001; Hill et al. 2006; c.f. Mallouf and Mandel 1997; Baker 1998; Smith et al.

2003; Byers et al. 2005).

Despite these findings, there are a growing number of archaeologists who argue that

the contribution of large game hunting to Paleoindian subsistence is exaggerated (e.g.,

Amick 1994; Johnson 1977; Bonnichsen et al. 1987; Cannon and Meltzer 2004).

According to these critics, empirical evidence and theoretical concepts raise questions

about whether such a high reliance on big game hunting is even possible for hunter-

gatherers. For example, these researchers note that several regions in North America,

such as the Northeast and Great Basin, have produced little direct evidence for the

exploitation of these types of fauna (Meltzer 1988; Cannon and Meltzer 2004). In

addition, they also question whether large game would have been abundant enough to

allow hunter-gatherers to exclude smaller game from their diet (Byers and Ugan 2005).

They suggest that the current perspective of Paleoindians as specialized big game hunters

is the result of biases introduced through the preferential preservation and archaeological

discovery of megafauna bone beds over accumulations of small-bodied animals (Grayson and Meltzer 2002; however, see contradictory conclusion by Hall et al. 2002). While site- based bone preservation bias is often thought to have biased evidence in favor of large game hunting, no one has claimed there are geographic biases to bone preservation within

the Great Plains and Rocky Mountains. 215

In summary, despite some excellent progress over the last decade in excavation of newly discovered sites (e.g., Hill and Hofman 1997; Bement 1999; Ferring 2001; Bement and Carter 2003; Hill et al. 2006) and the reanalysis of important curated zooarchaeological assemblages (e.g., Jones 1999; Hill 2001; Borreson 2002; Byers 2002;

Hill 2001, 2002a, 2002b, 2006; Hill and Hill 2002; Byerly et al. 2005; Meltzer 2006),

Paleoindian faunal studies have come to an impasse over what the diet of Paleoindians would have been. As a result, there has been little progress in either finding a consensus on this issue or generating a better understanding of the factors underlying the variation observed in the Paleoindian zooarchaeological record.

The current chapter tries to move beyond this impasse. By exploring geographic and diachronic variation in the human predation of differently sized game, it may be possible to begin to understand how Paleoindian occupation in different habitats influenced diet breadth and resource exploitation. Specifically, this chapter investigates the correspondence between the types of environmental settings occupied by Paleoindian foragers (grasslands, woodlands, alluvial valleys) and the predation strategies of those foragers (i.e., reliance on collecting small game versus hunting large game) (see Hill

2007). In this way, we may arrive at a clearer view of the balance of dietary opportunism versus specializations within Paleoindian subsistence adaptations.

216

PALEOENVIRONMENT, PREY ABUNDANCE, AND HUNTING PRESSURE

As outlined in chapter 2, understanding the pattern of the changing abundance of high-ranked prey over time is important for making inferences about Paleoindian faunal exploitation patterns. As I have proposed elsewhere (Hill 2007), the reorganization of terrestrial biotic communities at the Pleistocene-Holocene transition and the concurrent changes in human population demography likely had a strong influence on the abundance of fauna in the Great Plains and Rocky Mountains (FAUNMAP Working Group 1994;

Graham and Lundelius 1984). Below I summarize the current thinking concerning Late

Quaternary changes in large prey abundance.

During the Late Pleistocene, much of central and western North America was a broad expanse of grassland and parkland environments populated with a multitude of megafaunal species (Kurtén and Anderson 1980; Lundelius et al. 1983; Graham and

Mead 1987; Grayson 1993). Due to significant overlap in feeding strategies and likely high seasonal mortality, the absolute number of any particular species was probably fairly low (Guthrie 1984; Feranec and MacFadden 2000; Martin and Williams 2007). At the same time, human predation was probably relatively limited due to sparse human populations. Hypotheses of low population density during the early Paleoindian period are based on (1) the low density of early Paleoindian sites and diagnostic artifacts, (2) their extensive use of exotic raw material sources suggesting very high mobility, (3) analogies to the low population density of many modern hunter-gatherer groups, and (4) the high morphological variability among Paleoindian skeletal remains, suggesting very 217 small founding populations (e.g., Amick 1994; Barton et al. 2004; Fiedel 2002;

MacDonald 1998; Steele and Powell 2002; Wobst 1974; c.f. Martin 2005). Therefore, prior to the end of the Pleistocene, while foragers could expect to encounter a wide diversity of megafauna across most habitats, the absolute abundance of this prey class was low enough that it was unlikely that humans could exclude lower-ranked prey items from their diet (Byers and Ugan 2005).

The close of the Pleistocene and the extinction of most megafauna species wrought dramatic alterations on the landscape (Lundelius et al. 1983; Grayson and Meltzer 2002,

2003; Cannon 2004; c.f. Martin and Williams 2007). Bison was one of the few large land mammal species (e.g., moose, muskox) left in North America. However, these changes were not necessarily bad for human hunters. Released from much of their prior browsing competition and most predation pressure from large carnivores, bison populations seemed to undergo a rapid increase based on greater number of bison finds and significant decrease in body size (Chapter 4), albeit limited geographically to grasslands of the Great

Plains and adjacent areas (Guthrie 1984; MacDonald 1981; FAUNMAP Working Group

1994). This period also witnessed an expansion in populations of medium-sized artiodactyls, including deer, pronghorn, and elk, as they filled niches left open by the megafauna extinction (Martin and Williams 2007).

Some areas, such as the Southwest and California, previously were rich in bison, mammoths, and horses but became less productive hunting areas after the Pleistocene extinctions. Nowhere were bison so abundant that small and medium game could be totally ignored as a food source. Nonetheless, because human population density was 218 probably still quite low during this period and the there was an increasing abundance of bison in certain regions, especially the Great Plains, Paleoindian foragers probably found hunting large game rewarding at the end of the Pleistocene (MacDonald 1998; Amick

1996).

As the Holocene progressed, increases in mean temperature and monsoonal precipitation across the mid-continent of North America resulted in some important shifts in sediment deposition patterns (Holliday 1997; Olson and Porter 2002), the composition of biotic communities (Davis et al. 1986; Jodry 1999; Baker et al. 2000), and stream and pond hydrology (Neck 1995; Yansa 1998; Leonard and Reasoner 1999) that would have important consequences for large fauna abundance. These changes to the environment were slow to appear and their expressions varied across the landscape. Probably some time after approximately 9,500–9,000 14 C years BP bison population decreased to a point that would have affected human subsistence, at least in portions of the Rocky Mountains

(e.g., MacDonald 1981; Hughes 2003; Byers et al. 2005). Mass killing of bison was still possible at least periodically in the Great Plains, as seen at sites such as Heron Eden,

Horner, and Finley (Frison and Todd 1987; Corbeil 1995).

It was probably only during the Late Paleoindian period when foragers, especially in mountain settings, turned to intensively exploiting lower-ranked, lower risk resources

(e.g., plants and small game). During the Late Paleoindian period an increase in site density, an increased use of local lithic raw materials, and a diversification in artifact styles is used as evidence for increased occupational intensity and possibly the beginnings of regionalization of foraging populations, especially in alluvial valleys and 219

mountain settings (e.g., Husted 1969; Frison 1992; Bamforth et al. 2005; LaBelle 2005).

The combination effects of increased human population density and decreased mobility

likely would have depressed the size and structure of the faunal resource base that these

foragers relied upon.

METHODS FOR MEASURING DIET BREADTH AND INFERRING PREY RANK

Diet breadth can be measured as total number of food items in the diet, or greater proportional evenness among high- and low-ranked prey items in the diet, or both. This study uses multiple richness and evenness measures in the analysis of Paleoindian subsistence.

Prey rank is defined using two different classification systems. The first system groups prey according to package size and follows previously established body size categories (i.e., Klein 1976; Brain 1981): size class 1 (live body mass <22 kg; e.g., lagomorphs, birds, turtles), size class 2 (22–113 kg; e.g., bighorn sheep, pronghorn), size class 3 (113–340 kg; e.g., deer, elk), and size class 4 (>340 kg; e.g., bison, mammoth). In order to compare the relative abundance of the differently sized prey, an abundance index for each body size class was calculated, following procedures comparable to the previous work of Bayham (1979), Broughton (1994), and Szuter and Bayham (1989). The number of identifiable specimens (NISP) from each body size class was summed to calculate the proportional representation of the prey body size abundance indices (AI): large fauna index (Class 4), medium fauna index (Class 3 and 2) and small fauna index (Class 1). 220

Calculation of AI values performed by combining NISP counts for all taxa within

each body size category, and then comparing the contribution of each specific body size

class to the entire faunal assemblage. For example, the calculation of large fauna AI

follows the equation: Large Fauna AI = ∑ Size Class 4 NISP/( ∑ Size Class 1 NISP + ∑

Size Class 2 NISP + ∑ Size Class 3 NISP + ∑ Size Class 4 NISP). This resulting index generates values ranging between 0 and 1.0, with high values indicating greater representation of fauna of that size category.

The second classification system is a modified version of a scheme developed by

Stiner (Stiner et al. 2000; Stiner 2001, 2005), and groups the most abundant small- and medium-bodied prey species into three categories according to body size, predator avoidance strategy, and speed: (1) medium-sized ungulates, (2) small slow game, and (3) small, difficult-to-catch game (Table 6.2). The goal is to organize medium and small prey based on the difficulty or risk foragers face in harvesting these prey items. Although a few small- and medium-sized taxa in Paleoindian assemblages were not included in this classification system because they did not fit neatly into just one prey type category, none of these taxa were well represented. Excluding them does not influence the results presented here.

The Reciprocal of Simpson's Index (RSI) is used here as a measure of assemblage evenness. This index is calculated using the equation RSI = 1/ ∑(pi) 2, where p represents the proportion of each prey type for array i (Stiner and Munro 2002; adapted from

Simpson 1949; Levin 1968). As only three potential prey classes are used in this study,

the RSI ranges from 1 (least even) to 3 (most even). 221

Table 6.2. Summary of Life History, Predator Defense Strategy, and Capture Requirements for Common Medium and Small Prey in Paleoindian Assemblages Prey Reproductive Predator Defense Capture Classification Productivity Taxa Strategy Techniques A A deer, Rapid running, Medium bighorn freezing, slow Hunt/trap Ungulate sheep, charging pronghorn

Freezing, h iding, turtle, Small Slow slow Gather armor mussels

birds, Hiding, running, Small lagomorphs, subterranean difficult to fast Hunt/Trap rodents, escape, nocturnal catch prairie dog, activity beaver, fish A Classification and productivity systems used in this study are adapted from Stiner (2001, 2005; Stiner et al. 2000).

THE STUDY SAMPLE

Zooarchaeological information from 69 assemblages forms the basis of this study

(Table 6.3). Assemblages were selected from major habitat settings of the Great Plains

and Rocky Mountains (alluvial valley n = 16; foothill/mountain n = 30; plains/rolling

hills n = 23) and together span the period of approximately 11,500 to 7,200 14 C years before present. As already discussed in chapter 5, habitat settings used in this study chosen were because site occupants in these areas would have differential access to a

222

Table 6.3. Summary information for 69 Paleoindian zooarchaeological assemblages used in this chapter Small Medium Site Large AI AI %Medium %Small %Small Site Screen A NISP AI (Class Function B (Class 4) (Class Ungulate Slow Difficult 3/2) 1) Alluvial Valley Sites Allen-Occupation 2 N C 279 0.06 0.16 0.77 0.16 0.03 0.81 Allen-Intermediate Zone N C 473 0.15 0.24 0.60 0.28 0.02 0.70 Allen-Occupation 1 N C 1,763 0.47 0.07 0.45 0.12 0.10 0.78 Aubrey Y C 1,523 0.34 0.01 0.64 0.03 0.74 0.24 Big Black Y C 836 0.41 0.02 0.57 0.00 0.00 1.00 Cherokee Sewer- IIIa Y K 144 1.00 0.00 0.00 ------Frazier N K 1,188 0.99 0.00 0.01 0.25 0.00 0.75 Jurgens-Area 1 Y C 2,527 0.97 0.01 0.02 0.37 0.06 0.58 Jurgens-Area 2 Y C 155 0.61 0.37 0.03 0.93 0.00 0.07 Jurgens-Area 3 Y K 2,993 0.99 0.01 0.00 0.74 0.00 0.26 Lewisville Y C 466 0.05 0.04 0.92 0.04 0.04 0.92 Lime Creek-Zone I N C 564 0.26 0.22 0.52 0.14 0.03 0.83 Red Smoke-Strata 80-90 N C 6,165 0.55 0.35 0.10 0.60 0.07 0.33 Wilson Leonard-Units Isi, Isi/Icl, Igl/Isi, Isi/Isi- Y C 311 0.12 0.74 0.14 0.00 0.11 0.89 C, Isi/Icl/Isi-C Wilson Leonard-Unit II: Y C 606 0.01 0.66 0.33 0.01 0.10 0.89 Wilson Leonard-Upper Y C 430 0.00 0.75 0.24 0.06 0.29 0.65 Unit I Foothills/Mountain Sites 48SW8842 Y C 147 0.00 0.00 1.00 0.00 0.00 1.00 48SW13156 Y C 172 0.00 0.02 0.98 0.00 0.00 1.00 48UT375-Level 1 Y C 926 0.00 0.00 1.00 0.00 0.00 1.00 48UT375--Level 2 Y C 3,190 0.00 0.00 1.00 0.00 0.00 1.00 Cattle Guard Y C 3,517 1.00 0.00 0.00 ------Colby Y K 463 1.00 0.00 0.00 ------Finley Y K 8,326 1.00 0.00 0.00 ------Folsom N K 3,640 1.00 0.00 0.00 ------Hell Gap-Locality II Y C 1,318 0.99 0.01 0.00 1.00 0.00 0.00 (Agate Basin) Hell Gap-Locality V Y C 1,374 0.99 0.00 0.01 0.75 0.00 0.25 (Cody) Horner I Y C 2,157 0.99 0.01 0.00 0.81 0.19 0.00 Horner II Y K 7,097 1.00 0.00 0.00 1.00 0.00 0.00 Jerry Craig Y K 326 1.00 0.00 0.00 ------Linger Y C 462 1.00 0.00 0.00 ------Medicine Lodge Creek- Y C 65 0.03 0.35 0.62 0.37 0.00 0.63 Deep Deer Faunule Medicine Lodge Creek- Y C 3,880 0.00 0.01 0.99 0.01 0.00 0.99 Rodent Level Faunule Medicine Lodge Creek- Y C 95 0.00 0.28 0.72 0.09 0.00 0.91 Fire Pit Faunule 223

Table 6.3. Summary information for 69 Paleoindian zooarchaeological assemblages used in this chapter Small Medium Site Large AI AI %Medium %Small %Small Site Screen A NISP AI (Class Function B (Class 4) (Class Ungulate Slow Difficult 3/2) 1) Medicine Lodge Creek- Y C 52 0.02 0.48 0.50 0.49 0.00 0.51 Pryor Stemmed Faunule Mill Iron Y C 655 1.00 0.00 0.00 ------Mummy Cave-Layer 12 Y C 179 0.00 0.91 0.09 0.86 0.00 0.14 Mummy Cave-Layer 11 Y C 95 0.00 1.00 0.00 1.00 0.00 0.00 Mummy Cave-Layer 10 Y C 145 0.00 0.94 0.06 0.93 0.00 0.07 Mummy Cave-Layer 9 Y C 108 0.00 0.99 0.01 0.99 0.00 0.01 Mummy Cave-Layer 8 Y C 85 0.00 0.98 0.02 0.96 0.00 0.04 Mummy Cave-Layer 6 Y C 46 0.00 0.38 0.63 0.26 0.00 0.74 Mummy Cave-Layer 4 Y C 33 0.00 0.52 0.48 0.36 0.00 0.64 Mummy Cave-Layer 1 Y C 35 0.00 0.14 0.86 0.12 0.00 0.88 Myers-Hindman-Unit 1 Y C 37 0.38 0.62 0.00 1.00 0.00 0.00 Rattlesnake Pass Y K 424 0.99 0.01 0.00 0.80 0.00 0.20 Upper Twin Mountain Y K 222 1.00 0.00 0.00 ------Plains/Rolling Hills Sites 12 Mile Creek N K 871 1.00 0.00 0.00 ------Agate Basin-Area 2-AB Y K 3,179 1.00 0.00 0.00 ------Agate Basin-Area 2- Y C 1,330 0.78 0.22 0.00 1.00 0.00 0.00 Folsom Agate Basin-Area 3-Hell Y K 986 1.00 0.00 0.00 ------Gap Bonfire Shelter-Bone Y K 2,261 1.00 0.00 0.00 ------Bed 2 Casper Y K 7,146 1.00 0.00 0.00 ------Clary Ranch Y K 1,920 0.96 0.00 0.04 0.00 0.22 0.78 Cooper-Lower Kill Y K 900 1.00 0.00 0.00 ------Cooper-Middle Kill Y K 1,200 1.00 0.00 0.00 ------Cooper-Upper Kill Y K 1,450 1.00 0.00 0.00 ------Domebo N K 75 1.00 0.00 0.00 ------Frasca-Area I Y K 7,526 1.00 0.00 0.00 ------Heron Eden Y K 22,901 1.00 0.00 0.00 1.00 0.00 0.00 Jake Bluff Y K 167 1.00 0.00 0.00 ------Laird Y K 322 1.00 0.00 0.00 ------Lake Theo-Folsom Y C 6,919 1.00 0.00 0.00 ------Lamb Spring Y K 579 1.00 0.00 0.00 ------Lipscomb Y K 1,202 1.00 0.00 0.00 ------Miami N K 54 1.00 0.00 0.00 ------Olsen-Chubbuck N K 5,370 1.00 0.00 0.00 ------OV Clary Site-Middle Y C 1,033 0.97 0.00 0.03 0.00 0.39 0.61 Rex Rodgers Y K 1,886 1.00 0.00 0.00 ------Waugh Y C 1,403 1.00 0.00 0.00 ------A Y = Yes; N = No B C = camp locality; K = kill/processing locality

224

suite of important resources, as well as unique constraints (Kay 1998a; Kornfeld and

Osborn 2003). The foothill/mountain setting includes localities situated in (1) true alpine

settings, (2) high altitude intermountain grassland valleys (e.g., Middle Park and San Luis

Valley), and (3) foothill environments immediately adjacent to true mountain slopes.

Plains/rolling hills settings include both the High Plains proper and adjacent grassland

environments (e.g., Edwards Plateau, Rolling Hills of Oklahoma). Alluvial valley

habitats are restricted to very large, permanent river systems, sites must have been

located directly within a major drainage valley (or at least in a former floodplain terrace)

to be included in this group. Multifunction short-term camps (n = 40), as defined by lithic assemblages with numerous non-weaponry tools, debitage, and preforms in various stages of preparation, are slightly better represented than special-use kill/carcass- processing locales (n = 29), which are defined by having small chipped stone assemblages with a low diversity of tool types and a bias for weaponry, knives, and hide-

working tools.

Preservation and recovery quality were generally high for all assemblages. Systematic

matrix screening was practiced at more than 84% (n = 58) of the assemblages, although

the proportion of unscreened components was biased against alluvial (n = 6; 38%) and

plains/rolling hills (n = 4; 17%) settings. Just a single assemblage (3%) from

foothill/mountain settings was not screened. I have previously demonstrated that

screening treatment does not significantly influence outcome of the analyses of these

zooarchaeological assemblages; analysis outcome is similar regardless whether the 225

analysis examines only sites that had back dirt screened or all sites regardless of

screening treatment (Hill 2007).

A total of 130,365 specimens (NISP), of which more than 113,000 (87%) were

identified to genus or finer taxonomic distinction, are used in this study. Following the

procedure outlined by Grayson and Meltzer (2002) and Cannon and Meltzer (2004:1959–

1960) only taxa showing unequivocal evidence for human butchery and consumption

were included in this study. The criteria to determine whether a taxon was used and

consumed at a particular site include (1) presence of unambiguous butchery marks (e.g.,

cut marks or impact cones); (2) evidence of intentional burning or cooking; (3)

unquestionable and direct spatial association of projectile points and butchery tools with

faunal remains; and (4) association of faunal remains with constructed cultural features.

This study excluded any remains suggested to be of non-cultural origin by the original

analyst.

As recognized previously (Amick 1994; Hill 2007; Waguespack and Surovell 2003),

although more than 33 fauna taxa are represented in Paleoindian assemblages, just a few species account for the vast majority of faunal remains recovered (Table 6.4). The six most abundant species (bison, pronghorn, hare/rabbit, turtle, mammoth, and bighorn sheep) account for over 99% (n = 112,571) of identifiable remains. This bias for such a small suite of fauna is partially explained by the abundance of multi-animal bison kills and processing sites included in the database. However, even if kill sites are not considered, Paleoindian assemblages are still dominated by these same taxa (Hill 2007).

226

Table 6.4. Assemblage ubiquity, total NISP, and NISP/site for identifiable fauna reported from the sites listed in Table 6.1. Number of Common Name NISP NISP/Site Components (%) Bison 50 (83%) 108,230 2164.6 Pronghorn 16 (27%) 1,323 82.7 Hare/Rabbit 23 (38%) 1,310 57.0 Turtle 14 (23%) 693 49.5 Mammoth 3 (5%) 592 197.3 Bighorn Sheep 12 (20%) 423 35.3 Deer 23 (38%) 284 12.3 Prairie Dog 4 (7%) 254 63.5 Bird 21 (35%) 247 11.8 Beaver 11 (18%) 189 17.2 Snake 8 (13%) 182 22.8 Fish 10 (17%) 130 13 Canids 10 (17%) 114 11.4 Mussel/Mollusk 1 (2%) 87 87 Amphibian 5 (8%) 45 9 Badger/Weasel 6 (10%) 38 6.3 Elk 5 (8%) 24 4.8 Bear 3 (5%) 23 7.7 Horse 1 (2%) 13 13 Marmot/Woodchuck 4 (7%) 11 2.8 Porcupine 3 (5%) 7 2.3 Wolf 3 (5%) 6 2 Muskrat 2 (3%) 4 2 Raccoon 1 (2%) 4 4 Fox 2 (3%) 4 2 Lizard 2 (3%) 3 1.5 Moose 1 (2%) 2 2 Skunk 1 (2%) 2 2 Coyote 1 (2%) 2 2 Alligator 1 (2%) 1 1 Lynx/Bobcat 1 (2%) 1 1 Total 114,248

The conclusion drawn here is that while a broad spectrum of prey entered their diet,

Paleoindian foragers relied on only a small number of prey species for their primary subsistence. This first look at the data suggests that Paleoindian faunal assemblages tend 227 to be very uneven and their diets narrow overall. Below I attempt to explore potentially important spatial and temporal patterning in diversity and evenness in the Paleoindian diet. These analyses are conducted in the hope of identify factors underlying this patterning.

TRENDS IN PALEOINDIAN FAUNAL ASSEMBLAGES

The following sections explore several dimensions of the Paleoindian zooarchaeological record. The first section considers whether including kill assemblages in analyses influences our perspective on the importance of megafauna hunting to

Paleoindian subsistence. The second explores patterning in the use of prey species of different body sizes over time. Finally, evidence for resource intensification during the

Paleoindian period is discussed.

Large Game Abundance and Bias from Kill Sites

As noted above, bison remains from kill/processing localities represent a majority of the total fauna recovered from Paleoindian assemblages. But does this fact mean that kill site assemblages obscure the realities of Paleoindian subsistence? Figure 6.2 compares the distribution of large fauna AI values for kill/processing localities to those from camps. A simple comparison of median AI values shows that big game is much less abundant at camps (0.1; n = 40) than at kills (1.0; n = 29), and non-parametric analysis 228

indicates these differences are statistically significant (Kolmogorov-Smirnov D=.73; χ2 =

35.9; p<.0001).

Figure 6.2. Box plot comparison of large fauna (size class 4) abundance index (AI) between specialized kill/processing localities and multifunction short-term camps. Kolmogorov-Smirnov two-sample test (D = .731; χ2 = 35.9; p < .0001) rejects null hypothesis that two samples have the same distributions.

This result is not surprising, because mass death assemblages, such as those that make up the bulk of Paleoindian kills, should be dominated by the remains of a single species; whereas foragers are expected to transport a variety of prey types back to camp localities.

However, this is not to suggest that bison remains are absent from all camps or that the faunal assembles from camps in different habitats are similar to each other. For example, the faunal assemblages from the camps at Agate Basin-Area 2 (Folsom), Cattle Guard,

Horner I, Hell Gap-Localities II (Agate Basin) and V (Cody), Jurgens-Area 1, Linger,

Lake Theo, Mill Iron, OV Clary, and Waugh consist almost exclusively of bison remains.

In addition, large fauna AI values for kill assemblages are significantly higher than those 229 from camps in alluvial (Kolmogorov-Smirnov D = 1.0; χ2 = 9.80; p = .02) and foothill/mountain settings (Kolmogorov-Smirnov D = .78; χ2 = 13.2; p = .003) but not significantly higher than AI values of plains/rolling hill settings (Kolmogorov-Smirnov D

= .15; χ2 = .28; p > .99).

A model II two-way Analysis of Variance test of the effect of both site setting and site function on large fauna AI suggests that there is a significant interaction between site function and setting (F = 7.01; df = 2, 63; p = .002). In addition, based on this analysis it appears that site function explains more of the variance than setting. This seems to indicate that large fauna AI values between camps and kill/processes are significantly different in foothill/mountain and alluvial valley assemblages, but not in plains/rolling hills assemblages.

Based on these results, it seems reasonable to assume that assemblages from kills and processing sites could distort our views of the Paleoindian diet. In reality, non-parametric analyses indicate that the distributions of large fauna AI values by habitat do not significantly differ between samples that include both kills and camps and samples with just camps (Figure 6.3). Whether kills are included or not, the range of large fauna AI values for foothill/mountain and alluvial valley sites is wide, but median AI values remain consistently low. Plains/rolling hill assemblages, on the other hand, consist almost exclusively of remains of bison and mammoth. In other words, the habitat in which sites are situated had a strong influence on faunal diversity.

230

Figure 6.3. Box plot comparison of large fauna (size class 4) abundance index (AI) among alluvial valley, foothill/mountains, and plains/rolling hills with both (A) kill/processing localities and multifunction short-term camps, and (B) multifunction short-term camps only. Kolmogorov-Smirnov two-sample tests do not reject the null hypothesis that the distributions of large fauna AI values for sites in alluvial valley (D = .188; χ2 = 1.009; p > .99), foothill/mountain (D = .183; χ2 = 1.74; p = .84), and plains/rolling hills (D = .120; χ2 = .195; p > .99) are the same between samples that included both kills and camps and samples that included only camps.

231

These results suggest that kill sites are not "biasing" our perspective of Paleoindian subsistence. Rather, foragers appear to be targeting different types of prey depending on which habitat setting they occupy. In the more diverse environments of the alluvial valleys and foothill/mountains, foragers likely exploited the largest prey types whenever possible, but they also targeted a wide range of smaller faunal resources as well. In contrast, in the lower-diversity grassland environments, Paleoindians focused almost exclusively on hunting bison and mammoth and appeared to make very little use of small game.

Temporal Trends in Body Size of Prey Used

Based on the environmental history of the project area outlined above, I predict that the abundance of bison (and mammoth) taken by Paleoindian hunters should decrease through time as a result of increasing aridity and changes to the biotic communities during the early Holocene, and possibly as a result of increasing hunting pressure. I previously discussed some important regional differences in the extent to which foragers exploited prey of different body sizes through time (Hill 2007). This work showed that plains/rolling hills camp assemblages have very high proportions of large game for all time periods. In alluvial assemblages large fauna AI values decrease through time, while both the medium and small fauna AI ratios increase, although none of these trends were statistically significant. In contrast, foothill/mountain assemblages show a long-term pattern for significant decrease in large fauna AI values and a corresponding increase in 232 the medium fauna ratio values. Small fauna ratios in foothill/mountain assemblages showed no significant trend. As discussed in Hill (2007:430), these results are not a function of the size of the faunal sample, as measured by NISP.

While trends in AI values for different body size classes are informative for tracking how prey species of different sizes were used through time, it may ignore important shifts in the overall composition of diets. As a result, here I explore the degree of evenness through the use of the reciprocal of Simpson's Index calculated based on the three body size categories (small, medium, and large). Only camp assemblages are considered here because, unlike kills, they provide data on a broad spectrum of prey items. Also, as the sample of plains/rolling hill camps is very small, only results from alluvial valley and foothill/mountain camps are discussed.

Figure 6.4 shows that foothill/mountain camp assemblages generally yield low levels of evenness in diet breadth. Variation in the values for reciprocal of Simpson’s Index is not correlated with time (Spearman’s rho = .02; p = .94; n = 23). Interestingly, there was a significant inverse correlation between the reciprocal of Simpson's Index and sample size (Spearman’s rho = -.69; p < .001; n = 23). These results indicate that for most foothill/mountain assemblages a single body size class dominates. In early Paleoindian assemblages that prey is typically bison, but later assemblages consist of either small or medium fauna. In only a handful of foothill/mountain assemblages are more even mixtures of body sizes represented. Assemblages with a mixture of body sizes consist of only a few specimens, so these results may not be representative of human predation practices because of sampling biases. 233

These findings suggest that most foothill/mountain sites were not residential camps in

which broad suites of resources were pooled. Instead, these assemblages are probably the

end result of specialized collection of one or two prey types. Assemblages dominated by

medium-size ungulates probably represent seasonal hunting camps and usually consist of

bighorn sheep and deer (e.g., Hughes 2003). Assemblages dominated by small game are probably short-term resident localities where lagomorphs or various rodents were important resources (e.g., Walker 1975).

Figure 6.4. Evenness in the representation of three prey categories (large-, medium-, and small- bodied fauna) using the Reciprocal of Simpson's Index (3 = most even) versus radiocarbon age. Reciprocals of Simpson's Index from alluvial camp assemblages are significantly higher than those from foothill/mountains (Kolmogorov-Smirnov two- sample D = .609; χ2 = 12.309; p = .004).

Up to this point these findings are based on component-level analyses with multiple sites grouped according to habitat. Unfortunately, combining data from various sites in 234

different settings provides only a coarse resolution to the patterns of faunal use. This is

largely because environmental conditions around each site are unique, thus complicating

inter-site comparisons. To evaluate the evidence on a finer resolution, I will now compare

changes in evenness values within a single site through time. As seen at the stratified sites

of Mummy Cave and Medicine Lodge Creek, over time foragers’ diet breadth shifted

back and forth from narrow to broad (Figure 6.5).

Figure 6.5. Evenness in the representation of three prey categories (large-, medium-, and small-bodied fauna) using the Reciprocals of Simpson's Index (3 = most even) versus radiocarbon age for the foothill/mountain multifunction short-term camps of Medicine Lodge Creek and Mummy Cave.

Evenness values from alluvial camp assemblages are significantly higher than those

from foothill/mountains (Kolmogorov-Smirnov two-sample tests D = .609; χ2 = 12.309; p = .004), and show a slight, although insignificant, trend for increasing values through time (Spearman’s rho = .07, p = .80; n = 13). For alluvial camps, the reciprocal of

Simpson's Index does not correlate with sample size (Spearman’s rho = .31; p = .29; n = 235

13). Unlike in foothill/mountain settings, Paleoindian foragers occupying alluvial habitats preyed on a more equal mixture of different-sized game animals. This implies that alluvial camps were residential bases from which harvesting of a mix of fauna occurred and to which faunal resources were brought back. The poor temporal resolution at most sites makes it difficult to determine whether these different prey were collected concurrently or sequentially. The two stratified sites of Allen and Wilson-Leonard show a high degree of evenness through time and there does not appear to be any long-term trend

(Figure 6.6).

Figure 6.6. Evenness in the representation of three prey categories based on body size (large-, medium-, and small-bodied fauna) using the Reciprocal of Simpson's Index (3 = most even) by radiocarbon age for the foothill/mountain multifunction short-term camps of Allen and Wilson Leonard.

236

Use of Small- and Medium-Size Game and Evidence for Resource Depression

So far my analyses strongly suggest that there are significant regional differences in

Paleoindian prey exploitation patterns and, at least for alluvial valley sites, there is some evidence for dietary expansion through time. However, the factor(s) driving these patterns (such as environmental differences or human-induced resource depression) are not clear. Shifts in foragers’ use of different body size classes (which largely follow

Linnean taxonomic categories) may also be explained by climate-driven environmental changes or geographic variations in biotic communities (Stiner 2001; Wolverton 2005).

To track significant changes in the intensity of resource exploitation, Stiner suggests that an abundance of small-bodied, low-turnover prey use can serve as an indicator of threshold effects in predator-prey systems (Stiner et al. 1999, 2000; Stiner 2003). The greater representation of high- versus low-turnover prey with time suggests increasing harvesting pressure.

To explore whether similar trends are seen in the Paleoindian record, the most common small- and medium-sized faunal species were regrouped based on body size and predator avoidance strategies. A triangle graph depicting the proportions of these different prey categories reveals some interesting differences in exploitation of small and medium prey between habitat settings (Figure 6.7).

First, there are very few sites with evidence for use of small, slow game. Except for the Horner I assemblage, none of the foothill/mountain assemblages show evidence for predation on turtles or mollusks. Alluvial assemblages, on the other hand, generally show 237

Figure 6.7. Triangular graph of proportional representation of three prey categories based on forager harvesting strategies and prey body size and predator avoidance strategies for medium ungulate, small slow, small difficult-to-catch taxa for alluvial valley and foothill/mountain multifunction short-term camps. The presence of large fauna was excluded from consideration for this graph (MLC=Medicine Lodge Creek).

238 at least a minimal presence of these gatherable prey types. Allen-OC1, Red Smoke, and several levels at Wilson-Leonard show moderate use of these types of prey, but small, slow game only dominate the zooarchaeological assemblage at Aubrey. The paucity of these prey in foothill/mountain sites is probably a function of the fact that both turtle and mollusk populations preferentially inhabit lower-altitude riparian environments.

Second, because there is a paucity of small, slow fauna in these assemblages, the primary variation in faunal use occurs along the simpler continuum of sites dominated by large and medium prey and those dominated by small, difficult-to-catch prey. Sites distributed at the top of the graph (i.e., those with a high proportions of medium-size ungulate remains) all appear to be assemblages resulting from fairly efficient collection strategies. In all cases, these sites represent camps where the medium-size game was either the primary prey taken or a secondary resource acquired along with bison. The faunal assemblages from Myers-Hindman and a number of strata from Mummy Cave were dominated by bighorn sheep remains, which is not surprising given that both sites are situated in areas of rugged terrain, the preferred habitat for bighorn sheep (Shackleton et al. 1999).

At the other sites scattered at the top of this graph (Hell Gap-Localities II and V;

Horner I; Red Smoke, Rattlesnake Pass, and Jurgens-Area 2), the economic focus was apparently bison. The lower proportion of pronghorn and deer remains at these sites suggests that they were of secondary importance. Perhaps medium-sized ungulates were available in the same settings where the hunters searched for large game, or less likely, 239 these taxa were mixed in with bison being targeted and were killed as part of the larger kill event.

On the other hand, sites showing evidence for frequent harvesting of small game

(e.g., lagomorphs and rodents) likely represent a fundamentally different foraging approach, which required different strategies and equipment than those utilized in the capture of large- and medium-sized game (Frison 1991, 2004). For sites in the lower left corner of Figure 6.7, large- and medium-sized prey make up a relatively small proportion of the assemblage, if they are present at all. The faunal collections of these sites instead appear to result from systematic and intentional collection of low-ranked resources

(Walker 1975; Hughes 2003).

Greater dependence on small, difficult-to-catch prey likely developed because higher- ranked resources were not available in certain settings, or stores of large game remains were exhausted. These types of small game were probably available in the immediate vicinity of the sites, so collection probably occurred opportunistically by small foraging task-groups that were gathering other food and non-food resources (see Amick 1994).

The level of small game use represented at these sites, while substantial, was not at all comparable in scale to the systematic mass collection of rabbits, for example, seen at Late

Prehistoric Great Basin sites (e.g., Madsen 1993; Ugan 2005b)

Overall, the results presented here indicate a pattern for at least limited (local) human- induced resource intensification during the Paleoindian period (Jones 1999; Bamforth

2002a, 2002b). Support for this hypothesis comes from the fact that the most common types of small prey taken are not those that are slow reproducing (low turnover) or easily 240 caught types (e.g., turtle and shellfish), but instead are the quick reproducing (high turnover), difficult-to-catch types (e.g., birds, rabbits/hares, and subterranean rodents).

If harvesting pressure on small-bodied resources was low throughout the Paleoindian period, one would expect steady use of the easier-to-harvest small-bodied prey through time. Figure 6.8 shows that at least for alluvial valley sites this is not the case. Easily harvested small prey types are better represented in early Paleoindian assemblages than in later assemblages, where they are rare. The abundance of difficult-to-harvest small prey, on the other hand, generally increases through time. It is perhaps significant that the only locality where turtle remains are very abundant is the 11,500-year-old Aubrey site, which is the oldest site in the project area and one likely occupied at a time when human population density, and therefore also harvesting pressure, was at its lowest level.

While these patterns are suggestive, it must be acknowledged that the strength and direction of these trends are strongly influenced by just a few sites. For example, if the

Aubrey site is dropped from consideration, the slope of decline in the use of small, slow prey over time flattens significantly. Likewise, omission of the Red Smoke and Aubrey sites, which have a low frequency of small, difficult-to-catch prey, produces a graph with no obvious trend in the use of this prey group. The results of this study cannot prove that human hunting caused resource depression, but only suggest this may have been the case.

It is clear that more quantitative studies of small animal use are needed to clarify these findings.

241

Figure 6.8. Plot of (A) % slow small fauna and (B) % difficult small fauna versus radiocarbon age for multifunction short-term camp assemblages in alluvial valley habitats with best-fit regression lines.

242

DISCUSSION

This chapter explored several aspects of the nature of the Paleoindian

zooarchaeological record through a quantitative analysis of data from multiple sites.

Specifically, it evaluated whether bias introduced through collection strategies or site

selection influences our view of the role big game hunting played in Paleoindian

subsistence. It also explored how prehistoric diets varied through time and across space in

response to environmental differences or abundance of bison and mammoth populations.

The trends observed here match the predictions of the large prey abundance model

outlined earlier in this chapter. In addition, site location is a strong determining factor in

use of faunal resources (Hill 2007). As expected, early Paleoindian foragers focused

mostly on hunting bison and mammoths irrespective of the habitat occupied. Later in

time, when mammoths were extinct and bison became regionally more scarce, large game

hunting became less common, at least in foothill/mountain settings, although it continued

at moderate levels in alluvial valley settings and at very high levels in plains/rolling hills

environments. With the decrease in abundance of bison in foothill/mountain settings,

foragers shifted their attention to medium-sized ungulates and smaller game. While deer

and pronghorn were also exploited in plains/rolling hills and alluvial valley settings, they

appear to have been supplemental resources to bison.

With the resolution of available information I cannot determine whether Paleoindian foragers settled in alluvial valleys and foothill/mountain settings because these places provided a more diverse choice of foodstuffs and other resources than plains/rolling hills, 243 or if Paleoindian diet became diversified as occupation duration of these habitats increased. Nevertheless, it is possible to speculate about causes of the covariance between

Paleoindian diet and regional differences in Paleoindian land use, resource distribution, and group mobility.

The high reliance on large game in plains/rolling hill habitats is likely a response to the abundance of these taxa in this habitat. Also, because hunters must travel great distances to find their prey in these settings, there is a relationship between search time and prey choice. When search time increases or prey must be pursued over a long distance, foragers shift their criteria concerning which resource to exploit. The farther foragers travel, the more they preferentially focus on large game (Bettinger et al. 1997;

Schoener 1971). In addition, the farther foragers travel to find game the less likely they are to bring back evidence of the small game they do exploit.

This chapter and the previous one both suggest that Paleoindians spent more time and used a greater diversity of resources in alluvial valleys and foothill/mountain environments (Bamforth et al. 2005; Surovell 2003). It must be acknowledged that archaeologists do not really understand why groups decided to remain longer in these settings. Prehistoric people in these environments certainly spent time performing a variety of activities related to tool production, food preparation and processing, habitation. These locations would also encourage greater duration of use because of the availability of protected settings, wood, and potable water.

Even if we are unable to understand exactly why Paleoindian foragers were less mobile in these alluvial valleys and foothills/mountains, longer periods of habitat 244

residence are associated with greater reliance on alternative game types. Greater duration

of time spent in these habitats may have resulted in more resources being collected from

these areas (Knell 2007). The continued exploitation of these resources patches

eventually would have resulted in a decrease in the relative abundance of key resources

(e.g., Schoener 1971; Charnov et al. 1976; Cannon 2000). This explanation has been used

by other researchers to explain evidence for presumed intensification of resource

exploitation during the Paleoindian period (Jones 1999; Bamforth 2002a; Borresen 2002).

The results of this study suggest a general trend for Late Paleoindian period sites in

alluvial valleys, and to a lesser extent in foothill/mountain settings, toward an expansion

of diet in terms of number and type of prey items. Part of this decrease in foraging

efficiency was a shift from nearly exclusive use of large prey items to mixed use of large

and small prey items, all of which were costly to harvest. Unfortunately, the results of

this study cannot reveal meaningful trends on this issue. What is clear, is that when

Paleoindian foragers did not utilize large body game, they relied on fauna with high

reproductive rates and that were generally harder to catch. This suggests that either due to

environmental factors or previous hunting pressure, or both, the easiest to catch small game was not readily available for exploitation.

Up to this point discussions have been limited to patterning in the frequency of bones and teeth from differently-sized prey without mention of the quantity of kilograms of

meat, number of calories, or types of vitamins those skeletal remains represent. While it

is beyond the scope of this chapter to discuss the dietary contribution of different prey

taxa for all the sites studied, it is worthwhile to briefly compare how our perspective of 245

the contribution of different prey categories might change if our unit of analysis had been

useable meat represented instead of NISP values.

Table 6.5 presents the relative abundance of seven common prey taxa (bison,

pronghorn, deer, bighorn sheep, rabbit/hare, prairie dog, and box turtle) as measured by

NISP and the hypothetical mass (kg) of useable meat based on estimates of White (1953)

represented at three sites (Lime Creek-Zone I, Medicine Creek-Deep Deer Faunule, and

Medicine Creek-Pryor Stemmed Faunule). These sites are appropriate to use for this

analysis because they are some of the few sites considered here which have good

evidence for a mixed use of large, medium, and small game. The technique used to

calculate useable meat was developed by White (1953) and determines the amount of

meat provided by any given species based on the number of individuals represented at the

site (MNI) multiplied by the number of pounds of meat that one individual of that species would yield upon butchery. The meat yield is estimated based on a proportion, in this case 50%, of the total body weight of that animal.

These data are presented not as estimates of the literal amount of meat available to foragers at these sites, but only to emphasize that there are significant differences in foraging efficiency associated with preying upon large-bodied game versus small or medium game. Based on the White (1953) useable meat estimates, for a forager to obtain an equivalent amount of meat from killing one bison he would have to kill approximately

7 deer, 12 pronghorn, 270 rabbits/hares, or 430 prairie dogs. As seen below, the significance of this fact is that where large game is poorly represented in terms of NISP, it is possible that those taxa may still have provided the site residents with most of 246

Table 6.5. Comparison of relative taxonomic abundance estimates based on estimated mass of meat as calculated from site minimum number of individual (MNI) estimates versus number of individual specimens (NISP) for three Paleoindian sites. Estimated Mass of Meat % Useable Meat Taxon MNI represented, based on NISP %NISP Mass MNI (lbs) A Lime Creek-Zone I Bison 3 1200 91.2% 47 43.1% Pronghorn 2 110 8.4% 45 41.2% Rabbit/hare 1 2.4 0.2% 4 3.7% Box Turtle 1 0.7 0.1% 10 9.2% Bighorn sheep Deer Prairie Dog 2 3 0.2% 7 6.4 Total 1316.1 113 Medicine Creek-Deep Deer Bison 1 400 57.0% 2 6.90% Pronghorn Rabbit/hare 1 2.4 0.3% 4 13.79% Box Turtle Bighorn sheep Deer 3 300 42.7% 23 79.31% Prairie Dog Total 702.4 29 Medicine Creek-Pryor Stemmed Bison 1 650 49.9% 1 3.13% Pronghorn Rabbit/hare 1 2.4 0.3% 7 21.88% Box Turtle Bighorn sheep 3 300 37.4% 18 56.25% Deer 1 100 12.5% 6 18.75% Prairie Dog Total 802.4 32 A White (1953: Table 14) estimates of pounds of useable meats are bison (400 lbs, based on mature female), pronghorn (55 lbs), hares and rabbits (2.4 lbs, average of hare and rabbit), bighorn sheep (100 lbs), deer (100 lbs, mule deer), and prairie dog (1.5 lbs). A box turtle is estimated to provide .7 lbs of useable meat.

247 their food. In contrast, the numeric abundance of small game remains at many sites probably indicates that those taxa supplied foragers with only a modest amount of food.

CONCLUSIONS

Three major conclusions are drawn from the results of this study:

• Site function has a significant effect on the composition of Paleoindian faunal

assemblages. Kill assemblages are generally dominated by a single large game

taxon, while camp assemblages contain a larger diversity of prey. However, the

abundance of large game remains does not significantly differ between samples of

both kills and camps and samples of just camps, when examined by habitat setting.

• Prey use changes based on habitat setting. Paleoindian foragers targeted different

prey types depending on which habitat setting they occupied. In the more diverse

environments of the alluvial valleys and foothill/mountains, foragers exploited the

biggest prey whenever possible, but they also targeted smaller animals. In the more

uniform grassland environments, Paleoindians focused almost exclusively on

hunting bison and mammoth.

• There is limited evidence for increasing resource intensification, as measured by

increased exploitation of smaller game and therefore decreased foraging efficiency, 248 during the latter part of the Paleoindian period. Evidence for these changes is most apparent in alluvial valley and foothill/mountain settings. The differences likely relate to prehistoric land-use patterns and environmental conditions that affected abundance of big game on a larger scale.

249

CHAPTER 7: BISON KILLS: BONE BED COMPOSITION, SEASONAL

VARIATION, AND LONG-TERM TRENDS

As discussed in chapter 3, data on modern bison biology and behavior can be used to develop predictions concerning the size and composition of Paleoindian bison kills and the preferred season for bison hunting. These predictions were outlined in chapter 3, and are summarized here in Table 7.1. In this chapter, I evaluate how these predictions correspond to the zooarchaeological record, using data from 87 assemblages (Figure 7.1).

Table 7.1. Summary of key predictions related to the Paleoindian bison kills composition and seasonality, as developed in chapter 3. Topic Prediction • For mass kills, cow/calf herds will be preferred over bull herds. • For kills of isolated animals or small numbers of animals, Bone bed bulls will be preferred prey over cows or calves. Composition • The size of bison kills should vary according to seasonal and Size changes in bison herds. Therefore, late winter/early spring kills should be the smallest, late summer through early fall should be largest, and late fall though early winter should be intermediate in size. • Mass kills of cow/calves herds can occur at any time of the year. • Kills will be most frequent during late summer to winter. Seasonality • Kills will be least frequent during spring to early summer. • Kills of bulls and bull herds will be most common during spring and early summer.

Because the predictions relate to kill size, season of mortality, and sex distribution, I focus on these variables in my zooarchaeological analysis. I ask three primary questions:

In which season(s) did most bison hunting occur? Does seasonality influence the size of a 250

Figure 7.1 Map of the Great Plains and Rocky Mountains with sites mentioned in this chapter. 251 kill? Does the representation of male and female bison in an assemblage change according to the season of the kill?

PREVIOUS INTERPRETATION

For years archaeologists used analogies drawn from historic or Late Prehistoric bison hunters of the Great Plains to interpret Paleoindian bison procurement strategies (e.g.,

Bamforth 1988; Frison 1974; Moss 1951; Wheat 1972). Accounts of bison hunting by historic Plains Native American groups, in particular, provide a rich record for understanding strategies and techniques used to procure bison (e.g., Branch 1929; Catlin

1965; Gilmore 1931; Schoolcraft 1856; Weekes 1948; Wissler 1920). These accounts suggest that bison kills most commonly occurred during the late summer or early fall. In some historic-period bison hunts, a number of horse-mounted hunters surrounded a large herd of bison and drove them into some constructed enclosure where the kill itself was actually made. In other cases, once the hunters formed a tight circle around the herd, they began to shoot the bison with arrows from their mounts. The huge windfall of meat and fat procured by one of these mass kills provided rich food stores that hunters could use throughout the winter. Conducting these hunts during the late summer/early fall also took advantage of the large size of the herds during this time year, and corresponded to when many of the bison were in peak physical condition.

More recently, Frison (1982) and others (Bamforth 1988; Todd 1987; Todd et al.

1992) pointed out these accounts of historic and Late Prehistoric procurement patterns do 252

not have great explanatory value when it comes to earlier bison hunting for several

reasons. First, seasonality studies suggest that Paleoindian kills occurred in all four

seasons, meaning that bison were hunted even when they were not in the best physical

condition (Todd et al. 1992, 1996). Second, there are important differences in the type of

features and artifacts at Late Prehistoric sites and Paleoindian sites (Frison 1982:200).

Late Prehistoric groups utilized well-made hearths for intensive bone grease production.

Paleoindian-age hearths, such as those at Agate Basin-Folsom level (Hill 2001),

Stewart’s Cattle Guard (Jodry and Stanford 1992), Milnesand (Hill 2002a), Hanson site

(Frison 1978), and Waugh site (Hofman 1995), are shallow and not associated with

thermal rocks, and would not have been useful in grease production (Jodry and Stanford

1992; LaBelle 2005). Finally, there is no evidence for the intensive bone breakage

associated with bone grease extraction in Paleoindian faunal assemblages. Todd (1987,

1991; Todd et al. 1990) and others (Byers 2002; Hill 2001; Jodry and Stanford 1992)

have shown that although marrow extraction occurred at Paleoindian sites, such as

Stewart’s Cattle Guard, Casper, Clary Ranch, Agate Basin, Hell Gap, and Jones-Miller,

the degree and nature of this bone breakage is significantly different and less intensive

than seen at Late Prehistoric and historic period sites (see Binford 1981; Leechman 1951;

McKee 1988; Vehik 1977).

As an alternative explanation, Frison (1982) argues, based on seasonality studies from a number of Northern Plains bison kills, that procurement activities during the

Paleoindian period occurred predominantly during the winter. He suggests that the preference for cold season bison was to take advantage of natural refrigeration to store 253

meat. Hunting during the winter allowed carcasses to freeze and be preserved long

enough to provide hunters with stores of food and raw materials to sustain them

throughout the winter and early spring.

A decade later, Todd (1991) criticized Frison’s argument by showing that not all

Paleoindian-age kills fit this pattern. Paleoindian kills such as those at the Lipscomb and

Scottsbluff sites likely occurred during the summer, when it would not be possible to

freeze carcasses following the kill (Todd et al. 1990). Todd further observed that while

there may be a tendency for winter procurement, considering the vagaries of site

preservation, the sample of available sites may be too small to identify seasonal

differences in hunting tactics even if they existed.

Instead, Todd (Todd et al. 1990) and others (Hofman and Todd 2001; Hofman 1996;

Kelly and Todd 1988) claim that seasonality study of Paleoindian bison bone beds

indicates that bison (and presumably other large game) were killed year-round. This

differs significantly with the timing of bison hunts during the Late Prehistoric and

Historic periods, when bison hunting is usually confined to the late summer through fall.

Hofman (1992:198) suggests that this seasonality pattern indicates that Paleoindians practiced a strategy in which hunters moved steadily from kill to kill. Therefore, there should be no preferred season for bison hunting. This conclusion appears to contradict (1) what we know of the seasonal fluctuation in the condition of Holocene bison and other large game (Speth 1983, 1990, 1991), (2) known practices of modern hunter-gatherers who seasonally change the manner and intensity of procurement and processing of their faunal resources in response to variations in the physiological condition of their prey 254

(Speth and Scott 1985; Speth and Spielmann 1983), and (3) available seasonality data from Paleoindian sites (Bement 2003; LaBelle 2005; McCartney 1990).

Both Bement (2003) and McCartney (1990) present models for Paleoindian bison utilization relating bison hunting strategies to seasonal variation in bison behavior and condition. According to McCartney (1990), Paleoindian hunters would focus on whichever animals were in the best condition at that time of year, sometimes preying on nursery herds and other times hunting only bull herds. Kills at Casper, Olsen-Chubbuck, and Jones-Miller, which occurred during the early fall, represent nursery herds with low percentages of males. Late fall and winter kills, like Finley, Agate Basin, and Lamb

Spring, represent small bull herds. McCartney (1990:120) suggested that in the early fall,

Paleoindian hunters utilized nursery herds partly because they would have a higher biomass and higher fat content than male herds, which are in poor condition because of the late summer rut. Later in the year (winter and spring) the effects of pregnancy and lactation on female bison would make the small bull herds more attractive to hunters (see

Speth 1983).

More recently, Bement (2003) hypothesized that a dichotomy existed in the size and seasonality of Folsom period bison kills from the Southern High Plains. Small kills occurred in all seasons and resulted from ambush of bison near water sources. In contrast, large kills, which used natural features like arroyos to trap the bison, only occurred during the late summer or early fall when bison naturally aggregate into large herds. The dichotomy in kill size is also reflected in degree of carcass butchery at each site. At small 255

kills carcasses received extensive butchery, while at large kills carcasses underwent only

light butchery.

THE RECORD FOR PALEOINDIAN BISON BONE BEDS

Sufficient data on bison remains are available for 87 assemblages within the project

area representing at least 1988 bison (Appendix B). Included in this database are data on

size of the excavation area, dental age group (DAG), season(s) of occupation at the site,

the minimum number of individual bison, and the percentage of males and

females/immature animals.

There is a bias in this database for bison-dominated kill assemblage kills to be

preferentially located in grassland settings (Chapter 6). There are an additional five

localities for which site function was unknown. Plain/rolling hill assemblages (n = 59)

outnumber either alluvial valley (n = 14) or foothill/mountain assemblages (n = 14). In

addition, special-use kill/processing assemblages (n = 59) are significantly better

represented than multifunction short-term camps (n = 23) in this dataset.

There is also slight bias for sites dating after approximately 11,500 calBP (10,000 14 C yrs BP) (n = 57) than before this time (n = 37) this difference is not significant ( χ2 = 1.94;

df = 1; p = .16). These biases are not considered to be a significant problem, because the

primary focus of this analysis is understanding seasonal and long-term variation in the

timing and composition of bison kill events, rather than exploring regional differences in

the use of bison. 256

RESULTS

I present results here on kill seasonality, kill size, and herd composition.

Seasonality of Kill

I address here three questions about seasonality in Paleoindian procurement. First, during which season did most bison hunting occur? Second, does seasonality influence the size of a kill? Third, does the representation of male and female bison in the assemblage change according to the season of the kill?

The season of animal mortality of archaeological specimens is inferred from both mandibular tooth eruption stages and tooth occlusal attrition patterns (Frison 1978, 1982;

Todd and Hofman 1987; Todd et al. 1996). Such data from multiple animal bone beds can generate accurate assessments of the season in which an animal died. Dental age group values are presented in 0.1-year increments starting with an assumed middle of spring (i.e., late April) birth pulse (N) (Table 7.2).

This study utilizes a technique developed by Todd (1991:219) for depicting the distributions of dental age patterns from multiple sites. In this system “a frequency of occurrence per 0.1 year segments was tabulated by counting the number of assemblages spanning each increment.” For example, at sites such as Horner I, Horner II, Casper, and

Cooper, where dental age is limited to a single 0.1-year interval, just that interval is 257

counted; however, when dental ages span multiple increments the site or component is

counted in each 0.1 increment represented. For example, the dental age estimate for the

Plainview site ranges from N+0.6 to N+0.9 years, so this site would be counted in each the N+0.6, N+0.7, N+0.8, and N+0.9 intervals.

Table 7.2. Julian and calendar days associated with bison dental age groups. Dental Age Julian day(s) Calendar day(s) Group A N + .0 118–154 April 27–June 2 N + .1 155–190 June 3–July 8 Summer begins 173 June 21 N + .2 191–227 July 9–August 14 N + .3 228–264 August 15–September 20 N + .4 265–300 September 21–October 26 Fall begins 266 September 22 N + .5 301–337 October 27–December 2 N + .6 338–8 December 3–January 8 Winter begins 356 December 22 N + .7 9–45 January 9–February 14 N + .8 46–82 February 15–March 22 Spring begins 80 March 20 N + .9 83–117 March 23–April 26 A Assumes mean birth pulse date for bison of April 27 (Day 118)

Figure 7.2 depicts the seasonal distribution of kills events for all 84 sites with a

seasonality estimate. Although some bison hunting occurred throughout the year, hunting

is not uniform throughout the year. There is a weak but significant difference between the

observed distribution and the one expected if all intervals were equally represented ( χ2 =

25.9; df = 9; p = .002; Cramer's V = .13). Based on these data it would appear that late fall (N+0.6) is the most common time for kills to occur, while mid- spring (N+0) is the 258

least common time for kills. This implies that Paleoindians preferentially hunted bison

during certain times of the year.

Figure 7.2 Bar chart of relative frequency of 0.1-year dental age group intervals for 66 Paleoindian bison assemblages. Chi-square analysis ( χ2 = 25.9; df = 9; p = .002; Cramer’s V = .13) suggests significant difference between this distribution and equal representation of all dental age intervals.

Is grouping all Paleoindian sites together the best way to understand hunting scheduling? Figures 7.3 and 7.4 suggest that there may be important temporal and spatial differences in the timing of bison hunting. The distribution in the timing of bison kills during the early Paleoindian period is significantly different from the distribution during 259

the Late Paleoindian period (Kolmogorov-Smirnov D = .60; p = .03). During the early

Paleoindian period, the frequency of bison kills has a slightly bimodal distribution with

slight, but insignificant peaks during the summer (N + .3 – .4) and the late fall/early

winter (N + .6 – .7) ( χ2 = 11.9; df = 9; p = .22). In contrast, for Late Paleoindian-age bison bone beds show a significant bias for kills that occurred during late fall/early winter

(N + .6 – .7) ( χ2 = 22.1; df = 9; p = .009; Cramer's V = .15). This suggests that the beginning of cool-season bias in bison hunting, a trend well established later in time, occurred during the Late Paleoindian period. The similarities in the seasonal distributions between Figures 7.3 (Late Paleoindian) and Figure 7.4 (plains/rolling hills) are likely not accidental. The majority of sites used in this analysis are located in the plains/rolling hills it is also true that the majority of sites dating to the Late Paleoindian period are from plains/rolling hill settings.

That said, habitat differences in the timing of bison hunting are probably significant.

Unfortunately, because of small sample sizes within habitats, statistical evaluation of these patterns is only possible for plains/rolling hill sites. For foothill/mountain sites, bison hunting in general appears to have occurred at relatively low levels throughout most of the year, except for mid-summer through late fall (stages N + .3 – .6) when a small peak in bison hunting occurs. Although most sites in this setting are single- episode kills (e.g., Folsom, Horner I and II, Jerry Craig, Upper Twin Mountain), there is also evidence, such as at Hell Gap, for a nearly continuous or at least repeated occupation during multiple seasons. Winter use of foothill/mountain areas is indicated in spite of presumably severe climatic conditions (see Surovell 2003). 260

Figure 7.3 Bar chart of relative frequency of .1-year dental age group intervals for 23 Early (>11,500 calBP) and Late (<11,500 calBP) Paleoindian bison assemblages. Chi- square analyses suggest no significant difference between distribution for Early Paleoindian assemblages and equal representation of all dental age intervals (χ2 = 11.9; df = 9; p = .22), but significant difference for Late Paleoindian assemblages ( χ2 = 22.1; df = 9; p = .009; Cramer’s V = .15) 261

Figure 7.4 Bar chart of relative frequency of .1-year dental age group intervals for 9 alluvial valley, 12 foothill/mountain, and 45 plains/rolling hill Paleoindian bison assemblages. Due to small sample for these groups, statistical evaluation of these patterns is only possible with plains/rolling hill sites. Chi-square analysis ( χ2 = 22.8; df = 9; p = .00; Cramer's V = .15) suggests significant difference between distribution for plains/rolling hill assemblages and equal representation of all dental age intervals. 262

The record from alluvial sites is generally comparable to that from foothill/mountain sites, although here there was no evidence for bison hunting occurring during the late spring through early summer (N + .0 – .1) and more limited evidence for winter to early

(N + .7 – .9) spring hunting. The major of bison hunting in alluvial valleys appears to have occurred during the late summer through the fall (N + .4 – .6). The bison record alone would suggest that human occupation in these habitats was highly seasonal; however, Bamforth (2002b) and Jones (1999), using other faunal and floral seasonal indicators from sites such as Allen and Lime Creek, suggest that human occupation within alluvial valleys occurred in repeated visits throughout the year.

The timing of bison hunting events within plains and rolling hill settings differs from other habitat settings. For plains/rolling hill assemblages there is a weak but significant clustering of kill events during late fall through mid-winter (N + .6– .8) (χ2 = 22.8; df =

9; p = .00; Cramer's V = .15). Most sites in this region represent single-episode kills (e.g.,

Cooper, Olsen-Chubbuck, Scottsbluff, Bonfire Shelter, Lipscomb) although several sites

(Agate Basin, Jones-Miller, OV-Clary, Milnesand, and Plainview) have evidence for multiple or repeated uses during different times throughout the year. Currently, at no site in the plains/rolling hills is there evidence for year-round occupation.

Bison Abundance: MNI Values

This section examines the factors which influence bison abundance (as measured by

MNI) at these sites. For the sites considered here, bison vary from an MNI of one to 263 approximately 300 (Jones-Miller) (see LaBelle 2005). Inferences concerning the significance of these estimates are hampered by several weaknesses inherent in this database. First, published MNI values for some Paleoindian sites may not necessarily reflect the true number of animals originally represented at the sites, due to factors such as preservation and collection biases. For example, sites such as Linger, Milnesand, and

Nelson presently have relatively small MNI values; however, considering that these sites have poorly-preserved faunal assemblages it is likely that many more bison were actually killed at these sites than are reflected in the current MNI estimates (e.g., Hill 2002a). At other sites, such as Carter/Kerr-McGee, Laird, Lindenmeier, Norton, James Allen,

Blackwater Draw, and Winger, the limited size of the excavation area, incompleteness of excavations, or limited recovery of faunal remains by the excavators makes it nearly impossible to evaluate the accuracy of the MNI estimates for these sites (e.g., McCartney

1990). Further problems arise from the fact that some large kills sites, such as Jones-

Miller, Heron-Eden, Milnesand, and Plainview, are the result of multiple large kill events, and therefore the true size of each kill event is unknown.

Finally, additional difficulties in comparing sites arise because all the sites have unique taphonomic histories, and these histories are not well understood for a number of sites. Bone attritional processes may have adversely affected some sites, causing the MNI estimates for those sites to be misleadingly low. At sites such as Cattle Guard, Clary

Ranch, Lipscomb, Agate Basin, Horner, and Mill Iron, the taphonomic life histories of the collections have been intensively studied (Frison 1996; Hill 2001; Todd et al. 1990; 264

Todd 1987), whereas almost no information is available on the site formational histories of Jurgens, Carter/Kerr-McGee, Frasca, and Lindenmeier.

In spite of these problems, a number of conclusions can be drawn from the MNI data.

First, most Paleoindian kills included very few animals (Figure 7.5). Of the 84 assemblages for which MNI values are available the median value is just 7 animals. Even more striking, the 75 percentile for all sites is less than 30 animals. However, mass bison kills do exist. The largest single Paleoindian mass kills occurred at Olsen-Chubbuck (190 animals), Horner I (158 animals), and Casper (100 animals). Stanford (1999) reports an

MNI of approximately 300 animals from Jones-Miller; however, at least 3 kill events are represented at this site. Based on the infrequency of these huge kill events in this database, such kills were probably quite rare during the Paleoindian period. The more

“normal” size of kills made during this period was probably less than 15 animals.

Second, although site preservation and excavation area size certainly influence MNI values from sites, they are probably not determinant factors (LaBelle 2005). For example,

Figure 7.6 shows only a moderate, although significant, association between MNI and excavation size (r s = .42; p = .004).

These data suggest that although the largest kills are also the largest excavated sites, there are numerous sites exhibiting high bison MNI values with excavation areas smaller than 75 m 2 (e.g., Copper, Frasca, Lipscomb, Milnesand, Olsen-Chubbuck, and

Scottsbluff). Some sites with small MNI values have excavation areas greater than 125 m2 (e.g., Agate Basin, Aubrey, Allen, Cherokee Sewer, and Lime Creek). 265

Figure 7.5 Distribution of minimum number of individual (MNI) estimates for bison from 84 Paleoindian assemblages.

Figure 7.6 Plot of Log (MNI estimate) versus Log(excavation area [m 2]) for 84 Paleoindian bison assemblages with best-fit regression line. Best-fit regression line drawn to illustrate trend in data. Spearman’s rho shows a moderate, but significant, association between MNI and excavation size (r s = .42; p = .004). 266

In addition, I have shown that Paleoindian sites with poor preservation or incomplete recovery can still have relatively large MNI values if the analyst bases his skeletal part representation estimate on high-density skeletal parts, such as astragali or calcanei (Hill

2002a, 2002b).

Third, temporal and regional differences among sites can only explain a relatively small amount of the variation in MNI values from Paleoindian sites. For example, although the very largest kills date to after approximately 11,500 calBP, there is no significant difference between the distributions of MNI values from Early Paleoindian sites compared to Late Paleoindian sites (Kolmogorov-Smirnov D = .197; χ2 = 3.1; p =

.42) (Figure 7.7). In fact, the median value for Late Paleoindian sites (MNI = 6) is actually smaller than that for Early Paleoindian sites (MNI = 10).

A slight but insignificant difference also exists for MNI values of sites located in different habitat settings, with foothill/mountain sites generally exhibiting the highest

MNI values and alluvial valley sites the lowest MNI values (Kruskal-Wallis H = 4.45; df

= 2; p = .11) (Figure 7.8).

Seasonal changes in the size of kills may explain much of the variation observed in

MNI values, although high intra-group variation in seasonal MNI values makes it impossible to demonstrate statistically significant differences except in a few cases

(Kruskal-Wallis H = 7.5; df = 7; p = .38) (Figure 7.9; Table 7.3). The mean MNI for each season, and its associated 95% confidence limits, for the 58 assemblages with seasonality estimate confined to a single season. This analysis had to use four intermediate season intervals (spring/summer, summer/fall, fall/winter, and winter/spring) because a number 267

of sites were assigned to multiple seasons. The smallest mean MNI values are found in

spring sites. Spring sites also exhibit a narrow range of MNI values, so not only does this

time of year have a low mean, but all spring kills appear to consist of less than 11 bison.

Figure 7.7 Box plot comparison of minimum number of individuals (MNI) between 34 Early (>11,500 calBP) and 49 Late (<11,500 calBP) Paleoindian bison assemblages. Kolmogorov-Smirnov two-sample test (D = .12; χ2 = 3.1; p = .42) cannot reject null hypothesis that two samples have the same distributions. Largest outliers not shown here.

268

Figure 7.8 Box plot comparison of minimum number of individuals (MNI) between 13 alluvial valley, 15 foothill/mountain, and 55 plains/rolling hill Paleoindian bison assemblages. Kruskal-Wallis one-way analysis of variance by ranks test (H = 4.45; p = .11) cannot reject null hypothesis for equality of population median among groups.

269

Figure 7.9 Cross plot of mean minimum number of individuals (MNI) with 95% confidence interval over eight seasonal intervals for Paleoindian bison assemblages. Only sites with seasonal estimates confined to a single season were used in this study.

Table 7.3. Summary statistics for bison minimum number of individual (MNI) values for the eight seasonal estimates used in this study. Mean Standard Number Min. Max. Season MNI Deviation of sites MNI MNI Spring 5.2 3.2 6 2 11 Spring/Summer 23.3 15.2 4 1 34 Summer 33.8 49.9 13 1 190 Summer/Fall 27.5 14.3 6 7 41 Fall 39.3 53 11 2 158 Fall/Winter 15.9 15.6 9 4 47 Winter 21.3 15.8 6 5 44 Winter/Spring 39.3 32.7 3 2 63 Kruskal-Wallis one-way analysis of variance by ranks test (H = 7.5; df = 7; p = .38) cannot reject null hypothesis for equality of population median among groups.

270

Through the summer and fall, mean MNI trends are generally higher. These high mean MNI values are, however, associated with a wide range of values; therefore, while during these seasons there are a number of very large kills, there are also a number of small kills. Following this peak during the fall, the size of bison kills generally decreases through the rest of the year. During this time of year, most sites contain less than 30 bison; however, two very large winter kills (Hell Gap-Area II and Frasca) appear to be outliers to this trend for smaller late-season kills.

Herd Composition

Herd composition estimates are available for only 36 Paleoindian assemblages.

Because of this small sample size, this study did not separate Early and Late Paleoindian sites, but instead grouped all sites together. Except in four cases (Heron Eden,

Lindenmeier, 12 Mile Creek, and Lamb Springs) the remains of bison cows and calves outnumber those of bulls. As many of these sites represent single-event multi-animal kills, the abundance of females and immature animals suggests a predation strategy focused on attacking cow/calve herds. Figure 7.10 illustrates how the herd composition varies through the year. Cows and calf strongly dominate kill assemblages for nearly all seasons of the year. The only season when the frequency of bulls killed approaches the frequency of cows and immature animals killed is during the late winter/early spring.

Even then, the equity is not a result of equal numbers of bulls and cows/calves being represented in a single assemblage; instead, a number of the assemblages during this 271 period are bull-dominated.

As shown in Figure 7.11, kills with a high proportion of bulls are generally only smaller-size kills. Surprisingly, this trend does not even include isolated kills of a single animal, which consist entirely of cows or calves. Rather, when just sites with bulls present are considered, there is a general trend (but not statistically significant) for proportion of bulls to decrease with increasing size of kills (r s = -.24; p = .22).

Figure 7.10 Plot of mean %cow/calf (closed circle) and mean %bull (open squares) by eight seasonal intervals.

272

Figure 7.11 Plot of %males versus minimum number of individuals (MNI) for 34 Paleoindian bison assemblages. Spearman’s rho (r s= -.24; p = .22) shows weak negative relationship between %male and MNI for 28 Paleoindian bison assemblages with evidence for male animals present.

DISCUSSION

The most important finding of this study is that seasonality seems to strongly

influence the timing, size, and composition of bison kills, and that the actions of

Paleoindian hunters were greatly influenced by seasonal and long-term changes in bison

behavior and physiology. Overall, the findings of this analysis agree with the predictions

concerning bison kill composition, size, and season of mortality developed in chapter 3.

The available data indicates that Paleoindians hunted bison on a year-round basis,

although the intensity of this hunting was not uniform across the year. Mid-spring

through early summer was the least common period for bison kills, while mid-summer 273 through winter was the most common. The peak season of bison hunting changed over time and across habitat zones. During the Early Paleoindian (>11,500 calBP) period and in alluvial valleys and foothill/mountain settings, mid- to late summer was the preferred hunting season, whereas during the Late Paleoindian (<11,500 calBP) and in plains/rolling hill settings, late fall through winter was the most common time for bison hunting.

In addition, although very large multi-animal kill assemblages such as Olsen-

Chubbuck, are often perceived as “typical” Paleoindian kills, the reality is that most bison kill sites during this time consist of just a few animals. The majority of Paleoindian kills consist of less than 30 animals. Kill size did not significantly change over time or by habitat setting; the key variable for explaining kill size is apparently season of mortality.

In the spring, kill sizes were relatively small. During the late summer and fall, kill size tended to increase significantly. Through the winter and into the early spring, kill size trended back downward toward its spring minimum.

This analysis supports the earlier findings of McCarthy (1990), who suggested that

Paleoindians preferentially preyed upon female-dominated herds throughout most of the year. Hunting of the comparatively small male herds appears to have been rare and only seems to have occurred during or just prior to the calving season. This period would have been when pregnant or lactating females were in their poorest condition. It seems that for most of the year, the significantly large size of cow/calf herds made them a more attractive target for Paleoindian hunters, especially when cows were in equivalent or better condition than bulls. 274

Overall, these seasonal shifts in bison hunting intensity probably reflect hunters’

responses to changes in prey behavior and physiology. Kill sites are smaller and less

frequent in the spring, corresponding roughly to the calving period between late April and

early May (Frison and Reher 1970; McHugh 1958). Frison (1974, 2004) states that

during calving period, bison are very difficult to handle. In fact, he claims that hunters would be wise to leave cows with very young calves alone. The reactions of calves are simply too unpredictable when frightened. At the least provocation, calves bolt from the herd and their mothers are usually forced to follow them. With this erratic behavior of cows and calves during this time it would be nearly impossible for hunters to successfully

prey on bison.

Similarly, the rut is a period of frenzied activity for mature bulls. McHugh (1958)

noted that, during the rut, bull behavior is noticeably different from their behavior during

the rest of the year. Bulls are quick to fight and commonly become aggressive to any

outsider coming near the herd, even a human. The behavior of the bulls during this time

of year makes all members of the herd very skittish and quick to react to danger.

Paleoindian hunters probably understood that during the rut the rambunctious behavior of

the males made it difficult for hunters to successfully make a large kill (Frison 1974).

The major period of bison hunting was the late summer through winter. The late

summer through fall is when herd size is still fairly large and most bison (especially

females) are in peak condition. Even during the late winter, pregnant cows still probably

represented highly prized targets. In addition, as Frison (1982) notes, the abundance of

late-season kills, especially in the north, may indicate a strategy of over-winter freezing 275

of carcasses. Of course, this possibility alone does not explain bison hunting at other

times of the year.

Throughout my discussion, I assume that Late Pleistocene/Early Holocene bison

responded to environmental and social stimuli in ways comparable to modern bison.

Paleoindian-age bison may have differed from modern bison in many ways, including

gregariousness, defensive mechanisms, attunement to predation, cursorial ability, length

of growing season, and inter-specific competition (Guthrie 1989; Hofman 1996; Smiley

1978; Todd 1991). As a result, we cannot be certain whether Paleoindian-age bison herds

underwent the same seasonal fluctuations in size and condition as do modern bison.

However, considering that the rut and calving are biological processes that are not

determined solely by environmental conditions, Late Pleistocene/Early Holocene bison

may not have differed much from Late Holocene bison. As discussed in chapter 4, bison

populations increased dramatically following the Pleistocene extinctions of other

megafauna and appearance of large grasslands throughout the Plains at the end of the

Pleistocene (Guthrie 1990; McDonald 1981). While these herds probably never reached

the massive size of the Late Holocene herds, they were probably fairly large and likely

underwent some seasonal fluctuations in size. Based on the high degree of sexual

dimorphism of bison throughout the Pleistocene and Early Holocene, Guthrie (1990:156–

157) argues that North American bison have always had a polygamous social structure.

Therefore, as with many polygamous animals, bison bulls would likely have lived isolated from females for most of the year and only joined the nursery herds for the 276

purpose of mating. If this was the case, then Paleoindian-age bison underwent seasonal fluctuations in herd size broadly analogous to modern herds.

The one area where early bison probably differ from modern specimens is in degree

of annual variation in condition. As Todd (1991) states, “given current understanding of

the Late Pleistocene/Early Holocene seasonality, and particularly Guthrie’s model of

longer periods of plant productivity per annum, seasonal fluctuations in body fat

condition may not have been as extreme as they were by the Middle and Late Holocene.”

This is not to suggest that Late Pleistocene/Early Holocene bison did not face seasonal

stress. It simply implies that these periods may not have been as intense or as long-lasting

as they were during the Late Holocene. As a result, this may explain why bison were

taken during non-peak periods.

CONCLUSIONS

Paleoindians relied extensively upon bison, which would have provided a large

supply of protein and fat, as well as raw materials like skins, sinew, and bone for tools.

Paleoindian hunters responded to seasonal changes in bison physiology and behavior by

varying when they made a kill, how large a kill was made, and which type of herds to

focus their attention on. Such changes would have had important implications for (1)

labor requirements needed to make a kill and process the carcass, (2) landscape features

used by groups hoping to find bison to hunt, and (3) the seasonal role non-bison prey in

the diet of Paleoindians. 277

CHAPTER 8: NEW PERSPECTIVE AND IMPLICATIONS

Much has been made of the divided viewpoints on Paleoindian subsistence and land use. Some researchers see Paleoindians as highly mobile, specialized big game hunters who relied heavily on well-made bifacial tools created from exotic lithic raw materials.

Others instead propose that Paleoindian foragers practiced a more sedentary existence with considerable reliance on small game and plants and low-tech chipped stone tool kits made from local raw materials.

The Great Plains and Rocky Mountains have uneven distributions of biotic and abiotic (e.g., chipped stone and water) resources; this and long-term environmental changes should have affected Paleoindian subsistence, land use, and mobility. General principles from behavioral ecology and studies of modern bison behavior and biology have been used in this study to develop testable hypotheses on different aspects of prehistoric land use and subsistence (chapters 2 and 3).

I have argued that the competing perspectives outlined above simply describe the end-member expressions of a single adaptive system, which shifted in response to differences in the constancy and contingency of resource availability across the landscape.

This chapter summarizes the major findings from this study and then discusses the implications for (1) the importance and presence of communal kills and aggregation sites and (2) the organization of male and female labor during the Paleoindian period. Finally,

I compare two aspects of faunal use—big game hunting and resource intensification— 278 between Paleoindians and later populations living in the Great Plains and surrounding areas.

SUMMARY OF RESULTS

Land Use and Residential Mobility

As shown here and elsewhere (Hill 2007; Knell 2007; LaBelle 2005), there is a covariance between environmental zone (i.e., foothill/mountains, plains/rolling hills, and alluvial valleys) and patterns of forager land use. Paleoindian foragers structured their land use according to the presence and nature of the biotic and abiotic resources within major environmental zones. Data presented in this study seem to support the mobility models outlined in chapter 2, which predicted that foragers would move more slowly through rich, high-density environments than though environments where important resources are more scattered. In general, sites in grassland settings with low diversity of resources have lower artifact densities and are often dominated by exotic lithic raw materials. They apparently represent less intensive occupations than do sites in foothill/mountain or alluvial valley settings with ecologically high density and high diversity. These differences highlight the fact that certain environments have higher proportions of short-term camps than do other areas, and that camps have higher artifact density than do other types of sites. In addition, foragers clearly responded to proximity of high-quality lithic raw material sources (e.g., Edwards, Hartville Uplift, White River 279

Group silica, Knife River). Sites located near these types of lithic sources generally have

higher artifact densities than do sites that are farther from lithic sources or those near

other lithic sources (e.g., Alibates, Tecovas, Florence).

These differences in artifact density reflect, at least in part, site function differences

among settings. Multi-component camps and lithic workshops are by definition high-

artifact-density sites. Similarly, kill and processing sites generally are low-artifact-density

sites

The presence of a good quality lithic source in a particular region does not always

mean that sites in those areas will have very high artifact densities. For example, the

Alibates and Tecovas sources both are sources of good-quality cherts, but neither is

associated with large, dense sites. In addition, alluvial campsites associated with outcrops

of Niobrara chert are more common and have higher artifact densities than do camp sites near Niobrara cherts that are located in High Plains environments. One of the key issues here is not the chert source itself, but what other resources (plants, animals, and water) are nearby These areas did not provide abundant and reliable sources of wood, water, and a diversity of plant and animal resources in comparison to areas around the Edwards,

Hartville Uplift, or Knife River sources (see discussion in Hofman 1999). Foraging decisions are clearly based on the presence or absence of a suite of key resources, not simply one or two important resources.

While no statistically significant trends in occupational intensity were observed, the data do suggest changes in land use through time. Camp and workshop sites in most major environmental zones have reduced artifact densities with time, perhaps reflecting a 280 slight decrease in occupational intensity in most of the project area over the course of the

Paleoindian period, but no dramatic shifts in land use.

The cause for such a decrease in site occupation intensity is unclear but may be explained by groups moving slightly more rapidly through landscapes or by a modest decline in overall population size. Either change could result from decreased environmental productivity at the Pleistocene/Holocene transition due to increasing aridity. Such a change would result in lower carrying capacity and thus scarcer resources.

It is also possible that this trend for a decrease in artifact density through time relates to the fact that it is easier to separate individual occupation episodes at later sites than earlier sites. While this is an intriguing possibility, direct evidence for or against this idea is not immediately available and more work will be necessary to determine the cause(s) behind this patterning.

Diet Breadth and Diversity in Faunal Use

One important theme in this study is the effect of site function and location on our perception of prehistoric subsistence practices. Admittedly, researchers must look at both camp and kill assemblages to understand the full range of fauna being used. This fact, however, does not mean that our perspective on Paleoindian faunal use has been biased because of the presence of a high number of kill sites in the Paleoindian record, as has often been claimed. My estimate of large game abundance in the Paleoindian diet does not change whether the analysis considers kill and camp assemblages or camp 281 assemblages alone. Large game animals (bison and mammoth) dominate most

Paleoindian assemblages (see also Hill 2007). Undeniably, large game hunting was an important component in Paleoindian subsistence.

There are important links among Paleoindian land use, resource distribution, group mobility, and diet. In the low-diversity grasslands of the High Plains and Rolling Hills, prehistoric groups made nearly exclusive use of large fauna, especially bison, for the entire Paleoindian period. Such a strategy follows the predictions of behavioral ecology models in that it is more efficient for foragers to focus primarily on the highest ranked resources under conditions of high dispersion.

In more diverse environments, such as alluvial valleys and foothill/mountains, there is evidence for greater variety in the patterns of faunal use, largely the result of long-term changes in large prey availability. When large game animals were present in these settings, prehistoric hunters probably pursued them, and indeed, large game hunting was especially common nearly everywhere early in the Paleoindian period. Later in time, and probably associated with a decrease in abundance of bison outside of the Great Plains proper, smaller fauna appear in more frequency in alluvial valleys and foothill/mountain settings.

Although the archaeological records of both alluvial valleys and foothill/mountain areas show different predation strategies than those typically employed in plains/rolling hill habitats, uses of small and medium size game differed greatly between alluvial and foothill/mountain settings. Foothill/mountain camp zooarchaeological assemblages usually are dominated by a single prey category, generally bighorn sheep or deer at 282 specialized hunting camps, and lagomorphs or rodents at short-term residential camps.

Alluvial camp assemblages, on the other hand, exhibit a high degree of evenness among small- and larger-sized prey types.

Hunting Pressure and Human Impacts on Fauna

This study also examined two effects of Paleoindian human hunting on the physiology and population structure of prey populations. Using data on Late Quaternary body size changes and population age structure, chapter 5 considered the relative effects of human hunting and climate-driven environmental change on bison populations over the last 40,000 years. Chapter 6 examined whether the changes in diet breadth during the

Paleoindian period were a result of human-induced resource depression. Both issues are useful in exploring the nature and manner in which overhunting might have affected fauna in North America. These data may also shed light on the possible role human hunting played in the extinction of 35 genera of Rancholabrean land mammals.

This study’s finding that bison underwent dramatic body size change over the last

40,000 years is not surprising (McDonald 1981; Wilson 1980). There are multiple possible explanations for this body size decrease. Had there been intensive hunting pressure on Late Quaternary bison populations, it should have produced (1) a juvenile bias in bison population structure, and (2) a decrease in the mean body size of bison over time (Chapter 5). On the other hand, environmentally-driven shifts in carrying capacity, intraspecific competitions, or thermal regime would be reflected in a (1) decrease in 283 median and maximum body size, and (2) prime-age bias in age structure. The results of this study show that body size reduction was not a slow steady decline in the mean size value; instead, long periods of stability in body size were interrupted by short, rapid declines in both the median and maximum body sizes of bison. Body size decline was not associated with evidence for a juvenile bias in age structure. The results suggest that a number of environmental factors reduced the selective advantage of large body size for bison, including the expansion of C 4 grasses throughout North American grassland and relaxation of interspecific competition and predator pressure that came with the extinction of large predatory cats at the end of the Pleistocene.

Chapter 6 examined how changes in hunting intensity may have affected the proportion (evenness) and type of small game in Paleoindian faunal assemblages. Long- term trends in foraging efficiency and diet breadth generally co-vary with regional differences in land-use strategies. In alluvial valleys and foothill/mountain settings, where occupational intensity appears to have been the highest, sites exhibit evidence for an expansion in diet both in terms of the number and type of prey items over time. This change in diet breadth also involves a shift from easy-to-catch small prey to hard-to-catch small prey, meaning greater technological investment per unit meat obtained. In contrast, the diet in the plains/rolling hill settings, which had low occupational intensities through time, appears to have been more narrowly focused on large game throughout the

Paleoindian period.

284

Bison Hunting

Because bison generally dominate Paleoindian zooarchaeological assemblages,

variation in bison behavior and physiology exerted a very significant influence on the

behavior of Paleoindian hunters. I previously outlined testable predictions about the size and composition of Paleoindian bison kills and the preferred hunting season based on a review of modern bison ethology and biology (Chapter 3). Paleoindian bison hunting as seen archaeologically is generally is consistent with the predictions made. The most important variable in understanding the timing, size, and composition of Paleoindian bison kills is season of the year.

Bison kills occurred year-round, although the most active period was mid-summer to early winter. Timing of kills varied slightly by habitat setting. The most intensive hunting corresponds to seasons when herds are generally large, and cows in or near peak condition. Bison hunting is less common when herds are small or when cows were in poor condition.

Paleoindian hunters preferentially targeted cow/calf herds and mostly ignored bull herds. This makes a certain amount of sense considering that cow/calf herds are larger and females are less depleted from the mating season. Cow/calf herds therefore represent greater caloric return than do the small bull herds. In fact, the few examples of bull herd kills occurred during the late winter/early spring, when most cows would be in relatively poor condition just before or immediately after giving birth. The number of animals killed per hunting event (kill size) changed seasonally and generally mirrored natural 285

changes in herd size.

IMPLICATIONS

This section addresses how the findings of this dissertation study help us to understand more about the (1) the importance and/or presence of communal hunting and aggregation sites, and (2) the organization of male and female labor during the

Paleoindian period.

Communal Kills and Aggregation Sites

It is generally accepted that Paleoindian population density was very low and the

residential mobility was high, at least seasonally. Living in a largely empty landscape,

Paleoindians’ need for information exchange, economic cooperation, mates, group rituals or ceremonies, and access to trade goods was likely at a premium. Archaeologists have long assumed that Paleoindians satisfied these needs through periodic and temporary aggregations (i.e., fusion) of different groups (e.g., Bamforth 1988, 1991; Greiser 1985;

Wilmsen and Roberts 1978).

Bamforth (1991) suggests Paleoindian social aggregations were tied to communal

bison hunts. Although Driver (1990) defines communal hunts as hunts involving two or

more hunters, a more appropriate definition is a larger-than-normal hunting party created

by a temporary of fusion individuals from multiple social groups (see Hofman 1994: 286

348). Bamforth (1988, 1991) and others (Bement 1999, 2003) also see the fusion of numerous groups as a component of preparation for a large kill, because it is the only way to obtain the necessary labor to complete all the tasks associated with conducting a bison hunt. For example, before the kill hunters need an adequate stock of weapons, meaning that they must acquire the appropriate raw materials and manufacture large numbers of spear points and shafts ahead of time. To facilitate killing of a large number of animals in open settings, hunters must also construct or repair the corral or trap that will be used to contain the animals being hunted. In addition, after a successful kill, a large labor force is needed to butcher the carcasses and process the meat and hides before the carcasses begin to decompose (Frison 1991).

Hofman (1994) is critical of any perspective that argues for aggregations as a single type of site or event. He suggests that Paleoindian demography, subsistence, and land-use patterns may not fit any existing model as to why hunter-gatherer groups aggregate.

Instead, he proposes that hunter-gatherer aggregations were a mechanism by which different groups could solve problems that were beyond the capacities of any single group, such as finding spouses, conducting necessary spiritual rituals, and gaining information about very large geographic regions (Hofman 1994:342). As a result, he sees aggregations arising for a wide variety of reasons and under diverse conditions.

Therefore, there may be multiple archaeological expressions of aggregation localities depending on when, where, and why the fusion of different groups occurred.

The current study provides no new insights into what particular sites, if any, represent aggregation localities. However, this study’s emphasis on how Paleoindian land use and 287 occupational intensity related to the spatial and temporal predictability of key resources does provide some insights about where aggregations were likely to occur. I agree with

Hofman (1994) that if periodic aggregations did occur during the Paleoindian period, the diverse environments of the foothill/mountains and alluvial valleys would have been the most likely settings for this. Normally dispersed groups coming together would have found the protected settings, availability of water, chipped stone, and wood, and diverse and reliable food resources important for supporting the larger groups. This study shows that sites in these areas had high occupational intensity. In addition, as noted by Hofman

(1994:351), these habitats would also provide distinctive land features that could serve as navigation markers for communication where groups should meet in the future.

It seems unlikely that social aggregations would have involved large communal hunts. If they had, then the large kills common in plains/rolling hill environments would have been the loci of aggregation. This would have been problematic for several reasons.

First, in these grassland environments the dearth of major landscape features would have made it more difficult for groups to locate each other in a predictable manner. Second, except for bison there were few resources that would have attracted large groups of people or sustained aggregations for any length of time. In the event that the hunting expedition was unsuccessful, the entire group would have been put at risk because there are few alternative resources that could have been easily and quickly obtained. Third, there may be no correlation between the number of animals present at a kill site and the number of people involved in a hunt (Hofman 1994). A number of factors, including terrain, season, and hunters’ skill, greatly influenced the size of a kill. Finally, with few 288 exceptions, most large kill sites have low artifact densities. There is little empirical evidence for large numbers of people coming together in these settings.

Organization of Male and Female Labor

Throughout this dissertation I have largely discussed Paleoindian foragers as ungendered actors. I believe this approach is appropriate given the resolution of data the

Paleoindian records generally provide. Specifically, I see little empirical evidence for distinguishing the roles of men, women, and children. This said, if archaeologists hope to obtain a fuller understanding of the multifaceted nature of subsistence tasks, it is appropriate to examine prehistoric decision-making from the perspective of gendered decision-making.

Cross-cultural ethnographic and ethnoarchaeological research has repeatedly shown that female foragers make an important, if not dominant, contribution to the diet of hunter-gatherers (e.g., Hudecek-Cuffe 1998; Lee and Devore 1968; Waguespack 2003).

As discussed by Waguespack (2005), the contribution of women in foraging societies is not limited to gathering plant resources but also may include hunting of game and, in other instances, assistance to male hunters. However, because Paleoindian foragers have traditionally been viewed as male big game hunters, the contribution of women in these activities has largely been ignored (Gero 1993; Hudecek-Cuffe 1998; Waguespack 2003).

If, as the ethnographic record suggests, the existence of the division of labor is an empirical fact for all hunter-gatherer societies, than a few logical ties can be made 289 between the dependence on hunted versus gathered resources and the organization of male and female labor (Kuhn and Stiner 2006; Waguespack 2005). According to Kuhn and Stiner (2006) one of the few generalizations that can be made about modern foraging subsistence behavior is that males are normally responsible for obtaining large terrestrial mammals and large aquatic animals, while women and children are more likely to be responsible for collecting plants and small animals. While this very general of division of labor is nearly universal, Kuhn and Stiner (2006) are careful to note that individual deviations from this pattern are widespread in ethnographic accounts.

Despite gendered differences in subsistence activities, the foraging decisions of men and women are often complementary and together suggest a larger economic agenda for the long-term survival of the group. In her study of the subsistence practices of 71 hunter- gatherer groups, Waguespack (2005; see also Keeley 1988; Kuhn and Stiner 2001) drew several conclusions. First, there is a consistent relationship between the type of plant foods consumed and the contribution of meat to the diet. Specifically, in societies where dependence on meat is high only a small suite of plant resources, primarily fruits and roots, are consumed. In contrast, as dependence on meat decreases a greater range of plant products, including nuts and seeds, more commonly enter the diet. The general relationship appears to be that an increase in the importance of hunted animals correlates with the use of plants that require little processing. Second, a greater dependence on hunted resources is associated with a decrease in the amount of time women spend directly procuring food (Waguespack 2005). This is not to imply that in societies heavily dependent on hunting that women do not participate in locating and aggregating prey, 290 butchering/cooking the results from the kill, or manufacturing/repairing hunting equipment. However, there does appear to be a relationship between increasing meat dependence and female responsibility for non-subsistence related activities, such as house building, leather working, and burden carrying (Waguespack 2003:671). These findings suggest that when a group increases their reliance on large game hunting, women increasingly focus their time away from procuring food themselves, and more towards non-subsistence activities and tasks that facilitate male hunting.

The results from this study strongly suggest that Paleoindian hunters preferentially utilized the plains/rolling hills as a hunting ground for the high-ranked faunal resources such as bison and mammoths. As a result, it is unlikely that in these settings females contributed much towards the direct procurement of food, although women and possibly older children likely assisted men in locating and maneuvering game, as well as doing much, if not all, of the butchering of the carcasses of these large game. In contrast, the contribution of female foragers to feeding the group was probably greater in alluvial valleys and foothill/mountain settings, where there is greater evidence for residential activities and a greater range of resources being exploited, especially small animals.

There also appears to be seasonal variation in the intensity of use of large game, suggesting that the contributions of men and women to the meat diet may have shifted throughout the year. For example, during the late summer through late winter,

Paleoindian groups may have decided that risky, male-related activities (big game hunting) would provide the greatest economic return. As a result, the entire group or some subgroup (young hunters) probably moved to areas where they expected to 291

maximize men's encounters with large game (Jochim 1988; Elston and Zeanah 2002;

Winterhalder 1986). During the spring and early summer, big game hunting was expected to provide much lower returns, and groups may have decided to focus more on female- related low risk resources (small animals and plants) and preferentially positioned themselves in areas (e.g., alluvial valleys) where those resources were most abundant.

It is also possible that prehistoric groups practiced mixed mobility strategies to accomplish these different foraging goals. Mobility associated with foraging for small animals and plants would probably consist of short but repeated residential moves to new areas. When resources in that area became depleted, the group would move to a new previously unexploited area. When the groups wanted to go after large game, it may have required logistical forays undertaken by a subgroup of hunters and a few assistants to butcher the animals and carry material back to the rest of the group. In this case, women, children, and old people may have preferentially remained behind in ecologically rich areas, while young male hunters were more likely to travel into grassland settings in the search for large game. In this mixed strategy, the group as a whole reduces its chance of starving by relying heavily on low-risk resources, while simultaneously gambling on a big payoff if a kill is made .

AFTER THE PALEOINDIAN PERIOD

Many of the social and economic systems utilized by Paleoindians are very different from those used by later hunter-gatherer peoples. This section evaluates several aspects 292 of Paleoindian faunal use, such as big game hunting and long-term trends in resource intensification, to enable comparisons to later groups living in the Great Plains and Rocky

Mountains.

Big Game Hunting after Paleoindian Period

The economic importance of hunting of bison and other large game animals, which characterizes hunting during the Paleoindian period, was clearly significant well after the

Paleoindian period ended. There are some similar geographic and site function trends in large game use in later time periods. First, residential camps show the greatest evidence for diet breadth diversity. Second, sites located on great plains/rolling hill generally have higher representation of large game use, especially bison, than do sites in alluvial valleys or Rocky Mountains. Third, there appear to be east-west and south-north trends for greater use of bison (e.g., Bozell 2006) during all phases of the Paleoindian period.

It is likely that the Paleoindian-Archaic transition does not reflect as dramatic a shift in subsistence or land use as was once thought. Blackmar and Hofman (2006:47) state,

“some of the Plains Archaic complexes had economic strategies that, at least during some seasons and years, were comparable to those of the earlier Paleoindian bison hunters.”

Nothing highlights this pattern of gradual change more than the continuation of intensive bison hunting during the Archaic period. Large bison assemblages that date to the Early and Middle Archaic include Hawken, Head-Smashed-In, Hill, Simonsen, Kobold,

Cherokee-Sewer, and Spring Creek (Frison 1998; Widga 2006b). Small Archaic-period 293

bison kills also occurred in Great Plains environments seen at several components at

Lubbock Lake and Blackwater Draw (Johnson 1987; Hester 1972).

The species diversity of the Archaic period zooarchaeological assemblages and

seasonality of mortality do not differ significantly from Paleoindian faunas (Sheenan

2002). In addition, at least during the first half of the Archaic cases of bison carcass

processing is limited to cold marrow extraction, the same as seen during the Paleoindian

periods (Hill 2001, 2005).

This is not to say that there were no differences in large game predation from the

Paleoindian through Archaic periods. Sheenan’s (2002) study of 12 faunal assemblages

from Paleoindian, Early Archaic, and Middle Archaic assemblages found that bison

hunting was less common during the Early Archaic than during either the Paleoindian or

Middle Archaic periods. In addition, the bison hunting that did occur during the Early

Archaic primarily consisted of predation on individual animals or small groups. He suggests this temporary decline in bison hunting was a response to climatically-induced reductions in grassland productivity and forage quality during the Altithermal, which decreased bison herds (Sheenan 2002). Fewer bison on the landscape meant less frequent kills and smaller kills when they did occur.

Widga’s (2006b) reanalysis of a number of Early Archaic faunal assemblages, however, results in a different conclusion. He argues for at least a seasonal emphasis on hunting bison during the Early Archaic period. Specifically, Widga shows that a number of the Early Archaic components at Spring Creek and Logan Creek represent single event mass kills of more than 15 bison. The presence of such large kills during this period calls 294

into question Sheenan’s suggestion that Early Archaic hunters only undertook small kills.

By the Late Archaic, large-scale bison hunting appears to have expanded in terms of

number of kills and sophistication of bison traps used. Classic bison kills dating to this

period include Bonfire Shelter, Kaplan Hoover, Mavrakis-Bentzen-Roberts, Kobold,

Yonkee, and Head-Smashed-In (Frison 1991, 1998; Todd et al. 2001).

The most significant developments during this period is (1) the adoption of bison

jumps as a standard technique for killing bison and (2) the beginning of pemmican

production (Byerly et al. 2005; Dibble and Lorrain 1968; Frison 1991; Quigg 1998;

Reeves 1990). These two developments indicate increased levels of labor investment in carcass processing and a commitment to place not seen at any of the earlier periods, and probably signify dramatic intensification in bison exploitation more generally.

The earliest bison jump site is probably the 5700 14 C years ago Head-Smashed-In site,

if not the earlier Bonebed 2 at Bonfire Shelter (Byerly et al. 2005). Nevertheless, it is in

the Late Archaic when this technique became the standard for high-quantity killing.

Bison jumps were certainly an effective means of killing a large number of bison at one

time, but the most significant new development associated with these sites is the

construction of cairn-marked drive lines that directed scattered bison and concentrated

them toward a narrow jump-off point (Brink and Rollans 1990). Construction,

maintenance, and use of these drive lanes represent extensive planning and close

cooperation of a large number of people, which had not been important previously.

The introduction of jump kills during the Late Archaic period did not mean that

natural traps and surrounds fell out of use. A number of Late Archaic Northern Plains 295 sites, such as Kaplan-Hoover, Powers-Yonkee, and Rourke, represent localities where large numbers of bison were killed in arroyo traps (Todd et al. 2001). Similarly, in the

Southern Plains, where appropriate jump localities are few (except for Bonfire Shelter), much bison hunting occurred during the Late Archaic (Buehler 1997; Dibble and Lorrain

1968; Quigg 1998). In general, these kills are small to moderate in size (< 50 animals), although the Late Archaic jump kill at Bonfire Shelter may have involved between 197 and 800 prey animals.

Late Archaic intensification in bison hunting did not take place everywhere, however.

In fact, part of the project area instead witnessed a decline in use of bison during this time. Climatic changes resulting in an expansion of oak-hickory woodlands through much of the eastern portion of the Great Plains correspond to a decrease in use of bison in eastern sites, such as at Nebo Hill (Kay 1998b). In addition, throughout the Archaic period deer and bighorn sheep appear to have been economically more important than bison in the foothill/mountain sites (Frison 1998).

The intensive use of the bison continued throughout most of plains/rolling hills and alluvial valleys of the Great Plains over the last 2000 years. Bison remains are common at many western Woodland (Bozell 1995, 2006; Frison 1991; Johnson and Johnson 1998),

Plains Village (Drass 1998; Roper 2006; Winham and Calabrese 1998), middle ceramic

(Browowske and Bevitt 2006), and other Late Prehistoric sites (Koch 1995; Logan 1998;

Ritterbush 2006; Scheiber 2006; Wedel 1959), although bison was not always the dominant species (see Bozell 1995).

As with many Paleoindian sites, later zooarchaeological assemblages exhibit a high 296

diversity of other terrestrial, aquatic, and avian fauna based on site function and habitat

setting. For example, at many eastern sites and foothill/mountain localities, bison is not

the dominant game animal; instead, deer and mountain sheep are more common (Bozell

1995; Johnson and Johnson 1998). Unlike Paleoindian sites, later sites show evidence for

heavy reliance on numerous wild plants and various cultigens. From about A.D. 1 until

the time of Spanish contact, the relative importance of bison across the project area

varied, possibly because of natural variations in bison population levels (Creel et al.

1990; Dillehay 1974).

Southern Plains. Vehik (2006) summarizes the record for the late prehistoric period

on the Southern Plains. Between A.D. 900 and 1200, most Southern Plains groups lived

in relatively sedentary horticultural villages, relying on a mixture of hunting, gathering,

and horticulture. While bison was hunted, deer were the prime large game species. From

A.D. 1200 to 1450, in association with an upturn in bison populations, reliance on horticultural products and bison seems to have increased at the expense of other wild resources. One example of this is in Antelope Creek/Odessa Phase villages in the Plains

Border area of Oklahoma and Texas, where seasonal mass kills probably supplied the bulk of meat consumed by the residents of these sites (Brosowske and Bevitt 2006; Drass

1998; Winham and Calabrese 1998). After A.D. 1450, a number of Southern Plains groups (Wheeler and Garza Phase) appear to have abandoned their heavy reliance on horticulture and become seasonally mobile bison hunting specialists. These groups used the byproducts of bison hunting to trade with village farming communities in the southwest and eastern woodlands. 297

Central Plains. For the Central Plains, Bozell (1995) suggests that between A.D. 1 and A.D. 1000, sites in the western Plains exhibit moderate quantities of bison remains, while those in the east have very few. Later, very few bison remains have been recovered

from Central Plains tradition sites (A.D. 1150 – 1375) (c.f. Logan 1998). After

approximately A.D. 1500, there is a dramatic increase in the representation of bison in

the Central Plains archaeological record.

Northern Plains and Rocky Mountains. Over the last two millennia, hunting

technology diversified and carcass processing intensified steadily with the introduction of

bow and arrows, the use of large scale jumps and corrals features, and developments of an almost industrial scale of bone grease production. This is when the bison kills in

Northern Plains of Wyoming, Montana, and Canada were probably the largest (hundreds or thousands) and most frequent of any location throughout history. The Late Prehistoric

Northern Plains can probably be thought of as the penultimate expression of a bison hunting culture (Frison 1991; Reeves 1990). Evidence for pemmican production and the presence of fewer-than-expected scapulae and metapodials at many of these sites may indicate these Northern High Plains kill/processing localities were supplying food and bone tools (e.g., scapula hoes) for populations living in more eastern villages (Scheiber

2006). However, others suggest that bison remains from eastern villages came from

small-scale localized hunting strategies (e.g., Bozell and Ludwickson 1998).

As discussed by Frison (1991), there were long-term fluctuations in the intensity of

bison hunting. Late Woodland (Besant and Avonlea) period bison kills are present at

Ruby, Muddy Creek, and Wardell. Few bison kills appear to date between A.D. 1000 and 298

A.D. 1500. It is only after A.D. 1500 at sites like Glenrock, Big Goose, and Vore that

very large communal kills return in any frequency (see Bozell 1995).

Long-term Trends in Resource Intensification and Diet breadth

The appearance of Late Archaic communal bison jumps represents a substantial intensification in bison hunting and is quite different from the record of the Paleoindian period and earlier portions of the Archaic period. Two other later (Late Archaic and

Woodland) developments also represent increases in resource intensification and diet

breadth over Paleoindian levels: use of plants and bone grease extractions. Interestingly,

site location is just as important a variable for interpreting evidence of resource

intensification and diet breadth at later sites as it is for Paleoindian locations.

Use of Plants . The best direct evidence for the role of plants in Paleoindian

subsistence is found at the Barton Gulch site in the mountains of southwest Montana

(Davis et al. 1989). At this site 36 economic species were identified, with weedy annuals,

grasses, and fruits dominating the assemblage. In addition, at both the Late Paleoindian

Schiffer Cave and at Medicine Lodge Creek rock shelter in the Bighorn Mountains of

Wyoming, Paleoindian deposits include numerous storage pits containing dried and

charred seeds from several economic species, including amaranth, chokecherry/wild

plum, prickly pear, sunflower, and wild rye (Frison 1991). Finally, Baker Cave in Lower

Pecos River valley of Texas produced Late Paleoindian age charred seeds from 16

species of seed, fruits, and wood (Hester 1983). 299

Preserved plant remains are uncommon in Paleoindian sites as a rule, and this may be

more a matter of poor preservation than Paleoindian economics. LaBelle’s (2005)

analysis of Paleoindian grinding stones, interpreted as tools for plant processing, suggests

at least modest reliance on plant remains at a number of sites in foothill/mountain settings

and in some sites in central Texas.

There is clear evidence for plant resources playing an increasingly important role in

the diet of populations throughout the Great Plains and surrounding areas after the

Paleoindian period (Adair 1988, 2006). At a number of Early and Middle Archaic period

campsites, there is an increase in the abundance of milling and grinding stones, the first

appearance of concentrations of fire-cracked rock, and more ubiquitous recovery of

charred plant remains (Johnson 1987; Frison 1998). This increase in plant use during the

Archaic period is not associated with an apparent decrease in use of fauna. Just as with

Paleoindian sites, the highest diet diversity in Archaic assemblages is seen at residential

camp localities located in large alluvial valleys.

The earliest appearance of cultigens, such as Cucurbita pepo , Chenopodium sp. , and Iva annua , occurs in Late Archaic sites after approximately 3700 years ago (Adair

2006). Maize does not appear in the Great Plains until after A.D. 200 and it becomes economically important only after A.D. 800. These early cultigens are usually found in alluvial valley sites. Smith (1992) argues that floodplain environments were ideal environments for the earliest appearance of cultivated plants. Floodplain weedy plants, which form the core of the earliest cultigens, were pre-adapted to domestication because they rapidly colonize disturbed environments along river margins. Just as importantly, the 300

high human occupational intensity along alluvial valleys, a tendency that began in the

Paleoindian period but became more extreme as human populations grew, caused similar

disturbances to the environments, creating more opportunities for human use of these

taxa.

The initial adoption of cultigens is not related to significantly decreased residential

mobility. In fact, residential mobility significantly declines only after approximately A.D.

500. The earliest evidence for long-term sedentism is seen at Woodland period sites and is associated with more intensive exploitation of wild plants and tending of various cultigens, including maize

Grease extraction. The production of pemmican (pounded dried meat mixed with

bone grease) represents a significant increase in processing effort over cold-marrow

extraction as a means for storing animal foods. Reeves (1990) suggests that pemmican

first appeared around 4800 years ago, but that it did not become common in forager diets

until about 3000 years ago. This processing technology likely has its origins in earlier

Archaic stone boiling and/or roasting techniques associated with processing plant remains

(Reeves 1990).

Although more labor intensive, pemmican production allows hunters more reliable

access to meat and fat by extending food preservation time. Food drying practices are

older than this, but shelf-lives of simple dried products were also shorter. Reeves

(1990:170) even suggests that because pemmican freed hunters from their constant search

for fresh meat for themselves and their dogs, greater mobility was possible, over-

wintering starvation was less common, and cultures could become increasingly elaborate. 301

Even at Late Prehistoric village sites a mixture of cold-marrow and grease extraction techniques were used for processing bison carcasses (Brosowske and Bevitt 2006; Quigg

1997, 1998). Whether these different strategies are employed during different seasons or for different goals has yet to be explored fully.

SOME CLOSING THOUGHTS

After more than a century of research, we know much more about Paleoindian life ways. Some debates over Paleoindian land use and subsistence strategies seem never to be resolved because of our insistence on oversimplifying a very complex adaptive strategy. Paleoindian foragers were not strictly super-mobile, high-tech, specialized hunters or strictly sedentary, low-tech, broad-spectrum foragers. They probably practiced a mixed strategy depending on the particular circumstance in which they found themselves. The central issue in this study has been the variation in Paleoindian responses to geographic and diachronic variations in the resource structure of the Great Plains and

Rocky Mountains.

Human behavioral ecology models and wildlife observations of prey species provide important insights for understanding hunter-gatherer land use and subsistence practices.

However, they provide only a starting point, and must be compared against complete archaeological data. The collective Paleoindian database is far from complete. We need both more detailed documentation of multiple Paleoindian zooarchaeological assemblages and the investigation of new sites in poorly represented regions. More 302

information is sorely needed, for example, for zooarchaeological assemblages from

Blackwater Draw, Carter/Kerr-McGee, Jurgens, and Lindenmeier and chipped stone

assemblages from Finley, Lubbock Lake, Medicine Lodge Creek, Mummy Cave, and

Red Smoke. The geographic distribution of known Paleoindian sites is still very uneven.

Although many Paleoindian sites have been identified in the High Plains and Rocky

Mountains, there are relatively few known sites in the Missouri Plateau, Rolling Hills,

Colorado Piedmont, and Raton Volcanics. More work should be done to identify and

investigate sites in these settings in order to more fully appreciate regional variation in

life ways. Archaeologists also need better baseline information on the structure of natural

environments that prehistoric hunter-gatherers inhabited. Specifically, we need more

paleoenvironmental studies that track regional and long-term changes in economically

important species. The general lack of these kinds of information has played a significant role in our underestimation of the adaptive flexibility of Paleoindian peoples. 303

APPENDICES

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean 12 Mile Creek site kill/process P High Plains 10,503 14SG516 site camp P Osage Plains Middle Rocky 48JO202 site camp F Mountains Middle Rocky 48JO303 Pryor stemmed camp F Mountains 48SW13156 I camp F Wyoming Basin 8725 48SW8842 5 camp F Wyoming Basin 8490 48UT375 component 1 camp F Wyoming Basin 8476 48UT375 component 2 camp F Wyoming Basin 4Moore-Neal site kill/process P High Plains lithic Adair-Steadman site A Osage Plains workshop Agate Basin Area 1-Agate Basin unknown P Missouri Plateau Agate Basin Area 2-Agate Basin kill/process P Missouri Plateau 10,430 Agate Basin Area 2-Folsom camp P Missouri Plateau 10,699 Agate Basin Area 3-Hell Gap kill/process P Missouri Plateau 10,445 Allen Intermediate Zone camp A High Plains Allen Occupation Level 1 camp A High Plains 10,493 Allen Occupation Level 2 camp A High Plains Anadarko Cache site cache P Osage Plains Middle Rocky Anzick site Burial F 10,680 Mountains Arenosa Shelter Stratum 38 camp P Edward Plateau 9950 Camp A and B; Pond Aubrey camp A Osage Plains 11,588 C/E1; Red Sand S Baker Cave Zone 1 camp P Edward Plateau 8942 Southern Rocky Barger Gulch Locality B camp F 10,543 Mountains Middle Rocky Barton Gulch Adler Complex camp F 9,410 Mountains Barton Gulch Middle Rocky Barton Gulch camp F 8,780 Complex Mountains Beidleman Ranch site kill/process P Osage Plains 304

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean lithic Benz site A Missouri Plateau 8504 workshop lithic Big Black site A Missouri Plateau workshop Big Lake site kill/process P High Plains 7917 Billy Ross site unknown A Osage Plains Southern Rocky Black Mountain site camp F 10,631 Mountains Blue Point site camp F Wyoming Basin 9540 lithic Bobtail Wolf site A Missouri Plateau workshop Bonfire Shelter Stratum A and B/C kill/process P Edward Plateau 9871 Middle Rocky Bottleneck Cave I camp F Mountains Middle Rocky Bottleneck Cave II camp F Mountains Middle Rocky Bottleneck Cave III camp F Mountains Burntwood Creek site kill/process P High Plains 9085 Busse site cache P High Plains Blackwater Draw Locality 2/3 kill/process P High Plains Blackwater Draw Locality 7A TMM kill/process P High Plains North Pit Blackwater Draw Locality 12B kill/process P High Plains Blackwater Draw Locality 12C kill/process P High Plains Blackwater Draw Locality 15F kill/process P High Plains Blackwater Draw Locality 11 kill/process P High Plains 10,780 Blackwater Draw Locality 18 kill/process P High Plains Blackwater Draw Locality 18 kill/process P High Plains Blackwater Draw Mitchell Locality camp P High Plains Blackwater Draw Locality 4A kill/process P High Plains Blackwater Draw Locality 4E kill/process P High Plains Southern Rocky Caribou Lake Site Area A camp P 8359 Mountains Carter/Kerr- Cody kill/process P Missouri Plateau McGee Carter/Kerr- Folsom unknown P Missouri Plateau 10,400 McGee 305

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean Casper site kill/process P Missouri Plateau 10,025 Southern Rocky Cattle Guard site camp F Mountains Southern Rocky Chance Gulch Stratum B camp F 8044 Mountains Cherokee Sewer IIIA kill/process A Dissected Till Plains 7183 Clary Ranch site kill/process P High Plains 9040 Claussen Horizon 1 unknown P Osage Plains Claussen Horizon 2 unknown P Osage Plains Claypool Area II camp P High Plains Middle Rocky Colby site kill/process P 10,961 Mountains Cooper Lower Kill kill/process P Plains Border 10,050 Cooper Middle Kill kill/process P Plains Border Cooper Upper Kill kill/process P Plains Border Custer County site kill/process P High Plains Deadman Wash Component 2 unknown F Wyoming Basin Dent site kill/process A Colorado Piedmont 10,754 Area A-strats 22-24; Devils Mouth camp A Edward Plateau Area C Domebo site kill/process P Osage Plains 11,091 Drake site cache P Colorado Piedmont Dunn site kill/process P Dissected Till Plains Eckles SITE unknown P Plains Border False Cougar Middle Rocky IIB camp F Cave Mountains Finley site kill/process F Wyoming Basin 9009 Fletcher site kill/process P Alberta Plains 9380 Folsom site kill/process F Raton Volcanics 10,450 Fowler-Parrish site camp P Colorado Piedmont Frasca Area II kill/process P Colorado Piedmont 8908 Frazier site kill/process A Colorado Piedmont 9650 lithic Gault Clovis P Osage Plains workshop lithic Gault Folsom P Osage Plains workshop Gault Late Paleoindian unknown P Osage Plains Gordon Creek site Burial F Colorado Piedmont 9623 Granite Creek Middle Rocky Pryor stemmed camp F Rockshelter Mountains 306

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean lithic Middle Rocky Hanson site F 10,251 workshop Mountains Middle Rocky Helen Lookingbill Layer 2/4 camp F 10,405 Mountains Middle Rocky Helen Lookingbill Layer 7/9 camp F 8658 Mountains Locality I-Agate Hell Gap camp F Missouri Plateau Basin-Folsom Hell Gap Locality I-Alberta camp F Missouri Plateau Locality I-Eden- Hell Gap camp F Missouri Plateau Scottsbluff Hell Gap Locality I-Folsom camp F Missouri Plateau Locality I-Folsom- Hell Gap camp F Missouri Plateau Goshen Hell Gap Locality I-Frederick camp F Missouri Plateau Hell Gap Locality I-Goshen camp F Missouri Plateau Hell Gap Locality I-Hell Gap camp F Missouri Plateau Locality II-Agate Hell Gap camp F Missouri Plateau Basin Hell Gap Locality II-Alberta camp F Missouri Plateau Hell Gap Locality II-Frederick camp F Missouri Plateau 8600 Hell Gap Locality II-Goshen camp F Missouri Plateau 10,118

Hell Gap Locality II-Hell Gap camp F Missouri Plateau Locality III-Agate Hell Gap camp F Missouri Plateau 10,280 Basin Hell Gap Locality III-Hell Gap camp F Missouri Plateau Hell Gap Locality II-Lusk camp F Missouri Plateau Hell Gap Locality I-Unknown camp F Missouri Plateau

Hell Gap Locality V-Cody camp F Missouri Plateau 8781 Heron Eden site kill/process P Alberta Plains Horace Rivers Stratum III camp A High Plains 9109 Horn Shelter No. Stratum 5A unknown A Edward Plateau 2 Horn Shelter No. Stratum 5C unknown A Edward Plateau 2 Horn Shelter No. Stratum 5D unknown A Edward Plateau 2 307

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean Horn Shelter No. Stratum 5F unknown A Edward Plateau 2 Horn Shelter No. Stratum 5G camp A Edward Plateau 10,021 2 Horn Shelter No. Stratum 6 unknown A Edward Plateau 2 Horn Shelter No. Stratum 8A unknown A Edward Plateau 2 Horn Shelter No. Stratum 8B unknown A Edward Plateau 8400 2 Horn Shelter No. Stratum 8C unknown A Edward Plateau 2 Middle Rocky Horner I site camp F 8788 Mountains Middle Rocky Horner II site kill/process F 9630 Mountains Howard Gully site kill/process P Osage Plains 10,214 Jake Bluff site kill/process P Plains Border 10,762 James Allen site kill/process P Wyoming Basin 8405 Southern Rocky Jerry Craig site kill/process F 9390 Mountains Agate Basin-Folsom- Jim Pitts camp F Black Hills Goshen Jim Pitts Cody camp F Black Hills Jim Pitts Goshen camp F Black Hills 10,160 Johnson site camp F Colorado Piedmont Jones Miller site kill/process P High Plains 10,020 Jurgens Area 1 camp A Colorado Piedmont Jurgens Area 2 camp A Colorado Piedmont Jurgens Area 3 kill/process A Colorado Piedmont 9070 Kanorado Clovis? unknown P High Plains Locality Kanorado Folsom? unknown P High Plains Locality Kanorado Folsom? unknown P High Plains Locality Kanorado Late Paleoindian? unknown P High Plains Locality Kanorado Lower Clovis unknown P High Plains Locality 308

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean Kanorado Upper Clovis? unknown P High Plains Locality Krmpotich Site camp F Wyoming Basin Laird site kill/process P High Plains 8495 Lake Theo Folsom camp P High Plains Lake Theo Plainview unknown P High Plains 9618 Lamb Spring site kill/process P Colorado Piedmont Lange-Ferguson site kill/process P Missouri Plateau 11,140 Levi Rockshelter Zone II camp A Edward Plateau Levi Rockshelter Zone IV camp A Edward Plateau 8262 Lewisville site camp A Osage Plains Lime Creek Zone I camp A High Plains 9120 Lindenmeier Cody unknown F Colorado Piedmont Lindenmeier Folsom camp F Colorado Piedmont 10,569 Southern Rocky Linger SI1977-1979 camp F Mountains Lipscomb site kill/process P Plains Border 10,820 Lone Wolf Creek site kill/process P Osage Plains Lubbock Lake FA12-4 unknown P High Plains Lubbock Lake FA2-1 kill/process P High Plains Lubbock Lake FA2-2 kill/process P High Plains Lubbock Lake FA4-1 kill/process P High Plains Lubbock Lake FA5-12 kill/process P High Plains Lubbock Lake FA5-13 kill/process P High Plains Lubbock Lake FA5-17 kill/process P High Plains Lubbock Lake FA5-3 unknown P High Plains FA5-5/TMM Station Lubbock Lake kill/process P High Plains M Lubbock Lake FA5-7/GA5-3 kill/process P High Plains Lubbock Lake FA5-8/10 kill/process P High Plains Lubbock Lake FA6-11 kill/process P High Plains Lubbock Lake FA6-15 kill/process P High Plains Lubbock Lake FA6-3 camp P High Plains 8560 Lubbock Lake FA6-8 kill/process P High Plains Lubbock Lake FA9-1 kill/process P High Plains Lubbock Lake GA12-3 kill/process P High Plains Lubbock Lake GA12-4 kill/process P High Plains Lubbock Lake GA12-5 kill/process P High Plains Lubbock Lake GA5-2 kill/process P High Plains 309

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean Lubbock Lake TMM-Near Station D kill/process P High Plains Lubbock Lake TMM-Near Station I kill/process P High Plains lithic Middle Rocky MacHaffie II Site F 8100 workshop Mountains Mammoth lithic Middle Rocky Meadows I-Level Level II F workshop Mountains II Mammoth lithic Middle Rocky Meadows I-Level Level III F workshop Mountains III Mammoth lithic Middle Rocky Meadows I-Level Level IV F 9390 workshop Mountains IV Middle Rocky Mangus I camp F 8645 Mountains Marks Beach site kill/process P High Plains 10,301 (Gibson Ranch) Mclean site kill/process P Osage Plains Medicine Lodge Middle Rocky 23 ft-deep deer camp F 9700 Creek Mountains Medicine Lodge bone midden-22.2- Middle Rocky camp F 9590 Creek 23.3 ft Mountains Medicine Lodge Middle Rocky Cody camp F 8830 Creek Mountains Medicine Lodge Middle Rocky Foothill-Mnt camp F 9360 Creek Mountains Medicine Lodge Middle Rocky Foothill-Mnt camp F 9354 Creek Mountains Medicine Lodge Middle Rocky Fish Fauna camp F 8520 Creek Mountains Medicine Lodge Middle Rocky North Paleo camp F 8394 Creek Mountains Medicine Lodge Middle Rocky Fire Pit level camp F 8240 Creek Mountains Medicine Lodge Middle Rocky Pryor stemmed camp F 8191 Creek Mountains Medicine Lodge Middle Rocky Unknown camp F 8600 Creek Mountains Medicine Lodge Middle Rocky Unknown camp F 8760 Creek Mountains Medicine Lodge Middle Rocky Unknown camp F 9030 Creek Mountains 310

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean Medicine Lodge Middle Rocky Unknown camp F 9510 Creek Mountains Medicine Lodge Middle Rocky Unknown camp F 9620 Creek Mountains Merserve site unknown P High Plains Miami site kill/process P High Plains Midland site Burial P High Plains Mill Iron site camp P Missouri Plateau 11,168 Milnesand Site kill/process P High Plains Mona Lisa Locality A kill/process A Alberta Plains 8080 Middle Rocky Mummy Cave Layer 1 camp F Mountains Middle Rocky Mummy Cave Layer 4 camp F 9244 Mountains Middle Rocky Mummy Cave Layer 6 camp F Mountains Middle Rocky Mummy Cave Layer 7 camp F Mountains Middle Rocky Mummy Cave Layer 8 camp F 8455 Mountains Middle Rocky Mummy Cave Layer 9 camp F Mountains Middle Rocky Mummy Cave Layer 10 camp F Mountains Middle Rocky Mummy Cave Layer 11 camp F Mountains Middle Rocky Mummy Cave Layer 12 camp F Mountains Middle Rocky Mummy Cave Layer 13 camp F Mountains Middle Rocky Mummy Cave Layer 14 camp F Mountains Mustang Springs site unknown P High Plains Middle Rocky Myers-Hindman Unit 1 camp F Mountains Nall Allen/Frederick camp P High Plains 7740 Nall Angostura unknown P High Plains 6870 Nall Plainview camp P High Plains 9670 Nelson Site unknown P Colorado Piedmont 7995 Niska Site unknown P Alberta Plains 7041 311

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean Norton Site Site kill/process P High Plains 9080 Olsen-Chubbuck Site kill/process P High Plains 9395 OV Clary Lower camp P High Plains OV Clary Middle camp P High Plains 9123 OV Clary Upper camp P High Plains Packard Site camp P Osage Plains 9406 Middle Rocky Painted Rock V Pryor stemmed camp F Mountains Zone 5 (Area 3 and lithic Pavo Real A Edward Plateau 4) workshop Perry Ranch Site kill/process P Osage Plains 8460 lithic Pine Springs site F Wyoming Basin workshop Plainview site kill/process P High Plains R-6 site camp Raton Volcanics Rattlesnake Pass site kill/process F Wyoming Basin 9860 Ray Long Area A camp P Missouri Plateau 9380 Ray Long Area B camp P Missouri Plateau 7163 Ray Long Area C camp P Missouri Plateau Red Smoke Stratum 78 camp A High Plains Red Smoke Stratum 80 camp A High Plains Red Smoke Stratum 83 camp A High Plains Red Smoke Stratum 88 camp A High Plains Red Smoke Stratum 90 camp A High Plains Rex Rodgers site kill/process P Osage Plains 9391 Perez Component Richard Beene camp A Edward Plateau 8704 (Blocks H, T, N, Q) Ryan Cache site cache P High Plains 9220 Sailor-Helton site cache P High Plains San Jon (Area 2) site unknown P High Plains 8360 Middle Rocky Schiffer Cave Pryor stemmed camp F Mountains Scottsbluff site kill/process P High Plains 8809 Seminole-Rose site kill/process P High Plains Sheaman site camp P Missouri Plateau 10,690 Shifting Sands Site camp P High Plains Middle Rocky Sister's Hill site unknown P 9650 Mountains Slim Arrow site kill/process P High Plains 312

Appendix A: Summary Information for Assemblages Used in this Study Pooled Physiographic Site Component Site Function Setting 14 C Section Mean Middle Rocky Sorenson I camp F Mountains Middle Rocky Sorenson II camp F 7680 Mountains Middle Rocky Sorenson III camp F Mountains Sutter Site camp P Dissected Till Plains 7822 lithic Tim Adrian Site P Plains Border workshop Upper Twin Southern Rocky site kill/process F 10,392 Mountain Mountains Walsh site cache P Plains Border Middle Rocky Wasden Bison kill/process F Mountains Middle Rocky Wasden Mammoth camp F 11,965 Mountains Waugh site camp P Plains Border 10,393 Clovis: Units Igl, Wilson Leonard camp A Edward Plateau 11,200 Igl/Isi, Isi, Igl/Isi/Icl Early Paleoindian Bonebed: Units Isi, Wilson Leonard camp A Edward Plateau 11,200 Isi/Icl, Igl/Isi, Isi/Isi- c, Isi/Icl/Isi-c Unassigned early Wilson Leonard camp A Edward Plateau 10,567 Paleoindian Unit I/II: (Wilson) Wilson Leonard camp A Edward Plateau 9445 Burial Unit II: (Golindria- Wilson Leonard Barber, St Mary's, camp A Edward Plateau 9294 Angostura) Upper Unit I: Wilson Leonard camp A Edward Plateau 9663 (Wilson) Winger site kill/process P High Plains 9080

313

Appendix B. Summary Information on bison remains used in this study. Excavation %Cows/ Site DAG 1 Season 3 MNI%Bulls Source Area (m 2) Calves 12 Mile Creek 40 Unk 13 66 34 Hill 2002b Agate Basin-Area 1-Agate .6-.9 Fal-Spr Hill 2001 Basin Agate Basin-Area 2-Agate 293 .7-.8 Win 53 11 89 Hill 2001 Basin Agate Basin-Area 2- 186 .9 Spr 11 36 64 Hill 2001 Folsom Agate Basin-Area 3-Hell 48 .6-.7 F/W 16 Hill 2001 Gap Bamforth Allen-Intermediate Zone 129 .2-.4 Sum 2 1 2002b Bamforth Allen-Occupation Level 1 129 .2-.4 Sum 2 9 2002b Bamforth Allen-Occupation Level 2 129 .2-.4 Sum 2 2 2002b Aubrey-Camp A And B; 193 Unk 5 0 100 Ferring 2001 Pond C/E1; Red Sand S Turpin et al. Big Lake 33 .3-.4 Sum 10 0 100 1997 Johnson and Blackwater Draw-Locality 17 .6-.7 F/W 2 4 25 75 Holliday 11 1997 Blackwater Draw-Locality 12 Unk 3 Hester 1972 12c Blackwater Draw-Locality 22 Unk 3 Hester 1972 15f Blackwater Draw-Locality 4 Unk 7 Hester 1972 7a Bonfire Shelter-Stratum A Byers et al. 185 .2-.4 Sum 24 33 66 And B/C 2005 Hill et al Burntwood Creek 184 .1-.2 S/S 32 18 82 1992 Carter/Kerr-McGee-Cody 80 .6-.7 F/W 47 Frison 1984 Carter/Kerr-McGee-Folsom 34 Unk 1 Frison 1984 Todd et al. Casper 1088 .6 Fal 100 9 91 1997 Cattle Guard 1438 .3-.4 Sum 49 Jodry 1999a 314

Appendix B. Summary Information on bison remains used in this study. Excavation %Cows/ Site DAG 1 Season 3 MNI%Bulls Source Area (m 2) Calves Whittaker Cherokee Sewer-IIIa 135 .6 Fal 8 0 100 1997 Clary Ranch 194 .3-.5 S/F 41 16 84 Hill 2001 Cooper-Lower Kill 12 .3 Sum 20 25 75 Bement 1999 Cooper-Middle Kill 22 .3 Sum 29 25 75 Bement 1999 Cooper-Upper Kill 24 .3 Sum 29 25 75 Bement 1999 Todd and Hofman Finley 79 .6 Fal 75 36 64 1987; this paper Wilson and Fletcher >24 .4-.5 S/F 3 MacWilliams 1987 Folsom 286 .4-.5 S/F 32 15 85 Meltzer 2006 Fulgham and Frasca-Area II 28 .7-.8 Win 2 63 29.4 70.6 Stanford 1982 Borresen Frazier 220 W/S 44 15.4 60 2002 Hanson 215 .6-.7 F/W 2 4 Frison 1991 Hell Gap-Locality II-Agate 55 .1-.9 Spr-Win 14 Byers 2002 Basin Knell et al. Hell Gap-Locality V-Cody 47 .1-.9 Spr-Win 7 2002 Heron Eden 82 .6-.7 F/W 37 53.7 46.3 Corbeil 1995 Frison and Horner I 770 .6 Fal 158 Todd 1987 Frison and Horner II 130 .6 Fal 65 20 80 Todd 1987 Hurst and Howard Gully 24 Unk 2 Carter 2005 Bement and Jake Bluff 40 .3 Sum 15 Carter 2003 James Allen 4.6 Unk 15 11.1 88.9 Berman 1959 Hill and Jerry Craig 18 .3-.5 S/F 7 Kornfeld 1999 315

Appendix B. Summary Information on bison remains used in this study. Excavation %Cows/ Site DAG 1 Season 3 MNI%Bulls Source Area (m 2) Calves Fall, Stanford Jones Miller 600 .5-.9 Win, 300 1999 Spr Hill and Hill Jurgens-Area 1 112 .4-.6 S/F 37 2002 Hill and Hill Jurgens-Area 2 52 Unk 2 2002 Hill and Hill Jurgens-Area 3 88 .4-.6 S/F 36 2002 Hofman and Laird 4 .6-.9 Fall-Spr 2 Blackmar 1997 Baxevanis Lake Theo-Folsom 171 .4-.5 S/F 12 1997 McCartney Lamb Spring 108 .7-.9 W/S 27 70 30 1983 Alexander Levi Rockshelter-Zone II 45.5 Unk 2 1963 Alexander Levi Rockshelter-Zone IV 45.5 Unk 1 1963 Lime Creek-Zone I 148 .5 Fal 3 Jones 1999 Wilmsen and Lindenmeier-Folsom 353 Unk 13 64 36 Roberts 1978 Linger-SI1977-1979 .3-.4 Sum 5 Jodry 1999a Todd et al. Lipscomb 64 .2-.4 Sum 56 25 75 1992 Johnson and Lubbock Lake-FA2-2 3 Unk 3 Holliday 1989 Johnson and Lubbock Lake- FA 5-12 Unk 4 Holliday 1985 Johnson and Lubbock Lake- FA 5-17 22 .9-1.1 Spr 2 4 0 100 Holliday 1989 Johnson and Lubbock Lake- FA 5-3 .6 Fal 2 Holliday 1989 316

Appendix B. Summary Information on bison remains used in this study. Excavation %Cows/ Site DAG 1 Season 3 MNI%Bulls Source Area (m 2) Calves Johnson and Lubbock Lake- FA 5-7/ .9-1.1 Spr 2 6 Holliday GA5-3 1989 Johnson and Lubbock Lake- FA 5-8/10 70 .9-1.1 Spr 2 2 0 100 Holliday 1989 Johnson and Lubbock Lake-FA6-11 93 .6 Fal 6 16.7 83.3 Holliday 1980 Johnson and Lubbock Lake-FA6-15 Unk 1 Holliday 1989 Johnson and Lubbock Lake-FA6-3 120 .6 Fal 7 0 100 Holliday 1981 Johnson and Lubbock Lake-FA6-8 94 .6 Fal 3 0 100 Holliday 1989 Johnson and Lubbock Lake-FA9-1 35 .6 Fal 5 20 80 Holliday 1989 Johnson and Lubbock Lake-GA12-5 .9-1.1 Spr 2 3 Holliday 1989 Johnson and Lubbock Lake-GA5-2 Unk 1 Holliday 1989 Marks Beach (Gibson LaBelle and 158 Unk 1 Ranch) Meltzer 1996 Widga per. Merserve Unk Unk 4 comm. 2004 Mill Iron 140 .0-.2 S/S 34 Frison 1996 Hill 2002a Milnesand 60 .7-1.1 W/S 33 29 71 Mona Lisa-Locality A 30 .6-.7 F/W 2 Wilson 1974 Kornfeld et. Nelson 1 .7-.9 W/S 5 al. 2006 Hofman et Norton Site 23 .6-.8 F/W 8 al. 1995 317

Appendix B. Summary Information on bison remains used in this study. Excavation %Cows/ Site DAG 1 Season 3 MNI%Bulls Source Area (m 2) Calves Wheat 1972; Olsen-Chubbuck 75 .4 Sum 190 24 76 Frison 1991 Sum, Hill et al. OV Clary-Middle 9 .3-1.1 Win, 5 2006 W/S Hofman and Perry Ranch 7 .7 Win 2 0 100 Todd 1997 Plainview 131 .6-.9 Fal-Spr 84 47 53 This paper Smith and Rattlesnake Pass 60 Unk 2 McNees 1990 Rex Rodgers 17.5 .6-.9 F/W 6 33 67 This paper Hill et al. San Jon (Area 2) 62 Unk 5 1995 Todd et al. Scottsbluff 10 .1-.3 S/S 26 14 86 1990; this paper Collins et al. Seminole-Rose 16 .1 Spr 5 1997 Frison and Sheaman .1-.2 S/S 1 Stanford 1982 Kornfeld et Upper Twin Mountain 26 .5-.7 F/W 15 al. 1999 Hill and Waugh 30 .8-.9 W/S 6 50 50 Hofman 1997 Wilson Leonard-Early Paleoindian Bone bed: 95 Unk 2 Baker 1998 Units Isi, Isi/Icl, Igl/Isi, Isi/Isi-C, Isi/Icl/Isi-C Mandel and Hofman Winger 9 .6-.8 F/W 6 2003; Widga per. com. 2004 318

Appendix B. Summary Information on bison remains used in this study. Excavation %Cows/ Site DAG 1 Season 3 MNI%Bulls Source Area (m 2) Calves 1 DAG = dental age group 2 A DAG value for this seasonality was not provided in the original report, the DAG values used here is an estimation. 3 SPR = spring; S/S = Spring/Summer; SUM = summer; S/F = summer/fall; FAL = fall; F/W = fall/winter; WIN = winter; W/S = winter/spring; UNK = unknown 4 MNI = minimum number of individuals

319

REFERENCES CITED

Adair, Mary J. 1988 Prehistoric Agriculture in the Central Plains. University of Kansas Publication in Anthropology 16. Department of Anthropology, University of Kansas, Lawrence. 2006 Paleoethnobotanical research in Kansas. In Kansas Archaeology , edited by R.J. Hoard and W. E. Banks, pp. 248-263. University Press of Kansas, Lawrence, Kansas.

Agenbroad, Larry D., 1978 The Hudson-Meng Site: An Alberta Bison Kill in the Nebraska High Plains . University Press of America, Washington, D.C.

Agogino, George A. and E. Galloway 1965 The Sister's Hill Site: Hell Gap Site in North-Central Wyoming. Plains Anthropologist 10:190-195.

Agogino, George A., and A. Parrish 1971 The Fowler-Parrish Site: A Folsom Campsite in Eastern Colorado. Plains Anthropologist 16: 111-114.

Akersten, W.A., T.M. Foppe, and G.T. Jefferson 1988. New Source of Dietary Data for Extinct Herbivore. Quaternary Research 30: 92- 97.

Allen, C.R., A.S. Garmestani, T.D. Havlicek, P.A. Marquet, G.D. Peterson, C. Restrepo, C.A. Stow, and B.E. Weeks 2006 Patterns in Body Mass Distributions: Sifting Among Alternative Hypotheses. Ecology Letters 9: 630-643.

Allen, J. A. 1876 The American bison, living, and extinct. Memoir of the Geological Survey of Kentucky 1(pt. 2). University Press, Cambridge. 1877 History of the American bison, Bison americanus . Ninth Annual Report of the Survey for the Year 1875, United States Geological Surve y: 443-587.

Alexander, Jr., H. L. 1963 The Levi Site: A Paleo-Indian Campsite in Central Texas. American Antiquity 28: 510-528. 1982 The Pre-Clovis and Clovis Occupations at the Levi Site. In Peopling of the New World, edited by J. E. Ericson, R. E. Taylor, and R. Berger, pp. 133-146. Ballena Press, Los Altos.

320

Alroy, Jon 1998 Putting North America's End-Pleistocene Megafaunal Extinctions in Context: Large-Scale Analyses of Spatial Patterns, Extinction Rates, and Size Distributions. In Extinctions in Near Time, edited by R. MacPhee, pp. 105-143. Kluwer Academic/Plenum Publishers, New York. 1999 Putting North America's End-Pleistocene Megafaunal Extinction in Context: Large Scale Analysis of Spatial Patterns, Extinction Rates, and Size Distributions. In Extinctions in Near Time: Cause, Contexts, and Consequences , edited by MacPhee, R.D.E., pp. 105-143. Kluwer Academic/Plenum, New York. 2001 A Multispecies Overkill Simulation of the End-Pleistocene Megafaunal Mass Extinction. Science 292:1893-1896.

Amick, Daniel S. 1994 Folsom Diet Breadth and Land Use in the American Southwest. Unpublished Ph.D. dissertation, Department of Anthropology, University of New Mexico, Albuquerque. 1995 Patterns of Technological Variation among Folsom and Midland Projectile Points in the American Southwest. Plains Anthropologist 40:23-38. 1996 Regional Patterns of Folsom Mobility and Land Use in the American Southwest. World Archaeology 27:411-426. 2000 Regional Approaches with Unbounded System: The Record of Folsom Land Use in New Mexico and Texas. In The Archaeology of Regional Interactions: Religion, Warfare, and Exchange Across the American Southwest, edited by M. Hegman, pp. 119-147. University Press of Colorado, Boulder.

Anderson, A. E., D. L. Bowden , and D. E. Medin 1972 Indexing the annual fat cycle in a mule deer population. Journal of Wildlife Management 54(4): 550-556.

Anderson, D., and H. A. Semken, Jr. (editors) 1980 The Cherokee Excavations Holocene Ecology and Human Adaptations in Northwestern Iowa . Academic Press, New York.

Anderson, Elaine 1984 Who's Who in the Pleistocene: A Mammalian Bestiary. In Quaternary Extinctions: A Prehistoric Revolution , edited by P.S. Martin and R. G. Klein, pp. 40-89. The University of Arizona Press, Tucson.

321

Andrefsky, William, Jr. 1994 Raw Material Availability and the Organization of Lithic Technology. American Antiquity 56:216-234.

Arthur, G.W. 1975 An Introduction to the Ecology of early Historic Communal Bison Hunting Among the Northern Plains Indians . Archaeological Survey of Canada, Paper 37, Ottawa.

Ashton, K.G., M. C. Tracy, and A., de Queriroz 2000 Is Bergmann's Rule Valid for Mammals? The American Naturalist 156, 390-415.

Baker, B. W. 1998 Vertebrate Faunal Remains from the 1/4-inch and 1/8-inch Screens. In Wilson- Leonard: An 11,000-year Archaeological Record of Hunter-gatherers in Central Texas , vol. 5, edited by M. B. Collins, pp. 1463-1509. Studies in Archaeology 31. Texas Archaeological Research Laboratory and Archaeological Studies Program Report 10 Texas Department of Transportation, Austin.

Baker, Richard G., Glen G. Fredlund, Rolfe D. Mandel, and E. Arthur Bettis III 2000 Holocene Environments of the Central Great Plains: Multi-Proxy Evidence from Alluvial Sequences, Southeastern Nebraska. Quaternary International 67:75-88.

Baker, R. R. 1978 The Evolutionary ecology of Animal Migration . Hodder and Staughton London.

Ballard, W. B., and J. J. Whitman 1987 Marrow fat dynamics in Moose calves. Journal of Wildlife Management 51: 66- 69.

Bamforth, Douglas B. 1987 Historical documents and bison ecology on the Great Plains. Plains Anthropologist 32: 1-16. 1988 Ecology and Human Organization on the Great Plains . Plenum Press, New York. 1991 Flintknapping Skill, Communal Hunting, and Paleoindian Projectile Point Typology. Plains Anthropologist 36:309-322. 2002a High-tech Foragers? Folsom and Later Paleoindian Technology on the Great Plains. Journal of World Prehistory 16:55-98. 2002b The Paleoindian Occupation of the Medicine Creek Drainage, Southwestern Nebraska. In Medicine Creek: Seventy Years of Archaeological Investigations , edited by D. C. Roper, pp. 54-83. The University of Alabama Press, Tuscaloosa.

Bamforth, Douglas B. and Mark Becker 2000 Core/Biface Ratios, Mobility, Refitting, and Artifact Use-Lives: A Paleoindian 322

Example. Plains Anthropologist 45:273-290.

Bamforth, Douglas B., Mark Becker, and Joan Hudson 2005 Intrasite Spatial Analysis, Ethnoarchaeology, and Paleoindian Land-Use on the Great Plains: The Allen Site. American Antiquity 70:561-580.

Barbour, E. H., and C. B. Schultz 1932a The Mounted Skeleton of Bison occidentalis , and Associated Dart-Points. Bulletin of the Nebraska State Museum 32: 263-268. 1932b The Scottsbluff Bison Quarry and Its Artifacts. Bulletin of the Nebraska State Museum 34(1):283-286. 1936 Paleontological and Geological Consideration of Early Man in Nebraska. Bulletin of the Nebraska State Museum 45(1):431-449

Barlow, K. Renee, and Duncan Metcalfe 1996 Plant Utility Indices: Two Great Basin Examples. Journal of Archaeological Science 23:351-371.

Barton, C. Michael, Steven Schmich, and Steven R. James 2002 The Ecology of Human Colonization in Pristine Landscapes. In The Settlement of the American Continents: A Multidisciplinary Approach to Human Biogeography, edited by C M. Barton, G.A. Clark, D. R. Yesner, and G. A. Pearson, pp. 138-161. The University of Arizona Press, Tucson.

Baxevanis, S. E 1997 Collection Management and Analysis of the Lake Theo Folsom Bone Assemblage . Unpublished Master's thesis, Museum Science, Texas Tech University, Lubbock.

Bayham, Frank E. 1979 Factors Influencing the Archaic Pattern of Animal Utilization. Kiva 44:219-235.

Beaver, Joseph E. 2007 Paleolithic Ungulate Hunting: Simulations and Mathematical Modeling for Archaeological Inference and Explanation. Unpublished Ph.D. dissertation, Department of Anthropology, University of Arizona, Tucson.

Belovsky, G.E. 1987 Hunter-gatherer Foraging: A Linear Programming Approach. Journal of Anthropological Archaeology 6:29-76. 1988 An Optimal Foraging-Based Model of Hunter-Gatherer Population Dynamics. Journal of Anthropological Archaeology 7:329-372.

Bement, Leland C. 1999 Bison Hunting at Cooper Site: Where Lightning Bolts Drew Thundering Herds. 323

University of Oklahoma Press, Norman. 2003 Constructing the Cooper Model of Folsom Bison Kills on the Southern Plains. Great Plains Research 13:27-41.

Bement, Leland C., and Brian Carter 2003 Clovis Bison Hunting at the Jake Bluff Site, NW Oklahoma. Current Research in the Pleistocene 20:5-7

Bement, Leland C., Brian Carter, R. A. Varney, Linda Scott Cummings, and J. Byron Sudbury 2007 Paleo-environmental Reconstruction and Bio-Stratigraphy, Oklahoma Pandhandle, USA. Quaternary International 169-170:39-50.

Berger, Joel 1989 Female reproduction and its apparent evaluation by male mammals. Journal of Mammalogy 70: 347-358. 1992 Facilitation of Reproductive Synchrony by Gestation Adjustment in Gregarious Mammals: A New Hypothesis. Ecology 73:323-329.

Berger, Joel, and Steven L. Cain 1999 Reproductive Synchrony in Brucellosis-Exposed Bison in the Southern Greater Yellowstone Ecosystem and in Noninfected Populations. Conservation Biology 13:357-366.

Berger, Joel, and C. Cunningham 1994 Bison: Mating and Conservation in Small Populations . Columbia University Press, New York.

Berger, Joel, and M. Peacock 1988 Variability in Size-Weight Relationships of Bison bison . Journal of Mammalogy , 69: 618-624.

Bergmann, C. 1847 Uber die Verhaltnisse der Warmekonomie der Thiere zu ihrer Grosse. Gottingen Studien 3, 595-708.

Berman, J. E. 1959 Bison Bones from the Allen Site, Wyoming. American Antiquity 25 (1): 116-117.

Bettinger, Robert L. 1987 Archaeological Approaches to Hunter-Gatherers. Annual Review of Anthropology 16:121-142. 1991 Aboriginal Occupation at High Altitude: Alpine Villages in the White Mountains of Eastern California. American Anthropologist 93:656-679. 324

Bettinger, Robert L. and Martin A. Baumhoff 1982 The Numic Spread: Great Basin Cultures in Competition. American Antiquity 47:485-503.

Bettinger, Robert L., Ripan Malhi, and Helen McCarthy 1997 Central Place Models of Acorn and Mussel Processing. Journal of Archaeological Science 24:887-899.

Binford, Lewis R. 1968 Archaeological Perspectives. In New Perspectives in Archeology , edited by S. R. Binford, and L. R. Binford, pp. 5-32. Aldine, Chicago. 1978 Nunamuit Ethnoarchaeology . Academic Press, New York. 1980 Willow Smoke and Dog's Tails: Hunter-Gatherer Settlement Systems and Archaeological Site Formation. American Antiquity 45:4-20. 1981 Bones: Ancient Men and Modern Myths . Academic Press, New York. 1982 The Archaeology of Place. Journal of Anthropological Archaeology 1:5-31. 1983 In Pursuit of the Past: Decoding the Archaeological Record. Thames and Hudson, New York. 1987 Researching Ambiguity: Frames of Reference and Site Structure. In Methods and Theory for Activity Area Research , edited by S. Kent, pp. 449-512. Columbia University Press, New York. 2001 Constructing Frames of Reference: An Analytical Methods for Archaeological Theory Building Using Hunter-Gatherer and Environmental Data Sets . University of California Press, Berkeley.

Bird, Douglas W., and James F. O’Connell 2006 Behavioral Ecology and Archaeology. Journal of Archaeological Research 14:143-188.

Blackburn, T.M., K.J. Gaston, and N. Loder 1999 Geographic Gradients in Body Size: A Clarification of Bergmann's Rule. Diversity and Distributions 5:165-174.

Blackburn, T. M., and B.A. Hawkins, B.A., 2004 Bergmann's Rule and the Mammal Fauna of Northern North America. Ecology 27:715-724.

Blackburn, T.M., R. C. Stillwell, K.A. Young, C.W. Fox, and K.G. Ashton 2006 When Rensch Meets Bergmann: Does Sexual Size Dimorphism Change Systematically With Latitude. Evolution 60, 2004-2011.

Blackmar, Jeannette M. 2001 Regional Variability in Clovis, Folsom, and Cody Land Use. Plains 325

Anthropologist 46:65-94.

Blackmar, Jeannette M., and Jack L. Hofman 2006 The Paleoarchaic of Kansas. In Kansas Archaeology , edited by R.J. Hoard and W. E. Banks, pp. 46-75. University Press of Kansas, Lawrence, Kansas.

Blaxter, K.L. 1957 The effects of defective nutrition in farm livestock. Proceedings of the Nutrition Society (England and Scotland) 16: 52-58

Blurton Jones, N.G. 1987 Tolerated Theft, Suggestions about the Ecology and Evolution of Sharing, Hoarding, and Scrounging. Social Science Information 26:31-54.

Blurton Jones, N. G. K. Hawkes, and P. Draper 1994 Foraging Returns of !Kung Adult and Children: Why Didn’t !Kung Children Forager. Journal of Anthropological Research 50:217-248.

Bonnichsen, Robson, Dennis J. Stanford, and J. L. Fastook 1987 Environmental Change and Developmental History of Human Adaptive Patterns: the Paleoindian Case. In North America and Adjacent Oceans During the Last Deglaciation , edited by W. F. Ruddiman and H. E. Wright, pp. 403-424. Geological Society of America, Boulder.

Borresen, Jennifer A. 2002 A Faunal Analysis of the Frazier Site, An Agate Basin-Age Bison Kill-Butchery Site in Northeastern Colorado. Unpublished Master's thesis, Department of Anthropology, University of Tennessee, Knoxville.

Bousman, C. Britt, B. W. Baker, and A. C. Kerr 2004 Paleoindian Archeology in Texas. In The Prehistory of Texas , edited by T. K. Pertula, pp. 15-97. Texas A & M University Press, College Station.

Boutton, T.W., L.C Nordt, and D.D. Kuehn 1998 Late Quaternary Vegetation and Climate Change in the North American Great Plains: Evidence from δ13 C of the Palaeosol Organic Carbon. In Isotope Techniques in the Study of Environmental Change: Proceedings of an International Symposium on Isotope Techniques in the Study of Past and Current Environmental Changes in the Hydrosphere and the Atmosphere. pp. 653-662. International Atomic Energy Agency, Vienna.

Bozell, John R. 1995 Culture, Environment, and Bison Populations on the Late Prehistoric and Early Historic Central Plains. Plains Anthropologist 40:145-163. 326

2006 Plains Woodland Complexes of Western Kansas, and Adjacent Portions of Nebraska and Colorado. In Kansas Archaeology, edited by R. J. Hoard and W. E. Banks, pp. 93-104. University Press of Kansas, Lawrence, Kansas.

Bozell, John R. and John Ludwickson 1998 Central Plains Tradition. In Archaeology of Prehistoric Native America: An Encyclopedia , edited by G. Gibbson, pp. 556-557. Garland Publishing, New York.

Brain, C. K. 1981 The Hunters or the Hunted?: Introduction to African Cave Taphonomy. The University of Chicago Press, Chicago.

Branch, E. Douglas 1929 The Hunting of the Buffalo. D. Appleton and Company, New York.

Brantingham, P. Jeffrey 2003 A Neutral Model of Stone Raw Material Procurement. American Antiquity 68: 487-509.

Bright, Jason, Andrew Ugan, and Lori Hunsaker 2002 The Effect of Handling time on Subsistence Technology. World Archaeology 34:164-181.

Brink, John W. 1997 Fat content in leg bones of Bison bison , and Applications to Archaeology. Journal of Archaeological Science 24: 259-274.

Brink, John W. and Maureen Rollans 1990 Thoughts on the Structure and Function of Drive Lanes Systems at Communal Buffalo Jumps. In Hunters of the Recent Past , edited by L.B. Davis and B. O. K. Reeves, pp. 152-167. Unwin Hyman, London.

Broughton, Jack M. 1994 Declines in Mammalian Foraging Efficiency during the Late Holocene, San Francisco Bay. Journal of Anthropological Archaeology 13:371-401. 2002 Prey Spatial Structure and Behavior Affect Archaeological Tests of Optimal Foraging Models: Examples from Emeryville Shellmound Vertebrate Fauna. World Archaeology 34:60-83.

Broughton, J. M. and F. E. Bayham 2003 Showing Off, Foraging Models, and the Ascendance of Large-Game Hunting in the California Middle Archaic. American Antiquity 68:783-789.

327

Broughton, Jack M. and Donald K. Grayson 1993 Diet Breadth, Adaptive Change, and the White Mountains Faunas. Journal of Archaeological Science 20:331-336.

Brown, J.H., and P.F. Nicoletto 1991 Spatial Scaling of Species Composition: Body Masses of North American Land Mammals. The American Naturalist 138:1478-1512.

Browsowske, Scott D. and C. Tod Bevitt 2006 Looking South: The Middle Ceramic Period in Southern Kansas and Beyond. In Kansas Archaeology , edited by R.J. Hoard and W. E. Banks, pp. 180-205. University Press of Kansas, Lawrence, Kansas.

Buehler, Kent J. 1997 Where's the Cliff? Late Archaic Bison Kills in the Southern Plains. Plains Anthropologist 42, Memoir No. 29: 63-83

Bugalho, M. N., J. A. Milne, and P. A. Racey 2001 The Foraging Ecology of Red Deer ( Cervus elaphus) in the Mediterranean Environment: Is a Larger Body Size Advantageous? Journal of Zoology, London 255:285-289.

Burger, Oskar, Marcus J. Hamilton, and Robert Walker 2005 The Prey as Patch Model: Optimal Handling of Resources with Diminishing Returns. Journal of Archaeological Science 32:1147-1158.

Burgett, Galen R. 1990 The Bones of the Beast: Resolving Questions of Faunal Assemblage Formation Processes through Actualistic Research. Unpublished Ph.D. dissertation, Department of Anthropology, University of New Mexico, Albuquerque.

Butler, B. R., H. Gildersleeve, and J. Sommers 1971 The Wasden Site Bison: Sources of Morphological Variation. In Aboriginal Man and Environments on the Plateau of Western North America , edited by A.H. Stryd, and R.A. Smith, pp. 126-152. The University of Calgary Archaeological Association, Calgary.

Butler, Virginia L. and Susan K. Campbell 2004 Resource Intensification and Resource Depression in the Pacific Northwest of North America: A Zooarchaeological Review. Journal of World Prehistory 18:327-405.

328

Byerly, Ryan M., Judith R. Cooper, David J. Meltzer, Matthew E. Hill, and Jason M. LaBelle 2005 On Bonfire Shelter (Texas) as a Paleoindian Jump: An Assessment using GIS and Zooarchaeology. American Antiquity , 70:595-629. 2007 A Further Assessment of Paleoindian Site-Use at Bonfire Shelter. American Antiquity 72(2):373-381.

Byers, David A. 2002 Paleoindian Fat-Seeking Behavior: The Hell Gap Site Locality II Agate Basin Faunal Assemblage. Plains Anthropologist 47:359-377.

Byers, David A., C. S. Smith, and J. M. Broughton 2005 Holocene Artiodactyl Populations Histories and Large Game Hunting in the Wyoming Basin, USA. Journal of Archaeological Science 32:125-142.

Byers, David A. and Andrew Ugan 2005 Should We Expect Large Game Specialization in the Late Pleistocene? An Optimal Foraging Perspective on Early Paleoindian Prey Choice. Journal of Archaeological Science 32:1624-1640

Byers, F.M., and G.T. Schelling 1988 Nutrition in Growth. In: The Ruminant Animal Digestive Physiology and Nutrition, edited by D.C. Church, pp. 437-447. Waveland Press, Inc. Prospect Heights.

Caldwell, J. R. 1958 Trends and Traditions in the Prehistory of the Eastern United States . Scientific Paper 10, Illinois State Museum, Springfield

Calef, G. W., and J. Van Camp 1987 Seasonal distribution, group size and structure, and movements of bison herds. Canadian Wildlife Service Occasional Paper 63: 15-20.

Cannon, Kenneth P. and M. B. Cannon 2004 Zooarchaeological and Wildlife Management in the Greater Yellowstone Ecosystem. In Zooarchaeology and Conservation Biology , edited by R. L. Lyman, and K. P. Cannon, pp. 45-60. The University of Utah Press, Salt Lake City.

Cannon, Michael D., 2000 Large Mammals Relative Abundance in Pithouse and Pueblo Period Archaeofaunas from Southwestern New Mexico: Resource Depression among the Mimbres-Mogollon. Journal of Anthropological Archaeology 19:317-347. 2004 Geographic Variability in North American Mammal Community Richness during 329

the Terminal Pleistocene. Quaternary Science Review 23:1099-1123.

Cannon, Michael D., and David J. Meltzer 2004 Early Paleoindian Foraging: Examining the Faunal Evidence for Large Mammal Specialization and Regional Variability in Prey Choice. Quaternary Science Review 23:1955-1987.

Carbyn, L. N., and T. Trottier 1987 Responses of bison on their calving grounds to predation by wolves in Wood Buffalo National Park. Canadian Journal of Zoology 65: 2072-2078. 1988 Description of wolf attacks on bison calves in Wood Buffalo National Park. Arctic 41: 297-302.

Cassells, E. Steve 1983 The Archaeology of Colorado. Johnson Books, Boulder.

Catlin, George 1965 Letters and Notes on the Manners, Customs, and Conditions of North American Indians, 2 volumes . Rossand Haines, Minneapolis.

Cederlund, G. N., R. L. Bergstrom, K. Danell 1989 Seasonal variation in mandible marrow fat in moose. Journal of Wildlife Management 53: 587-592.

Chapman, J.A. and G.A. Feldhamer 1982 Wild Mammals of North America Biology, Management, and Economics . John Hopkins University Press, Baltimore.

Charnov, Eric L. 1976 Optimal Foraging, the Marginal Value Theorem. Theoretical Population Biology 9:129-136.

Charnov, Eric L. , Gordon H. Orians, and Kim Hyatt 1976 Ecological Implications of Resource Depression. The American Naturalist 110:247-259.

Christiansen, P., 2002 Locomotion in Terrestrial Mammals: The Influence of Body Mass, Limb Length and Bone Proportions on Speed. Zoological Journal of the Linnaean Society 136:685-714.

Christiansen, P. and J.M. Harris 2005 Body Size of Smilodon (Mammalia: Felidae). Journal of Morphology 266, 369- 384. 330

Collins, Michael B. 1976 Terminal Pleistocene Cultural Adaptation in Southern Texas. In Habitat Humains Antérieurs a L'Holocène en Amérique , edited by J. B. Griffin, pp. 102-135. Union Internationale des Sciences Préhistoriques et Protohistoriques. Nice, France. 1998 (editor) Wilson-Leonard: An 11,000-year Archaeological Record of Hunter- gatherers in Central Texas , 5 Volumes. Studies in Archaeology 31 Texas Archaeological Research Laboratory and Archaeological Studies Program Report 10 Texas Department of Transportation, Austin. 2002 The Gault Site, Texas and Clovis Research . Athena Review 3: 24-36.

Collins, Michael B., D.B. Hudler, and S.L. Black 2003 Pavo Real (41BX52): A Paleoindian and Archaic Camp and Workshop on the Balcones Escarpment, South-Central Texas . Texas Archeological Research Laboratory, University of Texas, and Texas Department of Transportation, Austin.

Collins, Michael B., Dennis J. Stanford, Jack L. Hofman, Margaret A. Jodry, R. O. Rose, Lawrence C. Todd, K. Kilber, and Jeannette M. Blackmar 1997 Cody Down South: The Seminole-Rose Site In West Texas. Current Research in the Pleistocene 14:15-18.

Coltman, D.W., P. O'Donohue, J.T. Jorgenson, J.T. Hogg, C. Strobeck, and M. Festa- Bianchet 2003 Undesirable Evolutionary Consequences of Trophy Hunting. Nature 426:655-658.

Coltrain, J.B., J.M. Harris, T.E. Cerling, J.R. Ehleringer, M-D. Dearing, J. Ward, and J. Allen 2004 Rancho La Brea Stable Isotope Biogeochemistry and its Applications for Palaeoecology of late Pleistocene Coastal Southern California. Palaeogeography, Palaeoclimatology, and Palaeoecology 205:199-219.

Cook, Harold J. 1925 Definitive Evidence of Human Artifacts in the American Pleistocene . Science 62: 459-460.

Corbeil, Marcel R. 1993 1992 Happenings at the Heron Eden Site (EeOi-11). Saskatchewan Archaeological Society Newsletter 14(1): 8. 1995 The Archaeology and Taphonomy of the Heron Eden Site, Southwestern Saskatchewan. Unpublished M.A. thesis, Department of Anthropology and Archaeology, University of Saskatchewan, Saskatoon.

331

Creel, Darrell, Robert F. Scott IV, and Michael Collins 1990 A Faunal Record from West Central Texas and its Bearing on Late Holocene Bison Population Changes in the Southern Plains. Plains Anthropologist 35:55- 69.

Cronk, Lee 1991 Human Behavioral Ecology. Annual Review of Anthropology 20:25-53.

Crook, W. W., Jr. and R. K. Harris 1957 Hearths and Artifacts of Early Near Lewisville, Texas and Associated Faunal Material. Bulletin of the Texas Archaeological Society 28:7-97.

Daleszczyk, Katarzyna 2004 Mother-Calf Relationships and Maternal Investment in European bison Bison bonasus. Acta Theriologica 49:555-566.

Damuth, J., and B. J. MacFadden (editors.) 1990 Body Size in Mammalian Paleobiology: Estimation and Biological Implications. Cambridge University Press, New York.

Dary, D. 1974 The Buffalo Book: The Full Saga of the American Animal . Swall Press/Ohio University Press, Athens.

Davis, E. Mott 1953 Recent Data from Two Paleoindian Sites on Medicine Creek, Nebraska. American Antiquity 18:380-386. 1962 Archaeology of the Lime Creek Site in Southwestern Nebraska. Special Publication No. 3. University of Nebraska State Museum, Lincoln.

Davis, J. L., P. Valkenburg, and D. J. Reed 1987 Correlation and depletion patterns of marrow fat in caribou bones. Journal of Wildlife Management 51: 365-371.

Davis, Leslie B., S. A. Aaberg, W. P. Eckerle, J. W. Fisher, and S. T. Greiser 1989 Montane Paleoindian Occupation of the Barton Gulch Site, Ruby Valley, Southwestern Montana . Current Research in the Pleistocene 6:7-9.

Davis, Leslie B. and S. T. Grieser 1992 Indian Creek Paleoindians: Early Occupation of the Elkhorn Mountains' Southeast Flank, West Central Montana . In Ice Age Hunters of the Rockies, edited by D. J. Stanford and J. S. Day, pp. 225-283, University Press of Colorado, Boulder.

332

Davis, Owen K., J. C. Sheppard, and S. Robertson 1986 Contrasting Climatic Histories for the Snake River Plains, Idaho, Resulting from Multiple Thermal Maximum. Quaternary Research 26:321-339.

Dayan, T., D. Simberloff, E. Tchernov, and Y. Yom-Tov, 1991 Calibrating the Paleothermometer: Climate, Communities, and the Evolution of Size. Paleobiology 17:189-199.

Dean, R. M. 2001 Social Change and Hunting During Pueblo III to Pueblo IV Transition, East- Central Arizona. Journal of Field Archaeology 28:271-285.

DelGiudice, Glenn D., Francis J. Singer, Ulysses S. Seal, and Gillian Bowser 1994 Physiological Responses of Yellowstone Bison to Winter Nutritional Deprivation. Journal of Wildlife Management 58:24-34.

Depperschmidt, J. D., S. C. Torbit, A. W. Alldredge, and R. D. Deblinger 1987 Body Condition Indices for starved pronghorns. Journal of Wildlife Management 51: 675-678.

Dibble, D. S., and D. Lorrain 1968 Bonfire Shelter: A Stratified Bison Kill Site, Val Verde County, Texas. Miscellaneous Paper No. 1. Texas Memorial Museum, The University of Texas, Austin.

Dick, Harold W., and Bert Mountain 1960 The Claypool Site: A Cody Complex Site in Northeastern Colorado. American Antiquity 26: 223-235.

Dickinson, C. E., and J. L. Dodd 1976 Phenological Pattern in the Shortgrass Prairie. American Midland Naturalist 96:367-378.

Dillehay, T. D. 1974 Late Quaternary Bison Population Changes on the Southern Plains. Plains Anthropologist 19:180-196.

Drass, Richard R. 1998 The Southern Plains Villagers. In Archaeology on the Great Plains , edited by W. R. Wood, pp. 415-455. University Press of Kansas, Lawrence, Kansas.

Driver, Jonathan C. 1990 Meat in Due Season: The Timing of Communal Hunts. In Hunters of the recent Past, edited by L. B. Davis and B. O. K. Reeves, pp. 1-33. Unwin Hyman, 333

London.

Dryden, F.D. and J.A. Marchello 1970 Influences of total lipid and fatty acid composition upon palatability of three bovine Muscles. Journal of Animal Science 31: 36-41.

Duffield, L. F. 1973 Aging and Sexing the Post-cranial Skeleton of Bison . Plains Anthropologist 18, 132-139.

Eastabrook, George F. and Arthur E. Dunham 1976 Optimal Diet as a Function of Absolute Abundance, Relative Abundance, and Relative Value of Available Prey. The American Naturalist 110:401-413.

Ebell, S. B. 1988 The Dunn Site. Plains Anthropologist 33(122): 505-530.

Eckels, C.H. 1916 The nutrients required to develop the bovine fetus. Missouri Agricultural Experimental Station Research bulletin No. 26

Edwards, W. E., 1967 The Late-Pleistocene Extinction and Diminution of Many Mammalian Species. In Pleistocene Extinctions: The Search for a Cause, edited by P.S Martin and H.E. Wright, pp. 141-154. Yale University Press, New Haven.

Egerton, P. J. M. 1962 Bison behavior. American Zoologist 2: 406.

Elston, R. G. and D. W. Zeanah 2002 Thinking Outside the Box: A New Perspective on Diet Breadth and Sexual Division of Labor in Prearchaic Great Basin. World Archaeology 34:103-130.

Emerson, A. M. 1993 The Role of Body Part Utility in Small-scale Hunting Under Two Strategies of Carcass Recovery. In From Bones to Behavior: Ethnoarchaeology and Experimental Contributions to the Interpretation of Faunal Remain , edited by J. Hudson, pp 138-155. Center for Archeological Investigations, Occasional Papers No. 21.

Emlen, J.M. 1966 The Role of Time and Energy in Food Preference. The American Naturalist 100:611-617.

334

Empel, W., and T. Roskosz 1963 Das Skellett der Gliedmassen des Wisents, Bison bonasus (Linnaeus, 1758). Acta Theriologica 7:259-297.

Engen, Steinar, and Nils. S. Stenseth 1989 Age-Specific Optimal Diets and Optimal Foraging Tactics: A Life-Historic Approach. Theoretical Population Biology 36: 281-295.

Everitt, J. H., M. A. Alaniz, and A. H. Gerbermann 1982 Chemical composition of Native Range Grasses Growing on Saline Soils of the South Texas Plains. Journal of Range Management 35: 43-46.

FAUNMAP Working Group 1994 FAUNMAP: A Database Documenting Late Quaternary Distribution of Mammal Species in the United States. Illinois State Museum Scientific Papers 25. Illinois State Museum, Springfield.

Fawcett, William B. 1987 Communal Hunts, Human Aggregations, Social Variation, and Climate Change: Bison Utilization by Prehistoric Inhabitants of the Great Plains . Ph.D. Dissertation, University of Massachusetts, University Micofilm International, Ann Arbor, MI.

Fenneman, Nevin M. 1928 Physiographic Divisions of the United States. Annals of the Association of American Geographers 18:261-353. 1931 Physiography of Western United States. McGraw-Hill Book Company, Inc., New York.

Feranec, Robert S. 2004 Isotopic Evidence of Saber-tooth development, growth rate, and diet from the adult canine of Smilodon fatalis from Rancho La Brea. Palaeogeography, Palaeoclimatology, Palaeoecology 206:303-310.

Feranec, Robert S., and Bruce J. MacFadden 2000 Evolution of the Grazing Niche in Pleistocene Mammals from Florida: Evidence from Stable Isotopes. Palaeogeography, Palaeoclimatology, Palaeoecology 162:155-169.

Ferring, C. Reid 2001 The Archaeology and Paleoecology of the Aubrey Clovis Site (41DN479) Denton County, Texas . University of North Texas. Submitted to U.S. Army Corp of Engineers, Fort Worth District. Fort Worth

335

Festa-Bianchet, M., D.W. Coltman, L. Turellli, and J.T. Jorgenson 2004 Relative Allocation to Horn and Body Growth in Bighorn Rams Varies with Resource Availability. Behavioral Ecology 15:305-312.

Fiedel, Stuart J. 2004 Rapid Migrations by Arctic Hunting Peoples. In The Settlement of the American Continents: A Multidisciplinary Approach to Human Biogeography, edited by C.M. Barton, G.A. Clark, D.R. Yesner, and G.A. Pearson, pp. 79-84. The University of Arizona Press, Tucson.

Figgins, J. D. 1933 A Further Contribution to the Antiquity of Man in America. Colorado Museum of Natural History Proceeding No. 12, Denver.

Flannery, Kent V. 1969 Origins and Ecological Effects of Early Domestication in Iran and the Near East. In The Domestication and Exploitation of Plants and Animals , edited by P.J. Ucko and G.W. Dimbley, pp. 73-100. Aldine, Chicago.

Forbis, R. G. 1968 Fletcher: A Paleo-Indian Site in Alberta. American Antiquity 33:1-10.

Forbis, R. G., and J. Sperry 1952 An Early Man Site in Montana. American Antiquity 18: 127-133.

Fortin, Daniel, John M. Fryxell, Lloyd O'Brodovich, and Dan Frandsen 2003 Foraging ecology of Bison at the Landscape and Plant Community Levels: the Applicability of Energy Maximization Principles. Oecologia 134:219-227.

Frison, George C. 1974 The Casper Site: A Hell Gap Bison Kill on the High Plains . Academic Press, New York. 1976 The Chronology of Paleoindian and Altithermal Cultures in the Big Horn Basin, Wyoming. In Cultural Change and Continuity, Essays in Honor of James Bennett Griffin , edited by C. E. Cheland, pp. 147-173. Academic Press, New York. 1978 Animal Population Studies and Cultural Inference. Plains Anthropologist Memoir 15:44-52. 1982 Paleo-Indian Winter Subsistence Strategies on the High Plains. In Plains Indian Studies: A Collection of Essays in Honor of J. C. Ewers and Waldo Wedel , edited by D. H. Ubelaker and H. J. Viola, pp. 193-201. Smithsonian Contribution to Anthropology No. 3. Smithsonian Institution, Washington, D.C. 1984 The Carter/Kerr-McGee Paleoindian Site: Cultural Resource Management and Archaeological Research. American Antiquity 49:288-314. 1991 Prehistoric Hunters of the High Plains , 2nd ed. Academic Press, San Diego. 336

1992 The Foothills-Mountain and the Open Plains: The Dichotomy in Paleoindian Subsistence Strategies between two Ecosystems. In Ice Age Hunters of the Rockies , edited by D. J. Stanford and J. S. Day, pp. 323-342. Denver Museum of Natural History or University Press of Colorado, Niwot. 1996 (editor) The Mill Iron Site . University of New Mexico Press, Albuquerque. 1998 The Northwestern and Northern Plains Archaic. In Archaeology on the Great Plains , edited by W. R. Wood, pp. 140-172. University Press of Kansas, Lawrence, Kansas. 2004 Survival By Hunting: Prehistoric Human Predators and Animal Prey . University of California Press, Berkley.

Frison, George C. and D. C. Grey 1980 Pryor Stemmed: A Specialized Paleo-Indian Ecological Adaptation. Plains Anthropologist 25:27-46.

Frison, George C. and Charles A. Reher 1970 Appendix 1: Age Determination of Buffalo by Teeth Eruption and Wear. In The Glenrock Buffalo Jump, 48CO304 , edited by G.C. Frison, pp. 46-50. Plains Anthropologist Memoir 7.

Frison, George C. and Dennis J. Stanford 1982 The Agate Basin Site: A Record of the Paleoindian Occupation of the Northwestern High Plains . Academic Press, New York.

Frison, George C., and Lawrence C. Todd (editors) 1986 The Colby Mammoth Site: Taphonomy and Archaeology of a Clovis Kill in Northern Wyoming. University of New Mexico Press, Albuquerque. 1987 The Horner Site: The Type Site of the Cody Cultural Complex . Academic Press, Orlando.

Fulgham, T., and Dennis J. Stanford 1982 The Frasca Site: A Preliminary Report. Southwestern Lore 48:1-9.

Fuller, W. A. 1960 Behaviour and Social Organization of the Wild Bison of Wood Buffalo National Park, Canada. Arctic 13:2-19.

Galm, Jerry, and Jack L. Hofman 1984 The Billy Ross Site: An Analysis of a Dalton Component from the Southern Arkansas Basin of Eastern Oklahoma. Bulletin of the Oklahoma Anthropological Society 33: 37-73.

Garretson, M. 1938 The American bison: The Story of its Extermination as a Wild Species and its 337

Restoration under Federal protection . New York Zoological Society, New York.

Geist, V. 1971 The Relation of Social Evolution and Dispersal in Ungulates during the Pleistocene, with Emphasis on the Old World Deer and the Genus Bison. Quaternary Research 1:283-315. 1987 On Speciation in Ice Age Mammals, with Special Reference to Cervids and Caprids. Canadian Journal of Zoology 65:1067-1084. 1989 Environmentally Guided Phenotype Plasticity in Mammals and Some of its Consequences to Theoretical and Applied Biology. In Alternative Life-History Styles of Animals , edited by M.N. Bruton, pp. 153-176. Kluwer Academic Publishers, Dordrecht. 1999 Adaptive Strategies in American Mountain Sheep. In Mountain Sheep of North America, edited by R. Valdez and P. R. Krausman, pp. 192-208. The University of Arizona Press, Tucson.

Gero, J. 1993 The Social World of Prehistoric Facts: Gender and Power in Paleoindian Research. In Women in Archaeology, A Feminist Critique , edited by H. du Cross and L. Smith, pp. 31-40. Occasional Papers in Prehistory No. 23, Department of Prehistory, Australian National University, Canberra.

Gilmore, Melvin R. 1931 Methods of Indian Buffalo Hunts. Papers of the Michigan Academy of Science, Arts, and Letters 16:17-32.

Goodyear, Albert C. 1989 A Hypothesis for the Use of Cryptocrystalline Raw Materials Among Paleoindian Groups of North America. In Eastern Paleoindian Lithic Resource Use, edited by Christopher Ellis and Jonathan Lothrop, pp. 1-9. Westview Press, Boulder.

Gould, Richard A. 1980 Living Archaeology . Cambridge University Press, Cambridge.

Graham, R. W. 1986 Plant-Animal Interactions and Pleistocene Extinctions. In Dynamics of Extinction , edited by D.K. Elliott, pp. 131-154. John Wiley & Sons, Inc. New York.

Graham, Russell W. and James I. Mead 1987 Environmental Fluctuations and Evolution of Mammalian Faunas during the last Deglaciation in North America. In North American and Adjacent Oceans during the last Deglaciation , edited by W.F. Ruddiman and H.E. Wright, pp. 371-402. Geological Society of America, Boulder.

338

Grayson, Donald K. 1984 Nineteenth-Century Explanations of Pleistocene Extinctions: A Review and Analysis. In Quaternary Extinctions: A Prehistoric Revolution, edited by P.S. Martin and R. G. Klein, pp. 5-39. The University of Arizona Press, Tucson, 1991 Late Pleistocene Extinctions in North America: Taxonomy, Chronology, and Explanations. Journal of World Prehistory 5:193-232. 1993 The Desert's Past: A Natural History of the Great Basin. Smithsonian Institution Press, Washington. D.C. 2001 The Archaeological Record of Human Impacts on Animal Populations. Journal of World Prehistory 15:1-68.

Grayson, D. K. and F. Delpech 1998 Changing Diet Breadth in the Early Upper Paleolithic of Southwestern France. Journal of Archaeological Science 25:1119-1129.

Grayson, Donald K. and David J. Meltzer 2002 Clovis Hunting and Large Mammal Extinction: A Critical Review of Evidence. Journal of World Prehistory 16:313-359. 2003 A Requiem for North American Overkill. Journal of Archaeological Science 30:585-593.

Green, W. C. H. 1986 Age-related difference in nursing behavior among American bison cows ( Bison bison ). Journal of Mammalogy 67: 739-741.

Green, W. C. H., and J. Berger 1990 Maternal investment in sons and daughters: Problems of methodology. Behavioral Ecology Sociobiology 27: 99-102.

Green, W.C.H., and A. Rothstein 1991 Trade-offs between Growth and Reproduction in Female Bison. Oecologia 18: 521-527. 1993 A synchronous parturition in bison. Journal of Mammalogy 74: 920-925.

Greiser, Susan T. 1985 Predictive Models of Hunter-Gatherer Subsistence and Settlement Strategies on the Central High Plains . Plains Anthropologist Memoir No. 20, Part 2. Li

Gremillion, K.J. 2002 Foraging Theory and Hypothesis Testing in Archaeology: An Exploration of Methodological Problems and Solutions. Journal of Anthropological Archaeology 21:142-164.

339

Guthrie, R. Dale 1980 Bison and Man in North America. In The Ice-Free Corridor and Peopling the New World , edited by N.W. Rutter and C.E. Schweger. Canadian Journal of Anthropology 1:55-73. 1984 Alaskan Megabucks, Megabulls, and Megarams: The Issue of Pleistocene Gigantism. In Contributions in Quaternary Vertebrate Paleontology: A Volume in Memorial to John E. Guilday , edited by H. H. Genoways, , M. R. Dawson, pp. 482-510. Special Publication in Carnegie Museum of Natural History No. 8. Pittsburgh. 1990 Frozen fauna of the Mammoth Steppe: The Story of Blue Babe. University of Chicago Press, Chicago. 1995 Mammalian Evolution in Response to the Pleistocene-Holocene Transition and the Break-up of the Mammoth Steppe: Two Case Studies. Acta Zoologica Cracoviensia . 38:139-154. 2003 Rapid Body Size Decline in Alaska Pleistocene Horses Before Extinction. Nature 426:169-171.

Hajic, E.R., R.D. Mandel, J.H. Ray, and N.H. Lopinot 1998 Geomorphology and Geoarchaeology. In The 1997 Excavations at the Big Eddy Site (23CE426) in Southwest Missouri , edited by N.H. Lopinot, J.H. Ray, and M.D. Conner, pp.. Special Publication No. 2, Center for Archaeological Research. Southwest Missouri State University, Kansas City.

Hall, Roberta L., Donna McCarthy, and Don A. Hall 2002 How wee North American Sites of the Pleistocene-Holocene Transition Discovered. North American Archaeologist 23:145-156.

Halloran, A. F. 1968 Bison (bovidae) productivity on the Wichita Mountain Refuge, Oklahoma. The Southwestern Naturalist 31: 23-26.

Hames, R. B. and W. T. Vickers 1982 Optimal Diet Breadth Theory as Model to Explain Variability in Amazonian Hunting. American Ethnologist 9:358-378.

Hammatt, H. H. 1970 A Paleo-Indian Butchery Kit. American Antiquity 35: 141-152.

Hannus, L. Adrian 1990 The Lange-Ferguson Site: A Case for Mammoth Bone Butchering Tools. In Megafauna and Man--Discovery of America's Heartland, edited by L. D. Agenbroad, J. I. Mead, and L. W. Nelson, pp. 86-99. Scientific Papers. No. 1. Mammoth Site of Hot Springs. Hot Springs, South Dakota.

340

Hansard, S. L., and R. K. Berry 1969 Fetal Nutrition. In Animal Growth and Nutrition , edited E.S.E. Hafez and I.A. Dyer, pp. 40-59. Lea and Febiger, Philadelphia.

Hanson, J. R. 1984 Bison ecology in the Northern Plains and a reconstruction of bison patterns for the North Dakota regions. Plains Anthropologist 29: 93-114.

Harrison, B. R., and K. L. Killen 1978 Lake Theo: A Stratified, Early Man Bison Butchering and Camp Site, Briscoe County, Texas. Special Archaeological Report No. 1, Panhandle-Plains Historical Museum, Canyon, Texas.

Hartwell, William T. 1995 The Ryan's Site Cache: Comparisons to Plainview. Plains Anthropologist 40:165- 184.

Haspel, Harold, and George C. Frison 1987 The Finley Site Bison Bone. In The Horner Site: The Type Site of the Cody Cultural Complex , edited by G. C. Frison and L. C. Todd, pp. 475-492. Academic Press, Orlando.

Haugen, A. O. 1974 Reproduction in the plains bison. Iowa State Journal of Research 49: 1-8.

Hawkes, Kristen 1991 Showing Off: Tests of an Hypothesis about Men’s Foraging Goals. Ethology and Sociobiology 12:29-54. 1996 Foraging Differences Between Men and Women: Behavioral Ecology of the Sexual Division of Labor. In The Archaeology of Human Ancestry: Power, Sex, and Tradition , edited by J. Steele and S. Shennan, pp. 283-305. Routledge, London.

Hawkes, K. and R. Bliege Bird 2002 Showing Off, Handicap Signaling, and the Evolution of Men's Work. Evolutionary Anthropology 11:58-67.

Hawkes, Kristen, Kim Hill, and James F. O'Connell 1982 Why Hunter Gather: Optimal Foraging and the Ache of Eastern Paraguay. American Ethnologist 9:379-398.

Hawkes, Kristen, James F. O'Connell, and N. G. Blurton Jones 1995 Hadza Children's Foraging: Juvenile Dependency, Social Arrangements, and Mobility among Hunter-Gatherers. Current Anthropology 36:688-700. 341

Hawkes, Kristen, James F. O'Connell, and L. Rogers 1997 The Behavioral Ecology of Modern Hunter-Gathers, and Human Evolution. Tree 12:29-32.

Haynes, Gary 2002a The Catastrophic Extinction of North American Mammoths and Mastodons. World Archaeology 33:391-416. 2002b The Early Settlement of the Americas: The Clovis Era. Cambridge University Press, Cambridge. 2006 Mammoth Landscapes: Good Country for Hunter-Gatherers. Quaternary International 142-143:20-29.

Haynes, Jr., C. Vance, 1995 Geochronology of Paleoenvironmental Change, Clovis Type Site, Blackwater Draw, New Mexico. Geoarchaeology 10:317-388.

Haynes, C. Vance, M. MacFaul, R. H. Brunswig, and K. D. Hopkins 1998 Kersey-Kuner Terrace Investigations at Dent and Bernhardt Sites, Colorado. Geoarchaeology 13: 201-218.

Heller, Rolf 1980 On Optimal Diet in a Patchy Environment. Theoretical Population Biology 17:201-214.

Hemming, E. Thomas 1970 Early Man in San Pedro Valley . Arizona . Unpublished Ph.D. dissertation, Department of Anthropology, University of Arizona, Tucson.

Hester, J. J. 1962 A Folsom Lithic Complex from the Elida Site, Roosevelt County, New Mexico . El Palacio 69: 9-13. 1972 Blackwater Locality No. 1: A Stratified, Early Man Site in Eastern New Mexico. Publication of the Fort Burgwin Research Center, No. 8. Taos, New Mexico.

Hester, T., R. 1983 Late Paleo-Indian Occupations at Baker Cave, Southwestern Texas. Bulletin of the Texas Archaeological Society 53:101-119.

Hildebrant, W. R. and K. R. McGuire 2002 The Ascendance of Hunting during the California Middle Archaic: An Evolutionary Perspective. American Antiquity 67:231-256.

342

Hill, Matthew E., Jr. 1996 The Paleoindian Bison Fauna from the 12-Mile Creek Site in Western Kansas. Plains Anthropologist 41(158): 359-372. 2002a The Milnesand Site: Site Formation Study of a Paleoindian bison Bonebed in Eastern New Mexico. Plains Anthropologist 47:323-337. 2002b The Folsom-Age 12 Mile Creek Bison Bonebed in Western Kansas. In TER-QUA Symposium Series vol. 3, edited by W. Dort, Jr., pp. 53-70. Institute for Tertiary- Quaternary Studies, University of Kansas, Lawrence. 2006 Rex Rodgers Bonebed: Reevaluation of Site Formation. Current Research in the Pleistocene 23 :108-110. 2007 A Moveable Feast: Variations in Faunal Resource Use Among Central and Western North American Paleoindian Sites. American Antiquity 72(3):417-438.

Hill, Jr., Matthew E. and Matthew G. Hill 2002 Season of Bison Mortality at the Jurgens Site, Weld County, Colorado. Current Research in the Pleistocene 19:104-106.

Hill, Matthew E., Jr. and Jack L. Hofman 1997 The Waugh Site: A Folsom-Age Bison Bonebed in Northwestern Oklahoma. In Southern Plains Bison Procurement and Utilization From Paleoindian to Historic , edited by L. Bement and K. J. Buehler, pp. 63-83. Plains Anthropologist Memoir 29. Lincoln.

Hill, Matthew E., Jr. Jack L. Hofman, and Larry D. Martin 1992 A Reinvestigation of the Burntwood Creek Bison Bonebed. Current Research in the Pleistocene 9: 99-102. 1993 Bone Attritional Processes at the 12-Mile Creek Site. Current Research in the Pleistocene 10:67-69.

Hill, Matthew G. 1996 Size Comparison of the Mill Iron Site Bison Calcanea. In The Mill Iron Site , edited by G.C. Frison, pp. 231-237. University of New Mexico Press, Albuquerque. 2001 Paleoindian Diet and Subsistence Behavior on the Northwestern Great Plains of North America . Unpublished Ph.D. dissertation, Department of Anthropology, University of Wisconsin, Madison. 2005 Late Paleoindian (Allen/Frederick Complex) Subsistence Activities at the Clary Ranch Site, Ash Hollow, Garden County, Nebraska. Plains Anthropologist 50:249-276.

Hill, Matthew G., Vance T. Holliday, and Dennis J. Stanford 1995 A Further Evaluation of the San Jon Site, New Mexico. Plains Anthropologist 4:369-39.

343

Hill, M. G. and M. Kornfeld 1999 Inferring Season of Kill for Cody Complex Bison Bonebed in Middle Park Colorado. Current Research in the Pleistocene 16: 3-32.

Hill, Matthew G., David W. May, and David J. Rapson 2006 A New, Stratified Late Paleoindian Locality on the Clary Ranch, Ash Hallow, Garden County, Nebraska. Current Research in the Pleistocene 23: 110-112 .

Hillerud, J. M., 1970 Subfossil High Plains Bison. Unpublished Ph.D. Dissertation, Department of Geology, University of Nebraska, Lincoln.

Hockett, Bryan and Jonathan Haws 2003 Nutritional Ecology and Diachronic Trends in Palaeolithic Diet and Health. Evolutionary Anthropology 12:211-216. 2005 Nutritional Ecology and the Human Demography of Neanderthal Extinction . Quaternary International 137:21-34.

Hofman, Jack L. 1991 Folsom Land Use: Projectile Point Variability as a Key to Mobility. In Raw Material Economies Among Prehistoric Hunter-Gatherers , edited by A. Montet- White and S. Holen, pp. 335-355. University of Kansas Publications in Anthropology 19, Lawrence. 1992 Recognition and Interpretation of Folsom Technological Variability on the Southern Plains. In Ice Age Hunters of the Rockies, edited by D.J. Stanford and J.S. Day, pp. 193-224. University Press of Colorado, Boulder. 1994 Paleoindian Aggregations on the Great Plains. Journal of Anthropological Archaeology 13: 341-370. 1995 Dating Folsom Occupations on the Southern Plains: The Lipscomb and Waugh Sites. Journal of Field Archaeology 22:421-437. 1996 Early Hunter-Gatherers of the Central Great Plains: Paleoindian and Mesoindian (Archaic) Cultures. In Archaeology and Paleoecology of the Central Great Plains , edited by Jack Hofman, pp. 41-100. Arkansas Archeological Survey Research Series No. 48. Arkansas Archeological Survey, Fayetteville. 1999 Unbounded hunters: Folsom bison hunting on the Southern Plains ca 10,500 BP, the lithic evidence. In Le Bison: Gibier et Moyen de Subsistance des Hommes du Paléolithique aux Paléoindiens des Grandes Plaines , edited by J. Jaubert, J.-P. Brugal, F. David, and J.G. Enloe, pp. 383-415. Editions APDCA, Antibes. 2003 Tethered to Stone or Freedom to Move: Folsom Biface Technology in Regional Perspective. In Multiple Approaches to the Study of Bifacial Technologies , edited by Marie Soressi and Harold Dibble, pp. 229-250. University of Pennsylvania Press, Philadelphia.

344

Hofman, Jack L., Daniel S. Amick, and Richard O. Rose 1990 Shifting Sands: A Folsom-Midland Assemblage From A Campsite in Western Texas. Plains Anthropologist 35(129): 221-253.

Hofman, Jack L. and Jeannette M. Blackmar 1997 The Paleoindian Laird Bison Bone Bed In Northwestern Kansas. The Kansas Anthropologist 18:45-58.

Hofman, Jack L. and Russell W. Graham 1998 The Paleo-Indian Cultures of the Great Plains. In Archaeology on the Great Plains , edited by W. R. Wood, pp. 87-139. University Press of Kansas, Tucson.

Hofman, Jack L., Matthew E. Hill, Jr., William C. Johnson, and Dean T. Sather 1995 Norton: An Early-Holocene Bison Bone Bed in Western Kansas. Current Research in the Pleistocene 12:19-21.

Hofman, Jack L. and Eric Ingbar 1988 A Folsom Hunting Overlook in Eastern Wyoming. Plains Anthropologist 33: 337- 350.

Hofman, Jack L. and Lawrence C. Todd 1997 Reinvestigation of the Perry Ranch Plainview Bison Bonebed, Southwestern Oklahoma. In Southern Plains Bison Procurement and Utilization From Paleoindian to Historic , edited by Leland Bement and Kent J. Buehler, pp. 101- 117. Plains Anthropologist Memoir 29. Lincoln. 2001 Tyranny in the Archaeological Record of Specialized Hunters. In People and Wildlife in Northern North America: Essays in Honor of R. Dale Guthrie , edited by S. C. Gerlach, and M. S. Murray, pp. 200-215. BAR International Series 944. Archaeopress, Oxford.

Holder, P. and J. Wilke 1949 The Frontier Culture Complex: A Preliminary Report of the Prehistoric Hunters Camp in Southwestern Nebraska. American Antiquity 14:260-266.

Holen, Steven R. and Mark P. Muniz 2005 A Flattop Chalcedony Clovis Biface Cache from Northeast Colorado. Current Research in the Pleistocene 22:49-51.

Holliday, Vance T. 1995 Stratigraphy and Paleoenvironments of Late Quaternary Valley Fills on the Southern High Plains . Geological Society of America Memoir 186. Geological Society of America, Boulder. 1997 Paleoindian Geoarchaeology of the Southern High Plains . University of Texas Press, Austin. 345

2000 The Evolution of Paleoindian Geochronology and Typology on the Great Plains. Geoarchaeology 15: 227-290.

Holliday, Vance T., C. Vance Haynes, Jr., Jack L. Hofman and David J. Meltzer 1994 Geoarchaeology and Geochronology of the Miami (Clovis) Site, Southern High Plains of Texas. Quaternary Research 41:234-244.

Holliday, Vance T., Eileen Johnson, and Thomas W. Stafford, Jr. 1999 AMS Radiocarbon Dating of the Type Plainview and Firstview (Paleoindian) Assemblages: The Agony and the Ecstasy. American Antiquity 64:431-443.

Holliday, Vance T. and David J. Meltzer 1996 Geoarchaeology of the Midland (Paleoindian) Site. American Antiquity 61: 755- 771.

Hornaday, W. T. 1889 The extermination of the American bison, with sketch of its discovery and life history. Report of the U.S. National Museum, Part 2. U.S. Government Printing Office 367-548, Washington D.C.

Houston, Alistair I. and John M. McNamara 1985 A General Theory of Central Place Foraging for Single-Prey Loaders. Theoretical Population Biology 28:232-262.

Hovde, D. and D. J. Blakeslee 1977 Archeological Phase II Tests at Site 14SG516--A Suggested Cody Complex Campsite in Sedgwick County, Kansas. Report on file at the Wichita State University Archaeological Research Laboratory, Wichita.

Hudececk-Cuffe, Caroline 1998 Engendering Northern Plains Paleoindian Archaeology. B.A.R. International Series 699. British Archaeological Reports, Oxford.

Hughes, D.T. 1989 Terminal Archaic Bison Kills in the Texas Panhandle. In In the Light of the Past Experience: Papers in Honor of Jack T. Hughes, edited by B.C. Roper, pp. 185- 204. Pandhandle Archeological Society Publication No. 5., Amarillo.

Hughes, Susan S. 1978 Bison Diminution on the Great Plains. Wyoming Contributions to Anthropology 1:18-47. 2003 Beyond the Altithermal: The Role of Climate Change in the Prehistoric Adaptation of Northwestern Wyoming. Unpublished Ph.D. dissertation, Department of Anthropology, University of Washington, Seattle. 346

Humphrey, J.D., and C.R. Ferring 1994 Stable Isotopic Evidence for Latest Pleistocene and Holocene Climatic Change in North Central Texas. Quaternary Research 41:200-213.

Hurst, S., and B. J Carter 2005 Excavation at the Howard Gully (34GR121) Site: Further Insights into Late Paleoindian Cultural Chronology on the Southern Plains, USA. Paper present at the Society for American Archaeology Conference, Salt Lake City.

Husted, W. M. 1969 Bighorn Canyon Archaeology. Smithsonian Institute River Basin Surveys Publications in Salvage Archeology No. 12. Lincoln.

Husted, W. M. and R. Edgar 2002 The Archaeology of Mummy Cave, Wyoming: An Introduction to Shoshoean Prehistory. Midwest Archeology Center Special Report No. 4. National Park Service, Lincoln.

Ingbar, Eric E. 1994 Lithic Material Selection and Technological Organization. In The Organization of North American Prehistoric Chipped Stone Tool Technologies , edited by P. J. Carr, pp. 45-56. Archaeological Series 7, International Monographs in Prehistory, Ann Arbor.

Irwin-Williams, Cynthia, Henry T. Irwin, George A. Agogino, and C. Vance Haynes, Jr. 1973 Hell Gap: Paleo-Indian Occupation on the High Plains. Plains Anthropologist 18:40-53.

Jennings, Jesse D. 1957 Danger Cave. Memoirs of the Society for American Archaeology 14.

Jochim, Michael A. 1981 Strategies for Survival: Cultural Behavior in an Ecological Context. Academic Press, New York. 1988 Optimal Foraging and the Division of Labor. American Anthropologist 90:130- 136.

Jodry, Margaret A. 1999a Folsom Technological and Socioeconomic Strategies: View from Stewart's Cattle Guard and the Upper Rio Grande Basin, Colorado . Unpublished Ph.D. dissertation, Department of Anthropology, American University, Washington, D.C. 1999b Paleoindian Stage. In Colorado Prehistory: A Context for the Rio Grande Basin , 347

edited by M. A. Martorano, T. Hoefer III, M. A. Jodry, V. Spero, and M. L. Taylor, pp. 45-114. Colorado Council of Professional Archaeologists, Denver.

Jodry, Margaret A. and Dennis J. Stanford 1992 Stewart’s Cattle Guard Site: An Analysis of Bison Remains in a Folsom Kill- Butchery Campsite. In Ice Age Hunters of the Rockies , edited by D. J. Stanford and J. S. Day, pp. 11-168. University Press of Colorado, Niwot.

Johnson, Ann M. and Alfred E. Johnson 1998 The Plains Woodland. In Archaeology on the Great Plains , edited by W. R. Wood, pp. 201-234. University Press of Kansas, Lawrence, Kansas.

Johnson, Eileen 1977 Animal Food Resources of Paleoindians. In Paleoindian Lifeways, edited by E. Johnson, Museum Journal (West Texas Museum Association): 15: 107-122. 1985 Current Development in Bone Technology. In Advances in Archaeological Method and Theory , vol. 8., edited by M. B. Schiffer, pp. 157-235. Academic Press, New York. 1987 (editor) Lubbock Lake: Late Quaternary Studies on the Southern High Plains . Texas A & M University Press, College Station.

Johnson, Eileen, and Vance T Holliday 1980 A Plainview Kill/Butchery Locale on the Llano Estacado-The Lubbock Lake Site. Plains Anthropologist 25: 89-111. 1981 Late Paleoindian Activity at the Lubbock Lake Site. Plains Anthropologist 26: 173-193. 1985 A Clovis-Age Megafaunal Processing Station at the Lubbock Lake Landmark. Current Research in the Pleistocene 2:17-19. 1989 Lubbock Lake: Late Quaternary Cultural and Environmental Change on the Southern High Plains, USA. Journal of Quaternary Science 4:145-165. 1997 Analysis of Paleoindian Bonebeds at the Clovis Site: New Data from Old Excavations. Plains Anthropologist 42:329-352.

Johnson, Jr., L. 1961 Devil's Mouth Site: A River Terrace Midden, Diablo Reservoir . Bulletin of the Texas Archeological Society 30: 253-285. 1989 Great Plains Interlopers in the Eastern Woodlands during Late Paleo-Indian Times. Office of State Archeologist Report 36. Texas Historical Commission, Austin.

Johnson, William C. and K.L. Willey 2000 Isotopic and Rock Magnetic Expression of Environmental Change at the Pleistocene-Holocene Transition in the Central Great Plains. Quaternary International 67:89-100. 348

Jones, C. J. 1899 Buffalo Jones' Forty Years of Adventure: A Volume of Facts gathered from Experience, by Hon. C. J. Jones, whose eventful life has been Devoted to the Preservation of the American bison and other Wild Animals . Compiled by Col. Henry Inman. Crane &Co., Topeka.

Jones, D. Chad 1999 Late Paleoindian/Early Archaic Archaeology and Subsistence: A Taphonomic Analysis of the Lime Creek Site (25ST41), Southwestern Nebraska . Unpublished Master's thesis, Department of Anthropology, Colorado State University, Fort Collins.

Jöris O. and B. Weninger 1998 Extension of the 14 C Calibration Curve to ca. 40,000 cal BC by Synchronizing Greenland 18 O/ 16 O Ice Core Records and North Atlantic Foraminifera Profiles: A Comparison with U/Th Coral Data. Radiocarbon 40:495-504.

Judge, W. J. 1973 Paleoindian Occupation of the Central Rio Grande Valley in New Mexico. University of New Mexico Press, Albuquerque.

Katz, Paul 1971 Archaeology of the Sutter Site in Northeastern Kansas. Plains Anthropologist 16:1-19.

Kay, Marvin 1998a The Great Plains Setting. In Archaeology of the Great Plains , edited by W. Raymond Wood, pp. 16-47. University of Kansas Press, Lawrence. 1998b The Central and Southern Plains Archaic. In Archaeology on the Great Plains , edited by W. R. Wood, pp. 173-200. University Press of Kansas, Lawrence, Kansas.

Keeley, Lawrence 1988 Hunter-Gatherer Economic Complexity and “Population Pressure”: A Cross- Cultural Analysis. Journal of Anthropological Archaeology 7:373-411.

Kelly, E. F., S.W. Blecker, C.M. Yonker, C.G. Olson, E.E. Wohl, and Lawrence C. Todd 1998 Stable Isotope Composition of Soil Organic Matter and Phytoliths as Paleoenvironmental Indicators. Geoderma 82, 59-81.

Kelly, E.F., C. Yonker, and B. Marino 1993 Stable Carbon Isotope Composition of Paleosols: An Application to Holocene. In Climate Change in Continental Isotopic Records , edited by P. Swart, K.C. 349

Lohman, J.A. McKenzie, and J. Savin. Geophysical Monographs 78. American Geophysical Union, Washington, D.C.

Kelly, Robert L. 1983 Hunter-gatherer Mobility Strategies. Journal of Archaeological Research 39:277- 306. 1988 The Three Sides of a Biface. American Antiquity 53:717-734. 1995 The Forging Spectrum: Diversity in Hunter-Gatherer Lifeways . Smithsonian Institution Press, Washington. 2002 Early Holocene Prehistory in the Intermontane West: Reinvestigation of the Pine Spring Site, Southwest Wyoming NSF99-73272, Final Report. Report on File at the Department of Anthropology, University of Wyoming, Laramie.

Kelly, Robert L. and Lawrence C. Todd 1988 Coming into the Country: Early Paleoindian Hunting and Mobility. American Antiquity 53:231-244.

Kirby, M. X., and Bruce MacFadden 2005 Was Southern Central America An Archipelago or a Peninsula in the Middle Miocene? A Test Using Land-Mammal Body Size. Palaeogeography, Palaeoclimatology, Palaeoecology 228:193-202 .

Klein, D. R., 1964 Range-Related Differences in Growth of Deer Reflected in Skeletal Ratios. Journal of Mammalogy 45, 226-235.

Klein, Richard G. 1976 The Mammalian Fauna of the Klasies River Mouth Sites, Southern Cape Province, South Africa. South African Archaeological Bulletin 31:75-98.

Knell, Edward J. 2007 The Organization of Late Paleoindian Cody Complex Land-Use on the Northern American Great Plains. Unpublished Ph.D. dissertation, Department of Anthropology, Washington State University, Pullman.

Knell, Edward J., Matthew G. Hill, A. Izeta, Marcel Kornfeld, C. Vance Haynes, Jr., and George C. Frison 2002 The Cody-complex Component at Locality V of the Hell Gap Site, Wyoming. Current Research in the Pleistocene 19:49-52.

Komer, P. E., F. Messier, and C. C. Gates 1993 Group structure in wood bison: nutritional and reproductive determinants. Canadian Journal of Zoology 71:1367-1371.

350

Knudson, Ruthann 2002 Medicine Creek is a Paleoindian Cultural Ecotone. In Medicine Creek: Seventy Years of Archaeological Investigations, edited by D.C. Roper, pp. 142-156. University of Alabama Press, Tuscaloosa.

Koch, Amy 1995 The McIntosh Fauna: Late Prehistoric Exploitation of Lake and Prairie Habitats in the Nebraska Sand Hills. Plains Anthropologist 40:39-60.

Koch, P.L., N.S. Diffenbaugh, and K.A. Hoppe 2004 The Effects of Late Quaternary Climate and pCO 2 Change on C 4 Plant Abundance in the South-Central United States. Palaeogeography, Palaeoclimatology, Palaeoecology 207:331-357.

Koch, W., 1935 The Age Order of Epiphyseal Union in the Skeleton of the European Bison ( Bos Bonasus L.). The Anatomical Record 61:371-376.

Kornfeld, Marcel 1996 Big Game Focus: Reinterpreting the Archeological Record of Cantabrian Upper Paleolithic Economy. Current Anthropology 37, 629-657. 2003 Affluent Foragers of the North American Plains: Landscape Archaeology of the Black Hills . BAR International Series 1106. John and Erica Hedge Ltd, and Archaeopress, Oxford.

Kornfeld, Marcel, and Julie E Francis 1991 A Preliminary Historical Outline of Northwestern High Plains Gender Systems. In Gender in Archeology, Proceedings of the 22nd Annual Conference of the Archeological Association of the University of Calgary, edited by D. Walde and N. D. Willow, pp. 444-451. University of Calgary, Alberta.

Kornfeld, M., and G. C. Frison 2000 Paleoindian Occupation of the High Country: The Case of Middle Park Colorado. Plains Anthropologist 45:129-153.

Kornfeld, Marcel, George C. Frison, Mary L. Larson, J. C. Miller, and J. Saysette 1999 Paleoindian Bison Procurement and Paleoenvironments in Middle Park of Colorado. Geoarchaeology 14:655-674.

Kornfeld, Marcel, Mary Lou Larson, Craig Arnold, Adam Wiewel, Mike Toft, and Dennis Stanford 2006 Preliminary Reports on the Nelson Cody Site Weld County Colorado. George C. Frison Institute of Archaeology and Anthropology Technical Report No. 40. University of Wyoming, Laramie. 351

Kornfeld, Marcel, Mary L. Larson, David Rapson, and George C. Frison 2001 10,000 Years in the Rocky Mountains: The Helen Lookingbill Site. Journal of Field Archaeology 28: 3/4: 307-324.

Kornfeld, M., and A. J. Osborn (editors) 2003 Island on the Plains: Ecological and Social, and Ritual Use of Landscapes. University of Utah Press, Salt Lake City.

Krasinska, Malgorzata and Zbigniew A. Kasinski 2002 Body mass and Measurements of the European Bison During Postnatal development. Acta Theriologica 47:85-106.

Krasinska, Malgorzata, Zbigniew A. Kasinski, and Aleksei N. Bunevich 2000 Factors Affecting the Variability in Home Range Size and Distribution in European Bison in the Polish and Belarussian Parts of Bialowieza Forest. Acta Theriologica 45:321-334

Krausman, P.R., A. V. Sandoval, and R.C. Etchberger 1999 Natural History of Desert Bighorn Sheep. In Mountain Sheep of North America , edited by R. Valdez and P.R. Krausman, pp. 139-191. The University of Arizona Press, Tucson.

Krebs, J.R. and N. B. Davies 1987 An Introduction to Behavioural Ecology, 2nd edition. Blackwell Scientific Publications, Oxford.

Kuhn, Steven L. 1991 "Unpacking" Reduction: Lithic Raw Material Economy in the Mousterian of West-Central Italy. Journal of Anthropological Archaeology 10: 76-106.

Kuhn, Steven L. and Mary C. Stiner 2001 The Antiquity of Hunter-Gatherers. In Hunter-Gatherers: An Interdisciplinary Perspective, edited by C. Panter-Brick, R. H. Layton, and P. Rowley-Conwy, pp. 99-142. Cambridge University Press, Cambridge. 2006 What's a Mother to Do?: The Division of Labor among Neandertals and Modern Humans in Eurasia. Current Anthropology 47:953-980.

Kurtén, Björn 1973. Geographic variation in size in the Pumas ( Felis concolor ). Commentationes Biologicae 63:1-8.

Kurtén, Björn and Elaine Anderson 1980 Pleistocene Mammals of North America. Columbia University Press, New York . 352

LaBelle, Jason M. 2001 Rediscovery of the Slim Arrow Paleoindian Bonebed: The Yuma Type Site that Wasn't Meant to be. Paper present at 59th Plains Anthropological Conference, Lincoln. 2005 Hunter-Gatherer Foraging Variability During the Early Holocene of the Central Plains of North America. Unpublished Ph.D. dissertation, Southern Methodist University, Dallas.

LaBelle, Jason M., Vance T. Holliday, and David J. Meltzer 2003 Early Holocene Paleoindian Deposits at Nall Playa, Oklahoma Panhandle, U.S.A. Geoarchaeology 18: 5-34.

LaBelle, J. M. and D. J. Meltzer 1996 Reinvestigation of Marks Beach: A Possible Early/Middle Holocene Campsite on Blackwater Draw. Current Research in the Pleistocene 13: 31-33.

Larter, Nicholas C., and C. Cormack Gates 1990 Home range of wood bison in an expanding population. Journal of Mammalogy 71: 604-607. 1994 Home-Range Size of Wood Bison: Effects of Age, Sex, and Forage Availability. Journal of Mammalogy 75:142-149.

Lee, Richard B. and Irven Devores (editors) 1968 Man the Hunter. Aldine, New York.

Leechman, Douglas 1951 Bone Grease. American Antiquity 17:355-356 .

Leonard, E. M. and M. A. Reasoner 1999 A Continuous Holocene Glacial Record Inferred from Postglacial Lake Sediments in Banff National Park, Alberta, Canada. Quaternary Research 51:1-13.

Leonhardy, F. 1966 Domebo: A Paleoindian Mammoth Kill in the Prairie-Plains. Contribution of the Museum of the Great Plains, No. 1. The Great Plains Historical Association, Lawton, Oklahoma.

Levins, R. 1968 Evolution in Changing Environments: Some Theoretical Explorations. Princeton University Press, Princeton.

353

Lewis, P.J., B. Buchanan, and E. Johnson 2005 Sexing Bison Metapodials Using Principal Component Analysis. Plains Anthropologist 50:159-172.

Litwinionek, Luc, Eileen Johnson, and Vance T. Holliday 2003 The Playas of the Southern High Plains: An Archipelago of Human Occupation for 12,000 Years on the North American Grasslands. In Islands on the Plains: Ecological, Social, and Ritual Use of Landscapes, edited by M. Kornfeld and A. J. Osborn, pp. 21-43. The University of Utah Press, Salt Lake City.

Logan, Brad 1998 The Fat of the Land: White Rock Phase Bison Hunting and Grease Production. Plain Anthropologist 43:349-366.

Lomolino, M.V. 2005 Body Size Evolution in Insular Vertebrates: Generality of the Island Rule. Journal of Biogeography 32:1683-1699.

Lomolino, M.V., D.F. Sax, B.R. Riddle, and J.H. Brown 2006 The Island Rule and a Research Agenda for Studying Ecogeographical Patterns. Journal of Biogeography 33, 1503-1510.

Lovvorn, M.B., G.C. Frison, and L.L. Tieszen 2001 Paleoclimate and Amerindians: Evidence form Stable Isotopes and Atmospheric Circulation. Proceedings of the National Academy of Sciences 98:2485-2490.

Lott, Dale F. 1974 Sexual and aggressive behavior of bison. In The Behavior of Ungulates and Its Relation to Management , edited by V. Geist and F. Walther. IUCN Publication 24: 382-294. 1981 Sexual behavior and intersexual strategies in American bison. Zeitschrift Tierpsychologica 56: 97-114. 1991 American bison socioecology. Applied Animal Behaviour Science 29: 135-145. 2002 American Bison: A Natural History. University of California Press, Berkeley.

Lott, D. F., and S. C. Minta 1983 Home ranges of American bison cows on Santa Catalina Island, California. Journal of Mammalogy 64: 161-162.

354

Lundelius, Jr., E. L. 1988 What Happened to the Mammoths? The Climatic Model. In Americans Before Columbus: Ice-Age Origins, edited by R.C. Carlisle, pp. 75-82. Ethnology Monographs 12, Department of Anthropology, University of Pittsburg. 1990 Vertebrates on Oceanic and Continental Islands. International Symposium on Biogeographical Aspects of Insularity. Atti dei Covegni Lincei 85 : 3-22

Lundelius, Jr., E.L., R.W. Graham, E. Anderson, J. Guilday, J.A. Holman, D. W. Steadman, and S. D. Webb 1983 Terrestrial Vertebrate Faunas. In Late Quaternary Environments of the United States . Vol 1. The Late Pleistocene , edited by S. C. Porter, pp. 311-353. University of Minnesota Press, Minneapolis.

Lyman, R. Lee 2003 The Influence of Time Averaging and Space Averaging on the Application of Foraging Theory in Zooarchaeology. Journal of Archaeological Science 30:595- 610. 2004a Aboriginal Overkill in the Intermountain West of North America: Zooarchaeological Tests and Implications. Human Nature 15:169-208. 2004b Late-Quaternary Diminution and Abundance of Prehistoric Bison ( Bison sp.) in Eastern Washington State, USA. Quaternary Research 62:76-85.

MacArthur, R. H. and E. R. Pianka 1966 On Optimal Use of a Patchy Environment. The American Naturalist 100:603-609.

McCartney, Peter H. 1983 An Archaeological Analysis of Bison Remains from the Cody Paleo-Indian Site of Lamb Spring . Unpublished Master's thesis, Department of Anthropology, University of Arizona, Tucson. 1990 Alternative Hunting Strategies in Plains Paleoindian Adaptation. In Hunters of the Recent Past , edited by L. B. Davies, and B. O. K Reeves, pp. 111-121. Unwin Hyman, Boston.

MacDonald, D. H. 1998 Subsistence, Sex, and Cultural Transmission in Folsom Culture. Journal of Anthropological Archaeology 17:217-239.

MacDonald, J. N. 1981 North American Bison: Their Classification and Evolution. University of California Press, Berkeley.

MacDonell, G. H. and L. Wandsnider 2003 The Western Niobrara River: An Inter-Island Passage on the Plains. In Islands on the Plains: Ecological, Social, and Ritual Use of Landscapes , edited by M. 355

Kornfeld and A. J. Osborn, pp. 89-107. The University of Utah Press, Salt Lake City.

McHugh, T. 1958 Social behavior of the American buffalo ( Bison bison bison ). Zoologica 43(1): 1- 40. 1972 The Time of the Buffalo . Knopf, New York.

McKee, David F., 1988 Bison Hunting and Processing at the River Bend Site (48NA202). The Wyoming Archaeologist 31:13-32.

MacKey, J. C., D. Kullen, S. D. Creasman, J. E. Sall, K. Harvey, and C. L. Armitage 1982 Paleoenvironmental reconstruction and Subsistence Change at the Deadman Wash Site in Southwestern Wyoming. Journal of Intermountain Archaeology 1: 11-65.

McNab, B.K. 1990 The Physiological Significance of Body Size. In Body Size in Mammalian Paleobiology: Estimation and Biological Implications , edited by J. Damuth, and B.J. MacFadden, pp. 11-23. Cambridge University Press, Cambridge.

McNamara, John 1982 Optimal Patch Use in a Stochastic Environment. Theoretical Population Biology 21:269-288.

Madsen, D. B. 1993 Testing Diet Breadth Models: Examining Adaptive Change in the Late Prehistoric Great Basin. Journal of Archaeological Science 20: 321-329.

McGuire, K. R. and W. R. Hildebrandt 2005 Re-Thinking Great Basin Foragers: Prestige Hunting and Costly Signaling During the Middle Archaic Period. American Antiquity 70:695-712.

Mallouf, Robert J. 1994 Sailor-Helton: A Paleoindian Cache from Southwestern Kansas. Current Research in the Pleistocene 11:44-46.

Mallouf, Robert J. and Rolfe D. Mandel 1997 Horace Rivers: A Late-Plainview Component in the Northeastern Texas Panhandle. Current Research in the Pleistocene 14:50-52.

Mandel, Rolfe 1995 Geomorphological Controls of the Archaic Record in the Central Plains of the United States. In Archaeological Geology of the Archaic Period in North 356

America , edited by E. A. Bettis, pp. 37-66. Special Paper 297, Geological Society of America, Denver,

Mandel, Rolfe D, and Jack L. Hofman 2003 Geoarchaeological Investigations at the Winger Site: A Late Paleoindian Bison Bonebed in Southwestern Kansas, U.S.A. Geoarchaeology 18:129-144.

Mandel, Rolfe D., Steven Holen, and Jack L. Hofman 2005 Geoarchaeology of Clovis and Possible Pre-Clovis Cultural Deposits at the Kanorado Locality, Northwestern Kansas. Current Research in the Pleistocene 22:56-57.

Mannino, M.A., and K.D. Thomas 2002 Depletion of a Resource? The Impact of Prehistoric Human Foraging on Intertidal Mollusc Communities and its Significance for Human Settlement, Mobility, and Dispersal. World Archaeology 33:452-474.

Marcus, L.F., and R. Berger 1984 The Significance of Radiocarbon Dates for Rancho La Brea. In Quaternary Extinctions: A Prehistoric Revolution , edited by P.S. Martin, and R.G. Klein, pp. 159-183. University of Arizona Press, Tucson.

Marquez, M., and B. E. Coblentz 1987 Metatarsal and mandibular marrow fat in black-tailed deer. Journal of Wildlife Management 51: 38-40.

Martin, Larry D. and Daniel R. Williams 2007 Late-Pleistocene Biodiversity in the Central Plains. Current Research in the Pleistocene 23:172-173 .

Martin, Paul S. 1984 Prehistoric Overkill: The Global Model. In Quaternary Extinctions: A Prehistoric Revolution , edited by P.S. Martin, and R. G. Klein, pp. 354-403. The University of Arizona Press, Tucson. 2005 Twilight of The Mammoths: Ice Age Extinctions and the Rewilding of America . University of California Press, Berkeley.

Martin, Paul S. and David W. Steadman 1999 Prehistoric Extinctions on Islands and Continents. In Extinction in Near Time, edited by R.D.E. MacPhee, pp. 17-52. Kluwer Academic/Plenum Publishers, New York.

Martin, Paul S. and Christine R. Szuter 2004 Revising the "Wild" West: Big Game Meets the Ultimate Keystone Species. In 357

The Archaeology of Global Change: The Impact of Humans on Their Environment , edited by C. L. Redman, S. R. James, P. R. Fish, and J. D. Rogers, pp. 63-88. Smithsonian Books, Washington.

Martin, R. A., and A. D. Barnosky 1993 Morphological Change in Quaternary Mammals of North America . Cambridge University Press, Cambridge.

Matheus, P. E. 2001 Pleistocene Predators and People in Eastern Beringia: Did Short-Faced Bears Really Keep Humans out of North America. In People and Wildlife in Northern North America: Essays in Honor of R. Dale Guthrie , edited by S.G. Gerlach, M.S. Murray, pp. 79-101. BAR International Series 944, Archaeopress, Oxford.

Mayr, E. 1956 Geographic Character Gradient and Climatic Adaptations. Evolution 10:105-108.

Mead, James I., and David J. Meltzer 1984 North American Late Quaternary Extinctions and the Radiocarbon Record. In Quaternary Extinctions: A Prehistoric Revolution , edited by P.S. Martin and R. G. Klein, pp. 440-450. The University of Arizona Press, Tucson.

Meagher, Margaret M. 1973 The Bison of Yellowstone National Park. National Park Service Science Monograph 1:1-161.

Meiri, S., T. Dayan, D. Simberloff, 2004 Carnivores, Biases and Bergmann's Rule. Biological Journal of the Linnaean Society 81:579-588. 2006 The Generality of the Island Rule Reexamined. Journal of Biogeography 33:1571-1577.

Melton, D. A., N. C. Larter, C.C. Gates, and J. A. Virgil 1989 The influence of rut and environmental factors on the behavior of wood bison. Acta Theriologia 34: 179-193.

Meltzer, David J. 1988 Late Pleistocene Human Adaptations in Eastern North America. Journal of World Prehistory 2:1-52. 1993 Is There a Clovis Adaptation. In From Kostenki to Clovis: Upper Paleolithic- Paleo-Indian Adaptations , edited by O. Soffer and N. D. Praslov, pp. 293-310. Plenum Press, New York. 1995 Clocking the First Americans. Annual Review of Anthropology 24:21-45. 2004 Peopling of North America. In The Quaternary Period in the United States, edited 358

by A.R. Gillespie, S. C. Porter, and B. F. Atwater, pp. 539-564. Elsevier, Boston. 2006 Folsom: New Archaeological Investigations of a Classic Bison Kill . University of California Press, Berkeley.

Meltzer, David J. and Bruce D. Smith 1986 Paleoindian and Early Archaic Subsistence Strategies in Eastern North America. In Foraging, Collecting, and Harvesting: Archaic Period Subsistence and Settlement in Eastern Woodlands, edited by S. Neusius, pp. 3-31. Center for Archaeological Investigations, Southern Illinois University, Carbondale.

Meltzer, David J., Lawrence C. Todd, and Vance T. Holliday 2002 The Folsom (Paleoindian) Type Site: Past Investigations, Current Studies. American Antiquity 67:5-36.

Messier, F. 1994 Ungulate Population Models with Predation: A Case Study with the North American Moose. Ecology 75:478-488.

Meyer, D. 1985 A Component in the Scottsbluff Tradition: Excavation at the Niska Site. Canadian Journal of Archaeology 9(1): 1-37.

Meyer, Michael W., and Robert D. Brown 1985 Seasonal Trends in the Chemical Composition of Ten Range Plains in South Texas. Journal of Range Management 38:154-157.

Mitchell, H. H. 1962 Comparative Nutrition of Man and Domestic Animals. Academic Press, New York.

Moen, Aaron N. 1973 Wildlife Ecology: An Analytical Approach. WH Freeman & Co., San Francisco.

Morgan, R. G. 1980 Bison movement patterns on the Canadian Plains: an ecological analysis. Plains Anthropologist 25: 143-160.

Morlan, R.E. 1991 Bison Carpal and Tarsal Measurements: Bulls Versus Cows and Calves. Plains Anthropologist 36:215-227.

Moss, John H. 1951 Early Man in the Eden Valley. The University Museum, University of Pennsylvania, Philadelphia. 359

Mulloy, William 1959 The James Allen Site near Laramie, Wyoming. American Antiquity 25:112-116.

Muniz, Mark P. 2004 Exploring Technological Organization and Burial Practices at the Paleoindian Gordon Creek Site (5LR99), Colorado. Plains Anthropologist 49: 253-280.

Munro, Natalie D. 2004 Zooarchaeological Measures of Hunting Pressure and Occupation Intensity in the Natufian. Current Anthropology 45, Supplement: S5-S33.

Mysterud, A., K. A. Bartón, B. Jedrezejewska, Z. A. Krasinski, M. Niedzialkowska, J. F. Kamler, N.G. Yoccoz, and N. C. Stenseth. 2007 Population Ecology and Conservation of Endangered Megafauna: The Case of European bison in Bialowieza, Primeval Forest, Poland. Animal Conservation 10:77-87.

Nagaoka, L. 2001 Using Diversity Indices to Measure Changes in Prey Choice at the Shag River Mouth Site, Southern New Zealand. International Journal of Osteoarchaeology 11:101-111.

Neck, R. W. 1995 Molluscan Remains. In Stratigraphy and Paleoenvironments of Late Quaternary Valley Fills on the Southern High Plains, by Vance T. Holliday, pp. 59-67. Geological Society of America Memoir 186, Boulder.

Nelson, K. L. 1965 Status and habit of the American buffalo (Bison bison) in the Henry Mountain areas of Utah . Publication No 65-2 Utah State Department of Fish and Game, Salt Lake City.

Niven, Laura B., and Matthew G. Hill 1998 Season of Bison Mortality at Three Plains Archaic Kills Sites in Wyoming. Plains Anthropologist 4:5-26.

Niven, Laura B., C.P. Egeland, and Lawrence C. Todd 2004 An Inter-Site Comparison of Enamel Hypoplasia in Bison: Implications for Paleoecology and Modeling Late Plains Archaic Subsistence. Journal of Archaeological Science 31:1783-1794.

360

Noble, R.C. 1984 Essential Fatty Acids in the ruminant. In Fats in Animal Nutrition , edited by J. Wiseman, pp. 185-200. Butterworth, Boston.

Nordt, L.C. 1998 Geoarchaeology of the Rio Grande and Elm Creek in the Vicinity of Site 41MV120. In 41MV120: A Stratified Late Archaic Site in Maverick County, Texas, edited by B.J. Vierra. Archaeological Survey Report No. 251. Center for Archaeological Research, University of Texas at San Antonio.

Nordt, L. C., T. W. Boutton, C. T. Hallmark, and M. R. Waters 1994 Late Quaternary Vegetation and Climate Changes in Central Texas Based on the Isotopic Composition of Organic Carbon. Quaternary Research 41:109-120.

Nordt, L. C., T. W. Boutton, J. S. Jacob, and R. D. Mandel 4 2002 C Plant Productivity and Climate-CO 2 Variations in South-Central Texas During the Late Quaternary. Quaternary Research 58:182-188.

Oaten, Allan 1977 Optimal Foraging in Patches: A Case for Stochasticity. Theoretical Population Biology 12:263-285.

O'Brien, Patricia J. 1984 The Tim Adrian Site (14NT604): A Hell Gap Quarry Site in Norton County, Kansas. Plains Anthropologist 29: 41-55.

O'Connell, James F. 1977 Aspects of Variation in Central Australian Lithic Assemblages. In Stone Tools as Cultural Markers: Change, Evolution, and Complexity, edited by R. V. S. Wright, pp. 269-281. Australian Institute of Aboriginal Studies, Canberra.

O'Connell, James F. and Kristen Hawkes 1981 Alywara Plant Use and Optimal Foraging Theory. In Hunter-Gatherer Foraging Strategies: Ethnographic and Archaeological Analyses , edited by B. Winterhalder and E. A. Smith, pp. 99-125. University of Chicago Press, Chicago.

Olson, C.G., and D.A. Porter 2002 Isotopic and Geomorphic Evidence for Holocene Climate, Southwestern Kansas. Quaternary International 87:29-44.

Orians, Gordon H. and Nolan E. Pearson 1979 On the Theory of Central Place Foraging. In Analysis of Ecological Systems , edited by D.J. Horn, G. R. Stairs, and R. D. Mitchell, pp. 155-177.

361

Osborn, Alan J. 1983 Ecological Aspects of Equestrian Adaptations in Western North America. American Anthropologist 85:563-591. 1993 Snowblind in the American Southwest: Moisture Islands, Ungulate Ecology, and Alternative Prehistoric Overwintering Strategies. Journal of Anthropological Research 49:135-164. 2003 "Places Where they Kill Deer": Moisture Islands, Winter Climate, and Ungulate Ecology in the Northwestern Great Plains. In Islands on the Plains: Ecological, Social and Ritual Use of Landscapes, edited by M. Kornfeld and A. J. Osborn, pp. 193-219

Owens-Smith, R. N. 1988 Megaherbivores: The Influence of Very Large Body Size on Ecology . Cambridge University Press, London.

Owsley, Doug W. and K. L. Brusweldheide 1997 Bioarchaeology Research in Northeastern Colorado, Northern Kansas, Nebraska, and South Dakota. In Bioarchaeology of the North Central United States , edited by D. W. Owsley and J. C. Rose, pp. 7-56. Arkansas Archeological Survey Research Series No. 49. Arkansas Archeological Survey, Fayetteville.

Owsley, Doug W. and D. R. Hunt 2001 Clovis and Early Archaic Period Crania from the Anzick Site (24PA506), Park County, Montana. Plains Anthropologist 46: 225-124.

Paruelo, J.M., and W.K. Lauenroth 1996 Relative Abundance of Plant Functional Types in Grasslands and Scrublands of North America. Ecological Applications 6:1212-1224.

Peden, D. G., G. M. Van Dyne, R.W. Rice, and R. M. Hansen 1974 The trophic ecology of Bison bison on shortgrass Plains. Journal of Applied Ecology 11: 489-498.

Peters, R. H. 1983 The Ecological Implications of Body Size . Cambridge University Press.

Peterson, M. R. 2001 Folsom Mobility and Technological Organization at the Krmpotich Site: An Analysis of the Lithic Artifact Assemblage. Unpublished Master's thesis, Department of Anthropology, University of Wyoming, Laramie.

Pianka, Eric R. 1983 Evolutionary Ecology. Harper & Row Publishers, New York.

362

Pitblado, Bonnie L. 1999 Late Paleoindian Occupation of the Southern Rocky Mountains: Projectile Points and Land Use in the High Country. Unpublished Ph.D. dissertation, Department of Anthropology, University of Arizona, Tucson. 2000 Living the High Life in Colorado: Late Paleoindian Occupation of the Caribou Lake Site. In This Land of Shining Mountains: Archeological Studies in Colorado's Indian Peaks Wilderness Areas, edited by E. S. Cassells, pp. 124-158. Research Report 8, Center for Mountain Archaeology, Ward, CO. 2003 Late Paleoindian Occupation of the Southern Rocky Mountains: Early Holocene Projectile Points and Land Use in the High Country. University Press of Colorado, Boulder.

Pitblado, Bonnie L., Beth A. Camp, and J. Stam 2001 Final Report. 2000 Test Excavations at the Chance Gulch Site (5GN817). Report Submitted to Colorado Historical Society, Denver.

Pitblado, Bonnie L. and Beth A. Camp 2003 2001-02 Excavations at the Change Gulch Site (5GN817). Report Submitted to Colorado Historical Society, Denver and Bureau of Land Management, Denver.

Potts, Richard, Anna K. Behrensmeyer, and P. Ditchfield 1999 Paleolandscape Variation and Early Pleistocene Hominid Activities: Members 1 and 7, Olorgesailie Formation, Kenya. Journal of Human Evolution 37:747-788.

Pucek, Z., I.P. Bleousova, M. Krasinska, Z. A. Krasinski, and W. Olech (compilers) 2003 European Bison Bison bonasus: Current State of the Species and an Action Plan for its Conservation . Mammal Research Institute, Polish Academy of Sciences, Bialowieza, Poland.

Purdue, J. R. 1989 Changes during the Holocene in the Size of White-tailed deer ( Odocoileus virginianus ) from Central Illinois. Quaternary Research 32:307-316. 1993 Decrease in the Body Size of White Tailed Deer (Odocoileus virginianus) during the Late Holocene in South Carolina Georgia. In Morphological Change in Quaternary Mammals of North America , edited by R.A. Martin, and A. D. Barnosky, pp. 281-298. Cambridge University Press, Cambridge.

Quigg, J. Michael 1997 Bison Processing at the Rush Site, 41TG346, and Evidence for Pemmican Production in the Southern Plains. Plains Anthropologist 42:145-161. 1998 A Late Archaic Bison Processing Event in the Texas Panhandle. Plains Anthropologist 43:367-383.

363

Raia, P., C. Barbera, and M. Conte, 2003 The Fast Life of a Dwarf Giant. Evolutionary Ecology 17, 293-312.

Ransom, A. B. 1965 Kidney and marrow fat as indicator of White-tailed deer condition. Journal of Wildlife Management 29: 397-398.

Rapport, David J. 1971 An Optimization Model of Food Selection. The American Naturalist 105:575- 587.

Rapson, David J. 1990 Pattern and Process in Intra-Site Spatial Analysis: Site Structural and Faunal Research at the Bugas-Holding Site. Unpublished Ph.D. Dissertation, University of New Mexico, Albuquerque.

Ray, C. N. and K. Bryan 1938 Folsomoid Point Found in Alluvium Beside Mammoth's Bones. Science 88: 257- 276.

Redder, A. J. 1985 Horn Shelter Number 2, the South End: A Preliminary Report . Central Texas Archaeologist 10: 37-66.

Reeves, Brian O. K. 1990 Communal Bison Hunters of the Northern Plains. In Hunters of the Recent Past , edited by L. B. Davis and B.O.K. Reeves, pp. 168-194. Unwin Hyman, Boston.

Reher, Charles A. 1970 Population Dynamics of the Glenrock Bison bison Population. In The Glenrock Buffalo Jump, 48CO304 , edited by G. C. Frison. Plains Anthropologist Memoir No. 7, Appendix II. 1973 Appendix II: The Wardell Bison bison Sample: Population Dynamics and Archaeological Interpretation. In The Wardell Buffalo Trap 48SU301: Communal Procurement in the Upper Green River Basin . Wyoming Anthropological Papers of the Museum of Anthropology, University of Michigan 48. 1974 Population Study of the Casper Site Bison. In The Casper Site: A Hell Gap Bison Kill on the High Plains , edited by G. C. Frison, pp. 113-124. Academic Press, New York. 1977 Adaptive Process on the Shortgrass Plains. In For Theory Building in Archaeology , edited by L. R. Binford, pp. 13-40. Academic Press, New York.

364

Reher, Charles A., and George C. Frison (editors) 1980 The Vore Site, 48CK302, A Stratified Buffalo Jump in Wyoming Black Hills . Plains Anthropologist Memoir No. 16, Lincoln.

Reynolds, H. W. R. M. Hansen, and D.G. Peden 1978 Diets of the Slave River lowland bison herd, Northwest Territories, Canada. Journal of Wildlife Management 43: 581-590.

Reynolds, H. W., R. D. Glaholt, and A. W. L. Hawley 1982 Bison-Bison bison. In Wild Mammals of North America , edited by J.A. Chapman, and G. A. Feldhamer, pp. 972-1007. John Hopkins University Press, Baltimore.

Reznick, D.N., and H. Bryga 1987 Life-history Evolution in Guppies ( Poecilia reticulata ): 1. Phenotypic and Genetic Changes in an Introduction Experiment. Evolution 41:1370-1385.

Reznick, D.N., F.H. Shaw, F.H. Rodd, and R. G. Shaw 1997 Evaluation of the Rate of Evolution in Natural Populations of Guppies ( Poecilia reticulata ). Science 275:1934-1937.

Rhodes, D.N 1973 Fatness of Beef And Eating Quality . Meat Research Institute Memorandum No 15. Langford Bristol Meat Research Institute.

Ritterbush, Lauren W. 2006 Late Prehistoric Oneota in the Central Plains. In Kansas Archaeology , edited by R.J. Hoard and W. E. Banks, pp. 133-150. University Press of Kansas, Lawrence, Kansas.

Roberts, Frank H. H. 1935 A Folsom Complex: A Preliminary Report on Investigations at the Lindenmeier in Northern Colorado. Smithsonian Miscellaneous Collections : 94:1-35. 1945 A Deep Burial on the Clear Fork of the Brazos River. Bulletin of the Texas Archeological and Paleontological Society 16: 9-30.

Robbins, C. T. 1983 Wildlife Feeding and Nutrition . Academic Press: New York.

Rodríguez, M. Á., I. L. López-Sañudo, and B.A. Hawkins 2006 The Geographic Distribution of Mammal Body Size in Europe. Global Ecology and Biogeography 15:173-181.

Roe, F.G. 1951 The North American Buffalo. University of Toronto Press, Toronto. 365

Roper, Donna 2006 The Central Plains Tradition. In Kansas Archaeology , edited by R.J. Hoard and W. E. Banks, pp. 105132. University Press of Kansas, Lawrence, Kansas.

Root, Matthew J. 1997 Production for Exchange at the Knife River Flint Quarries, North Dakota. Lithic Technology 22:33-50. 2000 The Archaeology of the Bobtail Wolf Site: Folsom Occupation of the Knife River Flint Quarry Area, North Dakota. Washington State University Press, Pullman.

Rutberg, A. T. 1984 Birth synchrony in American bison ( Bison bison ): Response to predation or season? Journal of Mammalogy 65: 418-423.

Sahlins, M. 1976 Culture and Practical Reason. University of Chicago Press, Chicago.

Saunders, Jeffrey J. and E. B. Daeschler 1994 Descriptive Analyses and Taphonomic Observations of Culturally-Modified Mammoths Excavated at "The Gravel Pit," Near Clovis, New Mexico in 1936. Proceedings of the Academy of Natural Sciences of Philadelphia 145:1-28.

Scheiber, Laura L. 2006 The Late Prehistoric on the High Plains of Western Kansas: High Plains Upper Republican and Dismal River. In Kansas Archaeology , edited by R. J. Hoard and W. E. Bank, pp. 133-150. University Press of Kansas, Lawrence.

Schiffer, Michael B. 1987 Formation Processes of the archaeological Record. University of New Mexico Press, Albuquerque.

Schoener, Thomas W. 1971 Theory of Feeding Strategies. Annual Review of Ecology and Systematics 2:369- 404.

Schoolcraft, Henry 1856 Information Respecting the History, Conditions, and Prospects of the Indian Tribes of the United States , J. B. Lippincott, Philadelphia.

Schultz, C. Bertrand 1932 Association of Artifacts and Extinct Mammals in Nebraska, Bulletin of the Nebraska State Museum 33(1):271-282. 1943 Some Artifact Sites of Early Man in the Great Plains and Adjacent Areas. 366

American Antiquity 8:242-249.

Scott, K.M. 1990 Postcranial Dimensions of Ungulates as Predictors of Body Mass. In Body Size in Mammalian Paleobiology , edited by J. Damuth, and B.J. MacFadden, pp. 301- 335. Cambridge University Press, Cambridge.

Sellards, E. H. 1938 Artifacts Associated with Fossil Elephant. Bulletin of the Geological Society of America 49:999-1010. 1952 Early Man in America . University of Texas Press, Austin. 1955 Fossil Bison and Associated Artifacts from Milnesand, New Mexico. American Antiquity . 2:336-344.

Sellards, E. H., G. L. Evans, G. E. Meade 1947 Fossil Bison and Associated Artifacts from Plainview, Texas. Bulletin of Geological Society of America 58:927-954.

Sellet, Frédéric. 1999 A Dynamic View of Paleoindian Assemblages at the Hell Gap Site, Wyoming: Reconstructing Lithic Technological Systems . Unpublished Ph.D. dissertation, Department of Anthropology, Southern Methodist University, Dallas. 2001 A Changing Perspective on Paleoindian Chronology and Typology: A view from the Northwestern Plains. Arctic Anthropology 38:48-63.

Senft, R. L., M. B. Coughenour, D. W. Bailey, L. R. Rittenhouse, O. E. Sala, and D. M. Swift 1987 Large Herbivore Foraging and Ecological Hierarchies. BioScience 37: 789-799.

Seton, E. T. 1910 Life-Histories of Northern Animals 2 vol. Scribner's, New York.

Shackleton, David M., Christopher C. Shank, and Brian M. Wikeem 1999 Natural History of Rocky Mountain and California Bighorn Sheep. In Mountain Sheep of North America , edited by R. Valdez and P.R. Krausman, pp. 78-138. The University of Arizona Press, Tucson.

Shapiro, B., A.J. Drummond, A. Rambaut, M.C. Wilson, P.E. Matheus, A.V. Sher, O.G. Pybus, M.T. Gilbert, I. Barnes, J. Binladen, E. Willerslev, A.J. Hansen, G.F. Baryshnikov, J.A. Burns, S. Davydov, J.C. Driver, D.C. Froese, C.R. Harington, G. Keddie, P. Kostintsev, M.L. Kunz, L.D. Martin, R.O. Stephenson, J. Storer, R. Tedford, S. Zimov, and A. Cooper 2004. Rise and Fall of the Beringian Steppe Bison. Science 306:1561-1565.

367

Sheenan, Michael S. 2002 Dietary Responses to Mid-Holocene Climatic Change. North American Archaeologist 23:117-143.

Sher, A.V. 1971 Pleistocene Mammals and Stratigraphy of the Far Northeast USSR and North America. International Geological Review 16:1-284.

Shult, M. J. 1972 American bison behavior patterns at Wind Cave National Park. Unpublished Ph. D. dissertation, University of Iowa, Iowa City.

Simms, S. R. 1987 Behavioral Ecology and Hunter-Gatherer Foraging: An Example from the Great Basin. BAR International Series 381. British Archaeological Reports, Oxford. 1988 Conceptualizing the Paleoindian and Archaic in the Great Basin. In Early Human Occupation in the Far Western North America: The Clovis-Archaic Interface, edited by J.A. Willig, C. M. Aikens, and J. L. Fagan, pp. 41-52. Nevada State Museum Anthropological Papers 21. Nevada State Museum, Carson City.

Simpson, E. H. 1949 Measurement of Diversity. Nature 163: 688 .

Simpson, G.G. 1941 Large Pleistocene Felines of North America. American Museum Novitates 1136:1-27.

Slonaker, J. R. 1925 The effect of copulation pregnancy, pseudopregnancy, and lactation on the voluntary activity and food consumption of the albino rat. American Journal of Physiology 71: 362-394.

Smiley, Francis E. 1978 Changes in the Cursorial Ability of Wyoming Holocene Bison. Wyoming Contribution to Anthropology 1:105-126.

Smith, Bruce D. 1992 River of Change: Essays on Early Agriculture in Eastern North America. Smithsonian Institution Press, Washington, D.C.

Smith, Craig S. and L. M. McNees 1990 Rattlesnake Pass Site: A Folsom Occupation in South Central Wyoming. Plains Anthropologist 35:273-289.

368

Smith, Craig S., T. P. Reust, and R. D. Russell 2003 Site 48UT375: Late Paleoindian Period Subsistence and Land Use Patterns in the Green River Basin, Wyoming. Plains Anthropologist 48:133-151.

Smith, Eric A. 1982 Anthropological Applications of Optimal Foraging Theory: A Critic Review. Current Anthropology 24:625-651.

Smith, R. L. 1966 Ecology and Field Biology. Harper & Row, New York.

Snapp, R.R. and S. Bull 1944 Effect on Pregnancy on quality of beef. Illinois Agricultural Station Bulletin No. 508: 429-451.

Sokal, R.R. and F.J. Rohlf 1995 Biometry: The Principles and Practice of Statistics in Biological Research . W.H. Freeman and Company, New York.

Soper, J. Dewey 1941 History, range, and home life of the northern bison. Ecological Monographs 11: 347-412.

Speer, Roberta 1978 Bison Remains from the Rex Rodgers Site. In Bison Procurement and Utilization: A Symposium , edited by Davis, L.B. and M. Wilson, pp. 113-127. Plains Anthropologist Memoir 14, Lincoln.

Speth, John 1983 Bison Kills and Bone Counts. University of Chicago Press, Chicago . 1990 Seasonality, Resource Stress, and Food Sharing in So-called "Egalitarian" Foraging Societies. Journal of Anthropological Archaeology 9: 148-188. 1991 Taphonomy and early hominid behavior: Problems in distinguishing cultural and non-cultural agents. In Human Predator and Prey Mortality , edited by M. Stiner, pp. 31-40. Westview press, Boulder.

Speth, John D., and S. L. Scott 1985 The Role of Large Mammals in Late Prehistoric Horticultural Adaptations: the View from Southeastern New Mexico. In Contributions to Plains Prehistory , edited by D. Burley, pp. 233-266. Archaeological Survey of Alberta Occasional Paper No. 26.

Speth, John D. and Katherine Speilmann 1983 Energy Source, Protein Metabolism, and Hunter-Gatherer Subsistence Strategies. 369

Journal of Anthropological Archaeology 2:1-31.

Speth, John D., and E. Tchernov 2002 Middle Paleolithic Tortoise Use at Kebara Cave (Israel). Journal of Archaeological Science 59:471-483.

Stanford, Dennis J. 1984 The Jones-Miller Site: A Study of Hell Gap Bison Procurement and Processing. National Geographic Research Reports: 1975 Projects 16:615-635. 1997 The Walsh Cache. In The Paleoindians of the North American Midcontinent, edited by A. Montet-White. Musee Department de Prehistoire de Solutre, Solutre. 1999 Analysis and Interpretation of Hell Gap Hunting Strategies at the Jones-Miller Site. In Le Bison: Gibier et Moyen de Subsistence des Hommes du Paleolithique aux Paleoindiens des Grandes Plaines, edited by J.-Ph. Brugal, A. Davis J. Enloe, pp. 437-454. APDCA, Antibes.

Stanford, Dennis J. and Margaret A. Jodry 1988 The Drake Clovis Cache. Current Research in the Pleistocene 5: 21-22.

Stanford, D. J., and R. Patten 1984 R-6. A preliminary Report of A Cody Site in North-Central New Mexico. Papers of the Philmot Conference on the Archaeology of Northeastern New Mexico. New Mexico Archaeological Council Proceedings 6(1): 189-199.

Steele, D. Gentry, Joseph F. Powell 2002 Facing the Past: A View of the North American Human Fossil Record. In The First Americans: The Pleistocene Colonization of the New World, edited by N. G. Jablonski, pp. 93-122. Memoirs of the California Academy of Sciences Number 27. California Academy of Sciences, San Francisco.

Stephens, D. W. and E. A. Charnov 1982 Optimal Foraging: Some Simple Stochastic Model. Behavioral Ecology and Sociobiology 10:251-263.

Stephens, D. W. and J. R. Krebs 1986 Foraging Theory. Princeton University Press, Princeton.

Stiner, Mary C. 1990 The Use of Mortality Patterns in Archaeological Studies of Hominid Predatory Adaptations. Journal of Anthropological Archaeology 9:305-351. 1991 Food Procurement and Transport by Human and Nonhuman Predators. Journal of Archaeological Science 18:455-482. 1994 Honor Among Thieves: A Zooarchaeological Study of Neanderthal Ecology . Princeton University Press, Princeton. 370

2001 Thirty Years on the "Broad Spectrum Revolution" and Paleolithic Demography. The Proceedings of the National Academy of Sciences 98:6993-6996. 2002 Carnivory, Coevolution, and the Geographic Spread of the Genus Homo . Journal of Archaeological Research 10:1-63. 2003 Zooarchaeological Evidence for Resource Intensification in Algarve, Southern Portugal. Promontoria 1:27-58. 2004 A Comparison of Photon Densitometry and Computed Tomography Parameters of Bone Density in Ungulate Body Part Profiles. Journal of Taphonomy 2:117- 145. 2005 The Faunas of Hayonim Cave Israel: A 200,000-year Record of Paleolithic Diet, Demography, and Society . American School of Prehistoric Research Bulletin 48. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge.

Stiner, Mary C., and Natalie D. Munro 2002 Approaches to Prehistoric Diet Breadth, Demography, and Prey Ranking Systems in Time and Space. Journal of Archaeological Method and Theory 9:175-208.

Stiner, Mary C., Natalie D. Munro, and Todd A. Surovell 2000 The Tortoise and the Hare: Small-game use, the broad-spectrum revolution and Paleolithic Demography . Current Anthropology 41:39-73. 2005 Computer Simulation Modeling of Small Game Predator-Prey Dynamics. In The Faunas of Hayonim Cave, Israel: 200,000-year Record of Paleolithic Diet, Demography, and Society , by M. C. Stiner, pp. 149-163. American School of Prehistoric Research Bulletin 48. Peabody Museum of Archaeology and Ethnology Harvard University, Cambridge.

Stiner, Mary C., Natalie D. Munro, Todd A. Surovell, E. Tchernov, and O. Bar-Yosef 1999 Paleolithic Population Growth Pulses Evidenced by Small Animal Exploitation. Science 283:190-194.

Struiver, Minze and Paula J. Reimer 1993 Extended 14 C Data Base and Revised Calib 3.0 14 C Calibration Program. Radiocarbon 35:215-230.

Suhm, Dee A 1960 The Beidleman Ranch Site: An Early Man Kill Site in Stonewall County, Texas. Bulletin of Texas Archeological Society 31: 207-212.

Surovell, Todd A. 2000 Early Paleoindian Women, Children, Mobility, and Fertility. American Antiquity 65:493-509. 2003 The Behavioral Ecology of Folsom Lithic Technology . Ph.D. dissertation, University of Arizona, Tucson. University Microfilms, Ann Arbor. 371

Surovell, Todd A., Nicole Waguespack, and P. Jeffrey Brantingham 2005 Global Archaeological Evidence for Proboscidean Overkill. Proceeding of the National Academy of Science 102, 6231-6236.

Surovell, Todd A., Nicole Waguespack, Marcel Kornfeld, and George C. Frison 2003 The First Five Seasons at Barger Gulch, Locality B, Middle Park Colorado . Technical Report No. 26. George C. Frison Institute of Archaeology and Anthropology, Laramie.

Surovell, Todd A., Nicole Waguespack, James H. Mayer, Marcel Kornfeld, and George C. Frison 2005 Shallow Site Archaeology: Artifact Dispersal, Stratigraphy, and Radiocarbon Dating at the Barger Gulch Locality B Folsom Site, Middle Park, Colorado. Geoarchaeology 20: 627-649.

Szuter, Christine R. and Frank Bayham 1989 Sedentism and Prehistoric Animal Procurement among Desert Horticulturalists of the North American Southwest. In Farmers as Hunters The Implications of Sedentism , edited by Susan Kent, pp. 80-95. Cambridge University Press, Cambridge.

Taylor, Walter 1964 Tethered Nomadism and Water Territoriality: An Hypothesis. Acts of the 35th International Congress of Americanists: 197-203.

Thoms, A.V., D. D. Kuehn, B. W. Olive, J E. Dockall, P. A. Claybaugh, and R. D. Mandel 1996 Early and Middle Holocene Occupations at the Richard Beane Site: The 1995 South Texas Archaeological Association Field School Project. La Tierra 32: 8-36.

Tieszen, L.L., B.C. Reed, N. B. Bliss, B.K. Wylie, and D.D. DeJong 1997 NDVI, C 3, And C 4 Production, and Distributions in Great Plains Grassland Cover Classes. Ecological Applications 7:59-78.

Todd, Lawrence C. 1986 Determination of Sex of Bison Upper Forelimb Bones: The Humerus and Radius. Wyoming Archaeologist 29: 109-123. 1987 Taphonomy of the Horner II Bone Bed. In The Horner Site: the Type Site of the Cody Cultural Complex , edited by G.C. Frison, and L.C. Todd, pp. 107-198. Academic Press, Orlando, 1991 Seasonality Studies and Paleoindian Subsistence Strategies. In Human Predators and Prey Mortality , edited by M. C Stiner, pp. 215-238. Westview Press, Boulder.

372

Todd, Lawrence C. and Jack L. Hofman 1987 Bison Mandibles from the Horner and Finley Sites. In The Horner Site: the Type Site of the Cody Cultural Complex , edited by G.C. Frison and L.C. Todd, pp. 493- 539. Academic Press, Orlando.

Todd, Lawrence C., Jack L. Hofman, and C. Bertrand Schultz 1990 Seasonality of the Scottsbluff and Lipscomb Bison Bonebed: Implications for Modeling Paleoindian Subsistence. American Antiquity 55: 813-827. 1992 Faunal Analysis and Paleoindian Studies: A Reexamination of the Lipscomb Bison Bonebed. Plains Anthropologist 37:137-165.

Todd, Lawrence C., Matthew G. Hill, David J. Rapson, George C. Frison 1997 Cutmarks, Impacts, and Carnivores at the Casper Site Bonebed. In Proceedings of the 1993 Bone Modification Conference, Hot Springs, South Dakota , edited by L.A. Hannus et al.. pp. 136-157. Occasional Papers No. 1. Archaeology Laboratory, Augusta College. Sioux Falls.

Todd, Lawrence, C., David C. Jones. Robert S. Walker, Paul C. Burnett, and Jeffrey Eighmy 2001 Late Archaic Bison Hunters in the Northern Colorado: 1997-1999 Excavations at the Kaplan-Hoover Bison Bonebed (5LR3953). Plains Anthropologist 46:125- 147.

Todd, Lawrence C., David J. Rapson, and Jack L. Hofman 1996 Dentition Studies of the Mill Iron and Other Early Paleoindian Bison Bonebed Sites. In The Mill Iron Site , edited by G.C. Frison, pp. 145-176. University of New Mexico Press, Albuquerque.

Toigo, C., J. Galliard, G. Van Laere, M. Hewison, and N. Morellet 2006 How Does Environmental Variation Influence Body Mass, Body Size, and Body Condition? Roe Deer as a Case Study. Ecography 29:301-308.

Torbit, S. C., L. H. Carpenter, R. M. Bartmann, A. W. Alldredge, and G. C. White 1988 Calibration of carcass fat indices in wintering mule deer. Journal of Wildlife Management 52: 582-588.

Towne, E. G. 1999 Bison Performance and Productivity on Tallgrass Prairie. The Southwestern Naturalist 44: 361-366.

Tunnel, Curtis 1975 Fluted Projectile Point Production as Revealed by Lithic Specimens from the Adair-Steadman Site in Northwest Texas . Special Report 18. Office of the State Archaeologist, Texas Historical Commission, Austin. 373

Turpin, S., Leland C. Bement, and H. H. Eling, Jr. 1997 Stuck in the Muck: The Big Lake Bison Kill Site (41RG13), West Texas. In Southern Plains Bison Procurement and Utilization From Paleoindian to Historic , edited by Leland Bement and Kent J. Buehler, pp. 119-133. Plains Anthropologist Memoir 29 , Lincoln.

Ugan, Andrew 2005a Climate, Bone Density, and Resource Depression: What is Driving Variation in Large and Small Game in Fremont Archaeofaunas? Journal of Anthropological Archaeology 24:227-251. 2005b Does Size Matter? Body Size, Mass Collecting and Their Implications for Understanding Prehistoric Foraging Behavior. American Antiquity 70: 75-90.

United States Fish and Wildlife Service 1977 The American Buffalo. Conservation Notes: 12. Department of Interior Fish and Wildlife Washington, DC.

Van Vuren, Dirk 1983 Diet of bison and cattle on a seeded range in southern Utah. Journal of Range Management 36: 499-500.

Van Vuren, Dirk and M. P. Bray 1986 Population Dynamics of Bison in the Henry Mountains, Utah. Journal of Mammalogy 67:503-511.

Vereschagin, N.K. 1959 The Mammals of the Caucasus. The History of the Fauna . Academy of Science, Moscow.

Vehik, Susan C. 1977 Bone Fragments and Bone Grease Manufacturing: A Review of Their Archaeological Use and Potential. Plains Anthropologist 22: 169-182. 2006 Wichita Ethnohistory. In Kansas Archaeology , edited by R.J. Hoard and W. E. Banks, pp. 206-218. University Press of Kansas, Lawrence, Kansas.

Vickery, J. R. 1977 Influence on carcass Lipids of the Condition of Cows at Slaughter . Division of Food Research and Technology, Paper No. 42. Melbourne.

Waguespack, Nicole M. 2003 Clovis Hunting and the Organization of Subsistence Labor. Ph.D. dissertation, University of Arizona, Tucson. University Microfilms, Ann Arbor. 374

2005 The Organization of Male and Female Labor in Foraging Societies: Implications for Early Paleoindian Archaeology. American Anthropologist 107:666-676. 2007 Why We’re Still Arguing About the Pleistocene Occupation of the Americas. Evolutionary Anthropology 16:63-74,

Waguespack, Nicole M. and Todd Surovell 2003 Clovis Hunting Strategies, or How to Make Out on Plentiful Resources. American Antiquity 68:333-352.

Walde, D. 2004 Distinguishing Sex of Bison bison bison Using Discriminant Function Analysis. Canadian Journal of Archaeology 28:100-116.

Walker, Danny N 1975 A Cultural and Ecological Analysis of the Vertebrate Fauna from the Medicine Lodge Creek Site (48BH499). Unpublished Masters Thesis. Department of Anthropology, University of Wyoming, Laramie.

Waller, S.S., and J.K. Lewis 1979 Occurrences of C 3 and C 4 Photosynthetic Pathways in North American Grasses. Journal of Range Management 32:12-28.

Wang, G. H. 1935 The changes in the amount of daily food-intake of the albino rat during pregnancy and lactation. American Journal of Physiology 71: 736-741.

Warnica, J. M. and T. Williamson 1965 The Milnesand Site--Revisited. American Antiquity 33:16-24.

Wedel, Waldo R. 1959 An Introduction to Kansas Archaeology . Bureau of American Ethnology, Bulletin 174. Smithsonian Institution, Washington D.C. 1961 Prehistoric Man on the Great Plains . University of Okalahoma Press, Norman.

Weekes, Mary L. 1948 An Indian’s Description of the Making of a Buffalo Pond. Saskatchewan History 3: 14-17.

Westerling, D.B., and H. B. Hendrick 1979 Fatty acid composition of bovine lipids as influenced by diet, sex, and anatomical location and relationship to sensory characteristics. Journal of Animal Science 48: 1343-1348.

375

Wheat, Joe B. 1972 The Olsen-Chubbuck Site: A Paleoindian Bison Kill. Society for American Archaeology, Memoir No . 26. Washington, D.C. 1978 Olsen-Chubbuck and Jurgens Sites: Four Aspects of Paleo-Indian Economy. In Bison Procurement and Utilization: A Symposium , edited by L. B. Davis and M. Wilson, pp. 84-89. Plains Anthropologist Memoir 14. Lincoln 1979 The Jurgens Site . Plains Anthropologists Memoir 15. Lincoln.

Wheeler, R. P. 1995 Archaeological Investigations in Three Reservoir Areas in South Dakota and Wyoming. Part I Angostura Reservoir . Reprints in Anthropology Volume 46. J & L Reprint Company, Lincoln

White, Theodore E. 1953 A Method of Calculating the Dietary Percentage of Various Food Animals Utilized by Aboriginal Peoples. American Antiquity 18:396-398.

Whitney, C., G. B. Grinnell, and O. Wister 1904 Musk-ox, Bison, Sleep, and Goat . Macmillian, New York.

Whittaker, William 1997 The Cherokee Excavation Revisited: Prehistoric Adaptation and Change in the Eastern Great Plains. Unpublished Master’s Thesis. Department of Anthropology, University of Iowa. Iowa City.

Widga, Chris C. 2003 Human Subsistence and Paleoecology in the Middle Holocene Central Great Plains: The Spring Creek (25FT31) and Logan Creek (25BT3) Sites. Unpublished M.A. Thesis, University of Kansas, Lawrence. 2004 personal communication January 27, 2005. 2006a. Niche Variability in Late Holocene Bison: A Perspective from Big Bone Lick, KY. Journal of Archaeological Science 33:1237-1255. 2006b Bison, Bogs, and Big Bluestem: The Subsistence Ecology of Middle Holocene Hunter-Gatherers in the Eastern Great Plains. Unpublished Ph.D. dissertation, Department of Anthropology, University of Kansas, Lawrence.

Widga, Chris C., Lynn M. Snyder, and John R. Bozell 2000 The Logan Creek Site Fauna: Early Archaic Adaptations on the Central Great Plains. Report on file at the Nebraska State Historical Society, Lincoln.

Wien, John A. 1976 Population Responses to Patchy Environments. Annual Review of Ecology and Systematics 7:81-120.

376

Willey, G. R. 1966 An Introduction to American Archaeology Volume 1: North and Middle America . Prentice-Hall, Englewood Cliffs, NJ.

Williams, J. D. (editor) 2000 The Big Black Site (32DU955C): A Folsom Complex Workshop in Knife River Flint Quarry Area, North Dakota . Washington State University Press, Pullman.

Willey, P.S., B.R. Harrison, and J.T. Hughes 1978 The Rex Rodgers Site. In Archeology at MacKenzie Reservoir, edited by J.T. Hughes and P.S. Willey, pp. 51-68. Texas Historical Commission, Austin.

Williams, Jerry D. (editor) 2000 The Big Black Site (32DU955C): A Folsom Complex Workshop in Knife River Flint Quarry Area, North Dakota . Washington State University Press, Pullman.

Wilmsen, Edwin N. and Frank H. H. Roberts Jr. 1978 Lindenmeier, 1934-1974: Concluding Report on Investigations. Smithsonian Contribution to Anthropology No 24. Smithsonian Institution Press, Washington, D.C.

Wilson, Gregory A., Wes Olson, and Curtis Strobeck 2002 Reproductive Success in Wood Bison ( Bison bison athabascae) Established Using Molecular Techniques. Canadian Journal of Zoology 80:1537-1548.

Wilson, Michael 1974 Fossil Bison and Artifacts from the Mona Lisa Site, Calgary, Alberta, Part 1: Stratigraphy and Artifacts. Plains Anthropologist 19:34-45. 1975 Holocene Fossil Bison from Wyoming and Adjacent Areas. Unpublished MA Thesis, University of Wyoming, Laramie. 1978 Archaeological Kill Site Populations and the Holocene Evolution of the Genus Bison , edited by XXX, pp. 9-22. Plains Anthropologist 82, Memoir 14. 1980 Morphological Dating of Late Quaternary Bison on the Northern Plains. In The Ice-Free Corridor and Peopling the New World, edited by N.W. Rutter and C.E. Schweger, pp. 81-85. Canadian Journal of Anthropology 1.

Wilson, M. C., and A. C. MacWilliams 1987 Preliminary Geoarchaeological Studies at the Fletcher Paleo-Indian Bison kill Site (Alberta/Scottsbluff), Southern Alberta. Alberta Archaeological Review 14:12-15.

Windham, R. Peter, and F. A. Calabrese 1998 The Middle Missouri Tradition. In Archaeology on the Great Plains , edited by W. R. Wood, pp. 269-307. University Press of Kansas, Lawrence, Kansas.

377

Winkler, D.A. 1982 Re-evaluation of the Vertebrate Fauna from the Lewisville Archeological Site, Denton County, Texas. University of North Texas. Report submitted to U.S. Army Corp of Engineers, Dallas.

Winterhalder, Bruce 1986 Diet Choice, Risk, and Food Sharing in a Stochastic Environment. Journal of Anthropological Archaeology 5:369-392. 1997 Gift Given, Gifts Taken: The Behavioral Ecology of Non-Market, Intra-Group Exchange. Journal of Archaeological research 5:121-168. 2001 The Behavioural Ecology of Hunter-Gatherers. In Hunter-Gatherers: An Interdisciplinary Perspective, edited by C. Panter-Brick, R. H. Layton, and P. Rowley-Conwy, pp. 12-38. Cambridge University Press, Cambridge.

Winterhalder, Bruce and Flora Lu, 1997 A Forager-Resource Population Ecology Model and Implications for Indigenous Conservation. Conservation Biology 11:1354-1364.

Winterhalder, Bruce, Flora Lu, and Bram Tucker 1997 Risk-Sensitive Adaptive Tactics: Models and Evidence from Subsistence Studies in Biology and Anthropology. Journal of Archaeological Research 7:301-348.

Winterhalder, Bruce and Eric A. Smith 1981 (editors) Hunter-Gatherer Foraging Strategies. University of Chicago Press, Chicago. 2000 Analyzing Adaptive Strategies: Human Behavioral Ecology at Twenty-five. Evolutionary Anthropology 51-72.

Wissler, Clark 1920 North American Indians of the Plains , 2 nd Edition. Handbook Series No. 1. American Museum of Natural History, New York.

Wobst, H. M. 1974 Boundary Conditions for Paleolithic Social Systems: A Simulation Approach. American Antiquity 39:147-178.

Wolfe, M.L. and J.F. Kimball 1989 Comparison of Bison Population Estimates with a Total Count. Journal of Wildlife Management 53:593-596.

Wolverton, Steven 2001 Environmental Implications of Zooarchaeological Measures of Resource Depression. Unpublished Ph.D. dissertation, Department of Anthropology. University of Missouri-Columbia. Columbia. 378

Wolverton, Steven, J. H. Kenney, and J.D. Cornelius 2007 A Paleozoological Perspective on White-Tailed Deer ( Odocoileus virginianus texana ) Population Density and Body size in Central Texas. Environmental Management 39:545-552.

Wood, J.D. 1984 Fat deposition and quality of fat tissue in meat animals. In Fats in Animal Nutrition , edited by J. Wiseman, pp. 407-435. Butterworth, Boston.

Wormington, H. M. 1957 Ancient Man in North America. Denver Museum of Natural History, Popular Series No. 4. Denver. 1988 The Frazier Site. In Guidebook to the Archaeological Geology of the Colorado Piedmont and High Plains of Southeastern Wyoming , edited by V.T. Holliday, pp. 82-84. Department of Geography, University of Wisconsin-Madison, Madison.

Wyckoff, Donald G. 1989 Accelerator Dates and Chronology at the Packard Site, Oklahoma . Current Research in the Pleistocene 6: 24-26.

Yansa, C. H. 1998 Holocene Paleoenvironment and Paleohydrology of a Prairie Pothole in Southern Saskatchewan, Canada. Journal of Paleolimnology 19:429-441.

Yom-Tov, Y., and J. Yom-Tov 2005 Global Warming, Bergmann's Rule and Body Size in the Masked Shrew Sorex cinereus Kerr in Alaska. Journal of Animal Ecology 74:803-808.

Young, C.M. 1973 Plains Bison in Northern Ontario. Ontario Field Biology 27: 29-32.

Young, R. A. 1976 Fat, Energy and mammalian survival. American Zoologist 16: 699-710.