TRENDS IN SUBSISTENCE FROM THE MIDDLE PALEOLITHIC THROUGH MESOLITHIC AT KLISSOURA CAVE 1 (PELOPONNESE, GREECE) by Britt Marie Starkovich ________________________ A Dissertation Submitted to the Faculty of the SCHOOL OF ANTHROPOLOGY In Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOPHY In the Graduate College THE UNIVERSITY OF ARIZONA 2011 2 THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Britt Marie Starkovich entitled Trends in Subsistence from the Middle Paleolithic through Mesolithic at Klissoura Cave 1 (Peloponnese, Greece) and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy _______________________________________________________________________ Date: Dr. Mary C. Stiner April 26th, 2011 _______________________________________________________________________ Date: Dr. Steven L. Kuhn April 26th, 2011 _______________________________________________________________________ Date: Dr. Vance T. Holliday April 26th, 2011 _______________________________________________________________________ Date: Dr. Barnet Pavao-Zuckerman April 26th, 2011 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: Dissertation Director: Dr. Mary C. Stiner April 26th, 2011 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: Britt Marie Starkovich 4 ACKNOWLEDGEMENTS It is difficult to express my thanks and appreciation for all of the people in my life who have helped make the completion of this project possible. My friends and family have provided all manner of love and support, and this work is a culmination of all of their efforts. First of all, thank you to my committee. I managed to find a group of very smart people, who are also nice to be around. Mary Stiner is a fantastic advisor and has opened doors to me I never thought imaginable. Steve Kuhn knows everything and more about statistics and stone tools, and is still on top of pop culture references. Barnet Pavao-Zuckerman patiently answered bone ID questions in the beginning. Vance Holliday is one of the most down-toearth academics I have ever met, and once scoured west Texas for a BBQ joint with vegetarian options. Thanks to the Klissoura Cave 1 team for their excellent work on the site. Margarita Koumouzelis welcomed me with open arms, and together with Janusz Kozlowski granted me access to the materials. Takis Karkanas is brilliant, and answers emails with lightening speed. Thanks also to Piotr Wojtal, Teresa Tomek, Zbigniew Bocheński, Malgorzata Kot, Niko Thompson, Valery Sitlivy, Krzysztof Sobczyk, Georgia Tsartsidou, Maria Ntinou, and our Greek workman. Thanks to my colleagues in Greece not affiliated with the Klissoura project. Sherry Fox is reason enough to work at the Wiener Lab; the faunal collection happens to be a nice perk. Thanks to Mary Voyatzis and David Romano, and everyone on the Mt. Lykaion project. Thanks to my undergrad advisors, who helped get me into forager archaeology. Funding for this study came from a variety of sources, including the NSF IGERT program, the Reiker Grant and William Shirley Fulton Scholarship funds, the Wiener Laboratory at the American School for Classical Studies at Athens, an NSF dissertation improvement grant, and INSTAP, who funded the excavations. Thanks to my friends, for giving me a mental outlet. This includes, but is not limited to: Susan Mentzer, Tiina Manne, Nicole Flowers, Mary Good, Natalie Munro, Matt E. Hill, Rebecca Dean, Amy Clark, Amy Margaris, Shane Miller, Bill Reitze, Katrina Erickson, Jerome Guynn, Scott St. George, Jake Sutton, Aly Thibodeau, Kevin Anchukaitis, Toby Ault, Ben Curry, Todd Pitezel, Jess Conroy, Jess Driscoll, Jesse Ballenger, Jenny Leijonhufvud, Liye Xie, Derek Anderson, Josh Reuther, Lisa Janz, Jon Scholnick, Pat Wrinn, Matt E. Hill, Joe Beaver, Deanna Grimstead, Jessica Cerezo, Tom Fenn, Dana Drake Rosenstein, Kelly Jenks, Lauren Jelinek, A.J. Vonarx, Melanie Dedecker Medieros, Katie MacFarland, Ashley Blythe, Anita Carrasco, Brandon Gabler, Daniela Klöker, Rob Jones, Lizzie May, Jess Munson, and Ashley Stinnett. Thanks to John Olsen and Barbara Mills. The tireless efforts of the Anthropology staff have also been critical. Thanks to Ann Samuelson, Catherine Lehman, Ellen Stamp, Ben Beshaw, Norma Maynard, and especially Cathy Snider. You are the all the glue that holds us together. Thanks also to my family. Mom, Dad, Ryan, Christa, aunts, uncles, cousins, Grandma(s) and Grandpa, I love you all so much. Special thanks, of course, to Mathew Devitt, who has made this project much less stressful in all ways. Without complaint he has taken care of my beloved pets when I am doing research, and looked up esoteric sources when I call him from the middle of nowhere. While I was writing, he also made sure to water and turn me to the sunlight periodically. And of course, thanks to my furry children (Maggie, Lewis, Tyler, et al.), who keep things in perspective. 5 TABLE OF CONTENTS LIST OF FIGURES ...................................................................................................................9 LIST OF TABLES ...................................................................................................................14 ABSTRACT .............................................................................................................................16 CHAPTER 1: INTRODUCTION – QUESTIONS IN PALEOLITHIC ARCHAEOLOGY AND THEORETICAL BACKGROUND .........................................................................18 INTRODUCTION .......................................................................................................18 THE MIDDLE AND UPPER PALEOLITHIC: BIG QUESTIONS ...........................20 APPLICATIONS FROM EVOLUTIONARY ECOLOGY ........................................29 Prey Choice Model ................................................................................................31 Patch Choice Model ...............................................................................................36 Central Place Foraging Models ..............................................................................39 Predictions of Hominin-Prey Relationships...........................................................48 THE SAMPLE .............................................................................................................49 DISSERTATION STRUCTURE.................................................................................50 CHAPTER 2: KLISSOURA CAVE 1 AND THE GREEK PALEOLITHIC .........................52 INTRODUCTION .......................................................................................................52 NORTHERN GREECE ...............................................................................................54 Asprochaliko ..........................................................................................................54 Theopetra Cave ......................................................................................................58 Kastritsa .................................................................................................................61 Klithi ......................................................................................................................63 Boïla .......................................................................................................................65 Grava ......................................................................................................................66 Sarakenos Cave ......................................................................................................67 Sidari ......................................................................................................................68 Cave of Cyclope .....................................................................................................69 SOUTHERN GREECE ................................................................................................72 Kalamakia Cave .....................................................................................................72 Lakonis I ................................................................................................................74 Apidima..................................................................................................................77 Kephalari Cave.......................................................................................................78 Franchthi Cave .......................................................................................................79 THE LATE MIDDLE PALEOLITHIC THROUGH MESOLITHIC ARCHAEOLOGICAL SEQUENCE IN GREECE ...............................................85 KLISSOURA GORGE AND KLISSOURA CAVE 1 ................................................88 Stratigraphy and Dating of Klissoura Cave 1 ........................................................92 Stratigraphy ......................................................................................................96 Lithic Assemblages ..............................................................................................100 Botanical Remains ...............................................................................................105 6 TABLE OF CONTENTS - CONTINUED Features ................................................................................................................106 Ornaments ............................................................................................................108 SUMMARY AND GOALS .......................................................................................110 CHAPTER 3: LATE PLEISTOCENE CLIMATE AND ENVIRONMENT IN THE EASTERN MEDITERRANEAN ....................................................................................112 INTRODUCTION .....................................................................................................112 ENVIRONMENTAL HISTORY OF THE MEDITERRANEAN: MIS 5A THROUGH 1 .............................................................................................................113 ENVIRONMENTAL HISTORY OF GREECE: MIS 5A THROUGH 1 .................121 Northern and Central Greece ...............................................................................121 Southern Greece ...................................................................................................127 Environmental Conditions during the Occupation of Klissoura Cave 1 ..............132 BIOTIC COMMUNITIES AND RESOURCE AVAILABILITY ............................135 CONCLUSIONS........................................................................................................139 CHAPTER 4: TAPHONOMIC PROCESSES - BONE ACCUMULATION AND ATTRITION ....................................................................................................................143 INTRODUCTION .....................................................................................................143 METHODS ................................................................................................................145 RESULTS ..................................................................................................................162 Input Analysis and Fragmentation .......................................................................162 Carnivore Damage and Weathering .....................................................................164 Density-Meditated Attrition .................................................................................168 DISCUSSION AND CONCLUSIONS .....................................................................173 CHAPTER 5: RELATIVE SPECIES ABUNDANCE AND DIVERSITY OF PREY TYPES .............................................................................................................................177 INTRODUCTION .....................................................................................................177 METHODS ................................................................................................................177 PREY REPRESENTATION IN THE MIDDLE PALEOLITHIC THROUGH MESOLITHIC LAYERS AT KLISSOURA CAVE 1 ........................................183 PREY DIVERSITY: RICHNESS AND EVENNESS ...............................................193 BIOMASS COMPARISON.......................................................................................200 CONCLUSIONS........................................................................................................204 CHAPTER 6: PREY SELECTION, TRANSPORT DECISIONS AND BUTCHERY DAMAGE ........................................................................................................................214 INTRODUCTION .....................................................................................................214 BURNING DAMAGE ...............................................................................................216 Burning Damage on Small Game ........................................................................222 Burning Damage on Ungulates ............................................................................224 7 TABLE OF CONTENTS – CONTINUED BUTCHERY DAMAGE AND OSSEOUS TOOLS: CUTS, CONES, IMPACTS, WORKING, AND OTHER FRACTURES .........................................................229 Butchery Damage on Small Mammals and Carnivores .......................................230 Butchery Damage on Ungulates ..........................................................................233 Worked Bone and Antler .....................................................................................237 BONE PROCESSING ...............................................................................................239 Ungulate Marrow Processing ...............................................................................241 PREY TRANSPORT .................................................................................................245 Utility Indices.......................................................................................................245 Body Part Profiles ................................................................................................252 Small Game Body Part Profiles ...........................................................................253 Ungulate Body Part Profiles ................................................................................259 MORTALITY PROFILES.........................................................................................265 CONCLUSIONS........................................................................................................276 CHAPTER 7: SPATIAL VARIATION WITHIN CULTURAL LAYERS ..........................280 INTRODUCTION .....................................................................................................280 METHODS ................................................................................................................281 RESULTS ..................................................................................................................284 Mesolithic layers 3-5a ..........................................................................................285 Epigravettian layers IIa-d .....................................................................................286 Mediterranean backed bladelet industry layer III’ ...............................................287 Upper Paleolithic (non-Aurignacian) industry layer III” .....................................289 Upper Aurignacian layers IIIb-d ..........................................................................290 Middle Aurignacian layers IIIe-g.........................................................................291 Lower Aurignacian layer IV ................................................................................293 Early Upper Paleolithic or Uluzzian layer V .......................................................296 Middle Paleolithic layer VIII ...............................................................................298 Middle Paleolithic layer X ...................................................................................299 Middle Paleolithic layers XI-XIV ........................................................................299 Middle Paleolithic layers XV-XVII .....................................................................301 Middle Paleolithic layers XVIII-XIX ..................................................................303 Middle Paleolithic layer XXa-b ...........................................................................304 DISCUSSION AND CONCLUSIONS .....................................................................305 CHAPTER 8: KLISSOURA CAVE 1 IN CONTEXT: LATE PLEISTOCENE SUBSISTENCE CHANGE IN GREECE AND THE MEDITERRANEAN BASIN .....308 INTRODUCTION .....................................................................................................308 KLISSOURA CAVE 1 ..............................................................................................308 Changes in Site Use and Occupation Intensity ....................................................313 RESULTS FROM THE PERSPECTIVE OF EVOLUTIONARY ECOLOGY .......317 Prey Choice ..........................................................................................................317 Central Place Foraging .........................................................................................318 8 TABLE OF CONTENTS – CONTINUED Results on Patch Choice based on Carcass Processing Intensity .........................321 KLISSOURA CAVE 1 WITHIN GREECE AND THE MEDITERRANEAN BASIN..................................................................................................................324 Intensification and Specialization ........................................................................335 CONCLUSION ..........................................................................................................338 APPENDICES .......................................................................................................................341 Appendix A: Hare survivorship and bone density values (values from Pavao and Stahl 1999) ..........................................................................................................................341 Appendix B: Fallow deer survivorship and bone density values (values from Lyman 1994) ..........................................................................................................................342 Appendix C: Tooth and bone-based MNE values by layer .............................................346 Appendix D: Shaft and end-based MNE values by layer ................................................347 Appendix E: Middle Paleolithic hare elements. Horizontal line indicates lowest structural density for cervids ......................................................................................................348 Appendix F: NISP and MNE values by layer for all taxon .............................................349 Appendix G: Biomass values for common taxa ..............................................................351 Appendix H: Utility indices for Rangifer, combined by anatomical region (values from Binford 1978) .............................................................................................................354 Appendix I: Percent MAU for fallow deer by layer ........................................................355 Appendix J: Mean FUI for fallow deer by layer ..............................................................358 Appendix K: Fallow deer anatomical representation.......................................................360 Appendix L: Hare anatomical representation ..................................................................362 Appendix M: Partridge anatomical representation ..........................................................363 Appendix N: Anatomical representation for red deer and ibex in layer IV .....................364 Appendix O: Tooth wear stages for ungulates by layer ...................................................365 Appendix P: Ungulate fusion stages (values from Reitz and Wing 2008 and Silver 1969) ................................................................................................................367 WORKS CITED ....................................................................................................................370 9 LIST OF FIGURES Figure 2.1. Map of Greece with Paleolithic and Mesolithic sites mentioned in the text .........55 Figure 2.2. Calibrated radiocarbon dates and ranges of dates for sites mentioned in the text. Cultural periods are listed above their respective dates. Where radiocarbon dates are unavailable, placement of cultural periods are estimated based on dated industries in Greece ................................................................................................................................57 Figure 2.3. Photograph of Klissoura Gorge, facing southwest ................................................89 Figure 2.4. Comparison of modern landmass (grey fill) to LGM shorelines (black line and white fill, indicating land inundated today) and paleolakes (black fill). Adapted from Petit-Maire et al (2005) ......................................................................................................90 Figure 2.5. Photograph of Klissoura Cave 1, facing northwest ...............................................91 Figure 2.6. Plan map of Klissoura Cave 1. Adapted from Karkanas (2010) ...........................92 Figure 3.1. Marine oxygen isotope chronology, adapted from Martinson et al. (1987). Bar at the top indicates period that were warm (w), cold (c) and mild (m), compared to surrounding periods .........................................................................................................115 Figure 3.2. Simplified schematic of Greek paleoclimate studies mentioned in the text. Dark bars = warm, wet (forest expansion), grey bars = mild (mixed forest-steppe), white bars = dry, open steppe. Core locations from studies mentioned in this figure are mapped in figure 3.3. Bottom axis corresponds to MIS stages in Figure 3.1 ....................................120 Figure 3.3. Map of Greece indicating locations of pollen cores discussed in the text, and in Figure 3.2 .........................................................................................................................122 Figure 4.1. Metallic blue patina on Middle Paleolithic long bone fragment. From layer XIX ..................................................................................................................................146 Figure 4.2. (a) Light carbonate fraction on great bustard tibiotarsus. From disturbed zone (layer 6a). (b) Heavy carbonate fraction on great bustard lumbar vertebrae. From disturbed zone (layer 6) ...................................................................................................147 Figure 4.3. Chemical weathering on Middle Paleolithic long bone fragment. From layer XIV ..................................................................................................................................167 Figure 4.4. Proportion of tooth to bone-based MNE for all layers at Klissoura Cave 1. Data from Table 4.6..................................................................................................................169 Figure 4.5. Proportion of shaft-based and end-based MNE for all layers at Klissoura Cave 1. Data from Table 4.8 .........................................................................................................172 Figure 5.1. NISP counts for the major taxa at Klissoura Cave 1 by layer. Taxa are in descending order of average mass ...................................................................................186 Figure 5.2. NISP counts for the major taxa at Klissoura Cave 1, grouped by body class .....187 Figure 5.3. Proportion of small game to ungulate NISP, plotted against time. The relationship is significant and positive. Data from Table 5.2 ..............................................................189 Figure 5.4. Relative NISP counts for small game animals by layer. White fill indicates small, slow-moving species, grey or black fill indicates small, fast-moving game ...................190 Figure 5.5. Proportion of small slow to small fast game NISP, plotted against time. The relationship is significant and negative. Data from Table 5.2 .........................................191 Figure 5.6. NISP counts for carnivore species at Klissoura Cave 1 by cultural layer ...........192 Figure 5.7. Ratio of logNISP to logNtaxa in the Klissoura Cave 1 assemblage. The relationship is significant and positive. Data from Table 5.3 ..........................................194 10 LIST OF FIGURES – CONTINUED Figure 5.8. Inverse Simpson’s index for all taxa plotted against time. No statistically significant relationship exists, but there is a general increase in variability between assemblages in the Upper Paleolithic. Data from Table 5.4 ............................................197 Figure 5.9. Inverse Simpson’s index for all ungulate taxa plotted against time. No statistical relationship is apparent, though certain Upper Paleolithic and later assemblages are significantly more even than the other layers. Data from Table 5.4 ................................198 Figure 5.10. Inverse Simpson’s index for small game species plotted against time. No temporal trend exists, but note the cluster of points that represent higher small game diversity. The earlier point indicates a more even representation of hares and tortoises, while the later points indicate layers with an even representation of hares and birds (great bustard and rock partridge). Data from Table 5.4............................................................199 Figure 5.11. Estimated biomass for each important taxa at Klissoura Cave 1 by layer. Note that small game biomass is calculated based on MNI, while large game biomass is calculated using MNE ......................................................................................................201 Figure 5.12. Proportion of small game to ungulate biomass, plotted against time. The relationship is significant and positive. Data from Table 5.5 ..........................................202 Figure 5.13. Proportion of small slow to small fast game biomass, plotted against time. The relationship is significant and negative. Data from Table 5.5 .........................................204 Figure 6.1. Frequencies of burned cranial and postcranial specimens at Klissoura Cave 1. In the Upper Paleolithic layers crania are more commonly burned .....................................228 Figure 6.2. Tortoise carapace fragment with crushing and an impact fracture. From Middle Paleolithic layer XVI .......................................................................................................232 Figure 6.3. Example of cut marks on medium ungulate remains. Specimen from Middle Paleolithic layer XII .........................................................................................................236 Figure 6.4. Cut frequencies on fallow deer elements in the Middle Paleolithic layers at Klissoura Cave 1. Includes elements with n > 10 cuts ....................................................237 Figure 6.5. Worked bone from Middle Paleolithic layer XIV ...............................................238 Figure 6.6. Percent of unopened first, second and third phalanges through the layers at Klissoura Cave 1. First phalanges, which contain the highest marrow content, are consistently processed through the layers. Second phalanges were opened more often in the late Middle Paleolithic and Aurignacian layers. Terminal phalanges were only intensively processed during the late Middle Paleolithic and Aurignacian layer IV (data from Table 6.11) ..............................................................................................................244 Figure 6.7. Scatter plots of %MGUI and %MAU values for different layers at Klissoura Cave 1 (following Binford 1978) ..............................................................................................250 Figure 6.8. Mean FUI for each layer, following Broughton (1999). Horizontal line indicates the average FUI for a complete cervid. Data from Appendix J .......................................252 Figure 6.9. Anatomical profiles for partridges in the Mediterranean backed-bladelet industry (III’) and upper Aurignacian (IIIb-d) (data from Appendix M) ......................................254 Figure 6.10. Anatomical profiles for hares in the Upper Paleolithic and later layers and Middle Paleolithic (XI-XIV). Note the lack of foot elements in the UP and Mesolithic and neck and axial elements in all layers (data from Appendix L) .................................255 11 LIST OF FIGURES – CONTINUED Figure 6.11. Plot of the foot region of hares from Klissoura cave 1 by layer. MAU (or standardized MNE) values for elements included in the foot are presented in order of descending structural density from the most dense scan site of each element. The expected MAU is presented based on the highest MNE value for all skeletal elements for each layer. Bone density values from Pavao and Stahl (1999) ........................................258 Figure 6.12. Plot of the axial region of hares from Klissoura cave 1 by layer. MAU (or standardized MNE) values for axial elements are presented in order of descending structural density from the most dense scan site of each element. The expected MAU is presented based on the highest MNE value for all skeletal elements for each layer. The head region is also plotted to indicate that in some layers it is present though other axial elements are not. Bone density values from Pavao and Stahl (1999) ..............................258 Figure 6.13. Body part profiles for fallow deer in the layers with large sample sizes, ranked by increasing GUI values (data from Appendix K) .........................................................261 Figure 6.14. Body part profiles for red deer and ibex in lower Aurignacian layer IV, ranked by increasing GUI values (data from Appendix N) .........................................................262 Figure 6.15. Body part profiles for fallow deer in the layers with large sample sizes, collapsed into regions with low, medium, high and highest GUI values (from Table 6.16). ................................................................................................................................264 Figure 6.16. Tripolar graph indicating living structure and mortally models in an ungulate death assemblage .............................................................................................................267 Figure 6.17. Proportions of ungulate age groups exploited at Klissoura Cave 1 during the Middle and Upper Paleolithic based on tooth eruption and wear. During the Middle Paleolithic, prime-aged adult animals were targeted. More juveniles were incorporated in Upper Paleolithic subsistence pursuits. Circles indicate 95% confidence intervals (following Weaver et al. 2011) (data from Table 6.17) ...................................................270 Figure 6.18. Proportions of fallow deer age groups exploited at Klissoura Cave 1 during the Middle and Upper Paleolithic based on tooth eruption and wear. During the Middle Paleolithic, prime-aged adult animals were targeted. Upper Paleolithic patterns are closer to the living structure of fallow deer. Circles indicate 95% confidence intervals (following Weaver et al. 2011) (data from Table 6.17) ...................................................270 Figure 6.19. Percentages of juveniles based on five selected elements, based on epiphyseal fusion, following Stiner (2005), figure 11.12. Selected elements fuse between 12 and 24 months (Kersten 1987). No major differences exist between the ages of juvenile animals exploited in the Middle and Upper Paleolithic (data from Table 6.17) ...........................272 Figure 6.20. Comparison of fallow deer fetal remains from Klissoura Cave 1 and a modern stillborn goat. (top) Fetal femur from Upper Paleolithic disturbed zone (6-7a), (bottom) Fetal humeri from Middle Paleolithic layers (VII) (bottom) and (XVa-XVI) (top) ........274 Figure 6.21. Proportion of fetal to adult fallow deer by MNI. Data from Table 6.19 ...........276 Figure 7.1. Plan maps for Mesolithic (3-5a), NISP = 256. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths ......................................................................286 12 LIST OF FIGURES – CONTINUED Figure 7.2. Plan maps for Epigravettian (IIa-d), NISP = 67. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths ..........................................................................287 Figure 7.3. Plan maps for Mediterranean backed-bladelet industry (III’), NISP = 1,359. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths..........................288 Figure 7.4. Plan maps for Upper Paleolithic (non-Aurignacian) industry (III”), NISP = 1,617. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths..........................289 Figure 7.5. Plan maps for Aurignacian (IIIb-d), NISP = 1,158. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths ..........................................................................291 Figure 7.6. Plan maps for Aurignacian (IIIe-g), NISP = 1,699. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths ..........................................................................292 Figure 7.7. Plan maps for Aurignacian (IV), NISP = 3,110. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths ..........................................................................293 Figure 7.8. Drafted plan maps in Aurignacian (IV) from 145 to 180 cm below datum by 5cm cuts. Hearth features are indicated in black and dark grey, limestone rocks in white. The light grey background represents sedimentary matrix. The shelter feature is apparent in cuts 145 to 170, where rocks are common and hearths are absent. From Stiner (2010) .295 Figure 7.9. Plan maps for Aurignacian layers (IV) containing the structure, NISP = 2,049. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. Units containing the structure are outlined in black...............................................................................................................296 Figure 7.10. Plan maps for Early Upper Paleolithic (Uluzzian V), NISP = 222. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths .......................................297 Figure 7.11. Plan maps for Middle Paleolithic (VIII), NISP = 1,569. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths...................................................298 Figure 7.12. Plan maps for Middle Paleolithic (X), NISP = 241. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters ..............................................................................................................300 Figure 7.13. Plan maps for Middle Paleolithic (XI-XIV), NISP = 1,779. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters ...........................................................................301 Figure 7.14. Plan maps for Middle Paleolithic (XV-XVII), NISP = 2,031. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters ...........................................................................302 13 LIST OF FIGURES – CONTINUED Figure 7.15. Plan maps for Middle Paleolithic (XVIII-XIX), NISP = 2,199. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters ...........................................................................303 Figure 7.16. Plan maps for Middle Paleolithic (XXa-XXb), NISP = 1,106. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning, (d) medium ungulate specimens by body region. Darker boxes indicate higher frequencies of listed Parameters ........................................................................................................................305 Figure 8.1. Illustration of changes through time in ungulate diversity, occupation intensity and small game use at Klissoura Cave 1 ..........................................................................315 Figure 8.2. Map of the four quadrants of the Mediterranean. Shaded areas indicate Mediterranean ecosystem. From Blondel and Aronson (1999) .......................................324 14 LIST OF TABLES Table 2.1. Stratigraphic layers for Klissoura Cave 1 and corresponding cultural units based on stone tool types. Layer groupings are used throughout the dissertation for analyzing faunal remains. From Kaczanowksa et al. 2010, Karkanas 2010, personal communication 2009, Kuhn et al. 2010 .......................................................................................................93 Table 2.2. Radiocarbon dates from Klissoura Cave 1. From Kuhn et al. (2010) ....................95 Table 3.1. Taxa expected in the vicinity of Klissoura Cave 1 by MIS stage. Note that the MIS stage of the lower layers are unknown, which is indicated by "/" and both options are included. MIS 3 is highly variable and local reconstructions are available for the UP layers but not the MP. P = present, A = absent, ? = unsure .............................................133 Table 4.1. Fallow deer metatarsal NISP counts and NISP:MNE ratios that include anterior groove fragments and corrected counts that exclude anterior groove fragments, compared to metacarpal NISP counts and ratios. Metacarpals were used for comparison because their structural density is similar and they are equally recognizable in an assemblage as are metatarsals. NISP counts for metatarsals exclude anterior grooves for taxonomic frequencies because they over-represent fallow deer frequencies. Only Middle Paleolithic layers are depicted because they are the most extreme, but cervid metatarsal anterior grooves (and other overly-recognizable fragments such as unidentifiable tooth and tortoise shell fragments) are excluded in all layers ..........................................................150 Table 4.2. Rates of faunal input (or faunal density) by layer ................................................151 Table 4.3. Mean fragment length for different body classes, by cultural layer .....................164 Table 4.4. MNE:NISP index values for medium ungulate cranial elements and all elements in the Klissoura Cave 1 assemblage. *Note small sample sizes ..........................................165 Table 4.5. Frequencies of gnawing and weathering damage on faunal specimens by cultural layer..................................................................................................................................166 Table 4.6. Tooth and bone-based MNE for all ungulates by layer ........................................169 Table 4.7. Spearman’s rank-order correlation between percent survivorship and bone density values for European hare and fallow deer/medium ungulate remains. Data from Appendices A and B. *Significant values........................................................................171 Table 4.8. End and shaft-based MNE values for medium ungulate long bones by cultural layer. Elements include humerus, radius, ulna, metacarpal, femur, tibia and metatarsal .........................................................................................................................172 Table 5.1. Taxa included in each body class category and their respective weight ranges and predicted ranking. Mass values from Nowack (1999) and Silva and Downing (1995) ...179 Table 5.2. Proportion of NISP counts per layer for small game: large game (ungulates), and small slow: small fast-moving prey types. *Small samples excluded from Figure 5.3 ...188 Table 5.3. LogNISP and logN-taxa values for all species, ungulates, and small game Species .............................................................................................................................195 Table 5.4. Inverse Simpson’s index per layer for all taxa, ungulates, and small game species. *Small sample sizes excluded from Figure 5.10 .............................................................196 Table 5.5. Proportion of biomass values per layer for small game: large game (ungulates), and small slow: small fast-moving prey types. *Small samples excluded from Figure 5.13...................................................................................................................................203 15 LIST OF TABLES – CONTINUED Table 5.6. Taxa found in Klissoura Cave 1 by cultural layer, with predictions from Table 3.1 in parentheses. P = present, A = absent, U = uncommon (NISP < 5). Grey shading indicates deviations from predictions. Roe deer and chamois are excluded because of small samples overall. *Epigravettian NISP = 67 ...........................................................206 Table 6.1. Burn frequency and degree for different classes of animals at Klissoura Cave 1 by layer..................................................................................................................................218 Table 6.2. Burn frequencies for tortoise shells in layers with larger sample sizes ................223 Table 6.3. Mean length for burned and unburned identifiable specimens by layer ...............224 Table 6.4. Burn frequency and degree for cranial and postcranial elements by layer for medium ungulates. *Small samples not included in Figure 6.1 ......................................226 Table 6.5. Butchery damage on small game at Klissoura Cave 1 by layer ............................231 Table 6.6. Transverse fractures on small animal limb bones (humerus, radius, femur and tibia/tibiotarsus) by layer .................................................................................................233 Table 6.7. Butchery damage on medium and large ungulate remains. *Note small sample sizes ..................................................................................................................................234 Table 6.8. Cut marks on medium ungulate elements in Middle Paleolithic layers ...............236 Table 6.9. Counts of all antler and worked antler by layer, and corresponding ungulate NISP counts for comparison ......................................................................................................239 Table 6.10. Percent of medium ungulate limb bones not opened prior to discard, by layer .242 Table 6.11. Percent of unopened medium ungulate phalanges for layers with MNE values larger than 10 ...................................................................................................................243 Table 6.12. Pearson’s correlation values between Binford’s (1978) utility indices. * Indicates significant relationships ...................................................................................................247 Table 6.13. Spearman’s rank-order correlation values for fallow deer %MAU and caribou utility indices (from Binford 1978). *Statistically significant relationships....................249 Table 6.14. K-S tests for hare body part profiles by layer. *Asymp. Sig. > 0.05, data do not differ from a uniform distribution ....................................................................................255 Table 6.15. K-S tests for ungulate body part profiles by layer. *Asymp. Sig. > 0.05, data do not differ from a uniform distribution ..............................................................................262 Table 6.16. Cervid anatomical regions (following Stiner 1991), ranked in order of GUI values derived for caribou (from Binford 1978) ..............................................................263 Table 6.17. Age distribution for ungulates in the Klissoura Cave 1 assemblages, based on tooth eruption data (see Appendix O for more detail) .....................................................269 Table 6.18. Fusion schedules for five elements for fallow deer. Note radii are excluded from Figure 6.19 because of small samples in the Upper Paleolithic layers ............................273 Table 6.19. Fetal elements identifiable to species by layer ...................................................275 16 ABSTRACT This study presents an analysis of the zooarchaeological remains from Klissoura Cave 1, a Middle Paleolithic through Mesolithic site in Peloponnese, Greece. Changes in subsistence patterns are evaluated across a long sequence (ca. 80,000-10,000 BP) against a backdrop of environmental change. Results are interpreted using models from evolutionary ecology, specifically prey choice, central place foraging, and patch choice models. Two major trends are apparent in the series. One is a decline in the exploitation of high-ranked ungulate species with an overall increase in lower-ranked small game animals. The second is an increase in low-ranked small, fast-fast moving animals (e.g., hares and partridges) at the expense of higher-ranked small, slow-moving animals (e.g., tortoises). These changes cannot be accounted for by environmental shifts alone, though shifts in ungulate diversity likely track the expansion and contraction of plant communities. The increase in use of low-ranked prey indicates human population growth and demographic pressure in southern Greece during the late Pleistocene and early Holocene. In addition to these overarching trends, there are changes in site use during the sequence. In the Middle Paleolithic, foragers used Klissoura Cave 1 more during the winter and overwhelmingly hunted prime-aged adult animals, maternal herds of fallow deer in particular. In the Upper Paleolithic and later periods, the site continued to be used during the winter, in addition to other times of year, but the mortality profiles reflect a natural fallow deer herd structure. There was an intense period of occupation during the Aurignacian period. This is evidenced by numerous clay-lined hearth features, a possible rock-lined structure, and increases in ornaments, as well as abundant lithic and faunal materials. The ungulate faunas are particularly rich during this 17 period, but there is evidence of resource intensification based on increased bone marrow processing and the transport of marrow-rich elements to the site. After this period there was a gradual decline in site use through the end of the Upper Paleolithic and into the Mesolithic, though the exploitation of low-ranked resources (e.g., small, fast-moving game) indicates that populations were on the rise in there region as a whole. 18 CHAPTER 1: INTRODUCTION - QUESTIONS IN PALEOLITHIC ARCHAEOLOGY AND THEORETICAL BACKGROUND INTRODUCTION On a global scale, the Middle and Upper Paleolithic were extremely dynamic periods, and change accelerated over time as hominin populations expanded into new regions and local cultural diversity increased. Human population densities fluctuated dynamically as anatomically modern humans left Africa and pushed into regions already occupied by Eurasian hominins. Stone tool technologies became more complex and locally distinct. Hominin subsistence strategies shifted in response to environmental change, tighter packing of humans on the landscape, and technological innovations that allowed for the safer hunting of large game (e.g., projectile weaponry) and the more efficient exploitation of small game (e.g., snares, nets and fish hooks). Durable artistic and symbolic expression appeared in Eurasia and became ubiquitous during the Upper Paleolithic, including shell and tooth ornaments, ochre-stained tools and bone, figurines, and eventually cave paintings. In general it was a time of tremendous social and cultural upheaval. As archaeologists have tried to make sense of variation within and between the Middle and Upper Paleolithic, their jobs are complicated by an archaeological record that is often geographically and temporally patchy, dating techniques with low levels of precision, and a lack of hominin skeletons that make it difficult to assign stone tool industries to specific populations. It is unrealistic to assume that any work, no matter how ambitious, could solve all of these problems, but every well-designed research project can add to our understanding of the final chapters of the evolutionary history of human foragers. This introduction serves as a 19 backdrop for a discussion of human dietary change across the late Pleistocene and early Holocene from Klissoura Cave 1, a small rock shelter in southern Greece. The archaeological record at Klissoura Cave 1 is important for several reasons. First is the overall lack of long, intact archaeological sequences in Greece, which is discussed in detail in Chapter 2. Second, is the importance of Greece in general to our understanding of hominin lifeways and population movements during the late Pleistocene. The geographic significance of Greece cannot be overemphasized. Along with Iberia and Italy, it is one of three major European peninsulas, and there is ample evidence that some of these regions served as refugia during the Pleistocene for plant, animal, and hominin populations (see below). Further, its location in eastern Europe makes Greece key in understanding the western expansion of modern human populations. The Klissoura Cave 1 record is equally important for understanding variation within cultural periods as it is for examining broad changes between periods. However, many changes concerning Paleolithic archaeology are framed in terms of understanding the socalled transition between the Middle and Upper Paleolithic. As such, I review some arguments and debates pertaining to difference between the two periods, and the difficulties of characterizing “transitional” archaeological phases in Europe. Often, these discussions revolve around the behavioral dichotomy (or not) between Neandertals and anatomically modern humans, whose cultural remains are presumably both represented at Klissoura Cave 1. An absence of hominin skeletal remains at the site, however, makes this dichotomy difficult to explore. Therefore, this study focuses on changes in diet and subsistence practices within and between cultural phases, regardless of whether they are associated with changes in hominin populations or lithic technologies. These changes are explored throughout this 20 dissertation using theoretical tools from evolutionary ecology, which are expanded upon in the latter part of this chapter. THE MIDDLE AND UPPER PALEOLITHIC: BIG QUESTIONS Many complex questions surround attempts to understand the Middle and Upper Paleolithic in Eurasia, most of which focus on the nature and meaning of the transition between the two periods. The nature of the transition is contentious and difficult to understand because there was no “moment” that indicates the shift to behavioral and cultural modernity (Clark 2002; Hovers and Belfer-Cohen 2006; McBrearty and Brooks 2000). The general timing of the transition seems to have varied, with evidence of behavioral modernity appearing earlier in Africa and later in southwest Asia and Europe (see McBrearty and Brooks 2000 and references therein). Another question seeks to determine exactly who the Neandertals were, a local population of Homo sapiens that contributed, albeit minimally, to the modern human gene pool (Ahern, et al. 2004; Clark 2002; Trinkhaus 2007; Zilhão 2006), or a behaviorally and technologically stagnant “other” that sometimes relied on anatomically modern humans for cultural innovations (d'Errico, et al. 1998; Mellars 1973, 1989, 1991). The abilities of Neandertals to hunt efficiently, create art, and innovate stone tool technologies have been widely questioned (Mellars 1973, 1989, 1991). The extent to which populations of modern Homo sapiens from Africa interacted with European Neandertals is also unclear, whether there was interbreeding (but see Green, et al. 2010), reluctant tolerance, the transmission of cultural knowledge, or outright hostility. It is also not known exactly how long Neandertal populations persisted in the refugia of the various southern European peninsulas (Carbonell, et al. 2000; Conard and Bolus 2003; d'Errico and Sanchez Goñi 2003; 21 d'Errico, et al. 1998; Finlayson, et al. 2006; Kuhn and Biette 2000; Tzedakis, et al. 2007; Walker, et al. 2008; Zilhão and d'Errico 1999; Zilhão and Pettitt 2006). What is apparent, however, is that human population densities increased around the world within the interval between the late Middle Paleolithic and the mid-Upper Paleolithic (e.g., Lahr and Foley 2003), and a range of cultural adaptations appeared, at least partly in response to population growth against a backdrop of climatic change. The current evidence for the emergence of durable forms of artistic expression, new stone tool chronologies, bioarchaeological data, and subsistence change are reviewed briefly below. One of the hallmarks of behavioral modernity is arguably the emergence of symbolic art. The lack of artistic expression at European Middle Paleolithic sites was long considered as evidence for cognitive and social differences between anatomically modern humans and Neandertal populations in particular (Mellars 1973, 1989, 1991). Ochre and black manganese occur throughout the Middle Paleolithic record, the uses of which are unknown. Evidence of artistic expression is very rare or absent in Middle Paleolithic contexts, though shell ornaments, ochre-stained artifacts and lumps of ochre are reported in a handful of sites outside of Africa, from Spain to the Levant and dating to between 135,000 and 50,000 BP (Bar-Yosef Mayer, et al. 2009; Vanhaeren, et al. 2006; Zilhão, et al. 2010). All early contexts containing Middle Paleolithic artistic expression, except those reported from Spain, are associated with anatomically modern humans. There is a difference between the sporadic artistic expression claimed for some Middle Paleolithic cases and its ubiquity and unmistakable form in Upper Paleolithic and later contexts (e.g. Kuhn and Stiner 2007; Kuhn, et al. 2001; Stiner 2010). It is no longer fair to assume that Neandertals and other Middle 22 Paleolithic populations were cognitively incapable of creating symbolic expression, but the lack of widespread durable examples is very difficult to explain or ignore. The chronology of the transition from the Middle to Upper Paleolithic varies regionally, beginning earlier in southwest Asia and occurring the latest in western Europe. The transition began between roughly 50,000 and 45,000 BP in the Levant (Bar-Yosef 2000; Bar-Yosef 2007), and gradually spread through Eurasia. The transition in the Caucasus, for example, occurs between about 35,000 and 34,000 BP (Adler, et al. 2006; Meshveliani, et al. 2004). The Middle Paleolithic persisted in Eastern Europe until between 47,000 and 40,000 (Kozlowski 2000; Kozlowski 2007), before the Eastern European Aurignacian became widespread. The transition to the Aurignacian occurred after 36,500 BP in France and Spain (d'Errico, et al. 1998; Zilhão and d'Errico 1999). Some areas of Europe, such as the southern Iberian and Italian peninsulas may have served as refugia for Neandertal populations (Carbonell, et al. 2000; Kuhn and Biette 2000). Late-persisting Neandertals still manufacturing Mousterian technology existed in southern Iberia as late as 28,000-30,000 BP (Carbonell, et al. 2000; d'Errico and Sanchez Goñi 2003; d'Errico, et al. 1998; Finlayson, et al. 2006; Zilhão and d'Errico 1999; Zilhão and Pettitt 2006). In Italy, it was initially thought that there was about a five thousand year difference between the appearance of the Aurignacian industry in the north and south, with Neandertal populations persisting until 32,500 BP (Kuhn and Biette 2000). However, recent re-dating of layers at Grotta di Fumane now suggest that transitional Uluzzian and Proto Aurignacian industries were deposited before 39,900 BP (Higham, et al. 2009; Peresani, et al. 2008). In fact, Higham et al. (2009) postulate that many sites dating to between 35,000 and 45,000 BP 23 in Italy are actually much more contemporaneous than traditional AMS dating techniques indicate. One of the major issues in European Paleolithic archaeology is the extent to which anatomically modern humans and Neandertal populations interacted, and how these interactions, or lack thereof, are reflected in early Upper Paleolithic stone tool industries. Some scholars argue that in France, Neandertal and anatomically modern human populations overlapped between about 38,000 and 33/34,000 BP. During this time Neandertals presumably “borrowed” technological attributes from anatomically modern humans, culminating in certain early Upper Paleolithic stone tool industries such as the Châtelperronian (Mellars 1999). Other authors reject this claim (Clark 2002; d'Errico, et al. 1998; Zilhão and d'Errico 1999). Clark (2002) points out that the Aurignacian industry is primarily based on blade production, while the twenty plus regional Early Upper Paleolithic industries do not necessarily include blade production. D’Errico and colleagues (1998; Zilhão and d'Errico 1999) also disagree with Mellars (1999) about Neandertal “enculturation.” They present a re-evaluation of Grotte du Renne in order to establish the cultural originality of the Châtelperronian Neandertal occupants of the site. They point to a stratigraphically sound Châtelperronian layer beneath the Aurignacian occupation, and debitage from bone tool and ornament manufacture in the former (but see Higham, et al. 2010 for recent dates that indicate the bone tools and ornaments are actually intrusive). D’Errico and colleagues (1998; Zilhão and d'Errico 1999) discuss differences in the reduction sequences, blank selection and tool manufacture between the two industries. Most notably, unlike the Aurignacian, the Châtelperronian contains no large-scale bladelet production. In general, degree of interaction is difficult to assess based on archaeological evidence because of the potential for 24 stratigraphic mixing, and unclear implications from interstratifications of cultural layers even in the case that they are found intact (e.g., did one population replace another, or did they simply use a site after it was abandoned by a different group). The bioarchaeological record, including hominin fossils and paleoDNA, have the potential to determine if there was any gene flow between Neandertal and modern human populations. Mitochondrial DNA studies have consistently indicated that there was little or no admixture between modern human and Neandertal populations (Briggs, et al. 2009; Krings, et al. 1997; Ovchinnikov, et al. 2000; Serre, et al. 2004). More recent studies of long sequences of nuclear DNA, however, suggest that small amounts of Neandertal DNA (1-4%) are present in the genomes of modern Eurasian populations (Green, et al. 2010). These genes are not present in modern African groups, which indicates that gene flow occurred after the initial divergence between modern human populations in Africa and European Neandertals, or that the lineages went extinct in Africa. Human fossils associated with “transitional” or Early Upper Paleolithic industries may indicate the authors of these stone tool types. A problem with the skeletal data is that early Upper Paleolithic skeletons are extremely rare and it is difficult to construct a model of what contemporary anatomically modern human populations looked like, even in Africa. A small handful of European skeletons from well-dated contexts have traits consistent with both Neandertal populations and the few contemporary modern humans (Duarte, et al. 1999; Trinkhaus 2007; Trinkhaus, Milota, et al. 2003; Trinkhaus, Moldovan, et al. 2003; Wild 2005; Zilhão 2006). This largely supports the nuclear DNA data results, namely that there was a small amount of gene flow between the two populations. 25 The paucity of hominin skeletal remains also makes it difficult to determine the manufacturers of key Early Upper Paleolithic stone tool industries (e.g., the Châtelperronian, Uluzzian and Szeletian), since stone tools and human remains are rarely found in direct association with one another. A reanalysis of the hominin remains from Vindija Cave in Croatia identifies Neandertal remains with Upper Paleolithic lithic and bone technology, though there is potential stratigraphic mixing in the associated layers (Ahern, et al. 2004). Neandertal remains are associated with the late Micoquian in Germany (Schmitz, et al. 2002), the French Châtelperronian (Hublin, et al. 1996, but see Bar-Yosef and Bordes 2010) and possibly the Italian Uluzzian (Churchill and Smith 2000, but see Riel-Salvatore 2009). There is consensus among many archaeologists that the Aurignacian industry is connected with anatomically modern humans, though bioarchaeological remains from this period are very rare (Gambier 1989; Mellars 1989, 1991; Stringer, et al. 1984; Svoboda, et al. 2002). Early questions concerning Paleolithic subsistence strategies have focused on whether Neandertals were successful hunters or obligate scavengers. Though this debate is fairly well exhausted at this point, it has important implications if Neandertals were similar to Africanorigin modern human populations in terms of foraging ecology, social structure, and language abilities. Differences in these capacities may explain why modern humans eventually out-competed Neandertals. Binford (1984b; 1985; 1988) was a major proponent of middle and early late Pleistocene hominins as scavengers based on faunal assemblages from Klasies River Mouth Cave in South Africa and Grotte Vaufrey in France. He interpreted the French assemblage as originating from carnivore kills that were later scavenged by humans and brought to the site for processing (Binford 1988). The African remains were dominated by head and foot 26 elements which were thought to be indicative of scavenging, though some small game hunting was also noted (Binford 1984b). Binford’s assertions drew early criticism, and many subsequent studies have disagreed with his conclusions (Chase 1989; Grayson and Delpech 1994; Marean and Kim 1998; Speth and Tchernov 1998; Stiner 1994). In a survey of European Paleolithic sites, Chase (1989) finds few differences in hunting practices between Middle and Upper Paleolithic hominins; Neandertal assemblages indicate foresight in their exploitation of migratory ungulate herds and occasional specialization in one large game species. Grayson and Delpech (1994) present a taphonomic reanalysis of the fauna from Grotte Vaufrey and find that Binford’s (1988) conclusion that the Middle Paleolithic inhabitants of the site primarily scavenged large game is not supported. In a comprehensive taphonomic and economic analysis of faunal remains from Middle Paleolithic sites in Italy, Stiner (1994) examines Neandertal subsistence strategies as compared to other top-tier late Pleistocene carnivores. She determines that based on prey choice, body part representation and age structure analysis, Neandertals were successful hunters of prime-aged adult large game, as well as occasional, opportunistic scavengers, not unlike many modern foragers (e.g. O'Connell, et al. 1988). Similar conclusions, that Neandertals and even earlier Eurasian hominins targeted prime-aged adult animals are noted in Middle and Lower Paleolithic sites in Israel (Speth and Clark 2006; Speth and Tchernov 1998; Stiner 2005; Stiner, et al. 2009), the Republic of Georgia (Adler, et al. 2006), and northern Europe (Gaudzinski 1995; Gaudzinski and Roebroeks 2000). There is also evidence that Neandertal groups in Gibraltar and Italy occasionally exploited marine mammals in addition to marine mollusks (Stiner 1994; Stringer, et al. 2008). As these studies suggest, 27 Neandertals were clearly successful hunters of large game, and the hunting/scavenging debate has been laid to rest. Another hunting behavior traditionally considered as a hallmark of the Upper Paleolithic is specialized hunting a single large game species. This idea, proposed by Mellars (1973; 1989; 1991) was based on some instances of monospecific faunas such as reindeer assemblages in Paleolithic sites in western France. Such hunting strategies were thought to have a link to cognition as an indicator of the capacity for planning, forethought and logistics that accompanies predicting the movements of gregarious taxa (Mellars 1989), though natural diversity and seasonality also play a part. Since specialization was thought to be unique to the Upper Paleolithic, Neandertals and other Middle Paleolithic hominins were apparently lacking these high-level cognitive functions. Monospecific faunas are documented at many Upper Paleolithic sites, particularly those that concern ibex hunting (e.g. Gamble 1997; Gamble 1999; Phoca-Cosmetatou 2004a; Straus 1987), but it turns out that they are not unique to the Upper Paleolithic. "Specialization" as defined by the presence of monotypic assemblages occurs as early as the Middle Paleolithic (e.g. Costamagno, et al. 2006; Gaudzinski 1995, 2006) and represents a range of hunting strategies, including mass kills, and long-term periodic exploitation of herd and solitary animals. In the Middle and Upper Paleolithic of southern France, certain species were targeted during seasonal migrations (Costamagno 2004; Costamagno, et al. 2006; Surmely, et al. 2003). Examining small game exploitation patterns during the Paleolithic, Stiner and colleagues (Stiner 2001; Stiner and Munro 2002; Stiner, et al. 2000) show that the relative proportions of small to large game exploited during the Middle and Upper Paleolithic periods 28 in Italy and Israel did not uniformly change, though the kinds of small game exploited drastically shifted. In the Middle Paleolithic, small, slow-moving game species such as tortoises and limpets were exploited, while in the Early Upper Paleolithic, game birds and small mammals became proportionally more common. By the Epipaleolithic, lagomorph abundance was especially high (Stiner, et al. 2000). Additionally, Stiner et al. (2000) present data on size diminution in limpets from Italian sites and tortoises from sites in Israel, which indicates increased levels of pressure on these slow-reproducing species as early as 44,000 BP. A similar situation may have occurred in Turkey at Üçağızlı Caves I and II, where ungulates dominate the Middle and Upper Paleolithic layers, though they decline by the Epipaleolithic. Small, slow-moving taxa are important in the Middle Paleolithic, and again in the Ahmarian (Stiner 2009; Stiner, et al. 2002). By the Epipaleolithic, small fast-moving prey vastly increases in dietary importance. This pattern is also beginning to emerge in southern Greece, particularly by the Epipaleolithic and Mesolithic periods (Starkovich 2009; Starkovich and Stiner 2010; Stiner and Munro 2011). Small game was less important in southern France and areas to the north, as compared to the eastern Mediterranean (Costamagno 2004; Costamagno, et al. 2006; Surmely, et al. 2003). Clearly, zooarchaeology has much to tell us about the nature of hunting patterns and human adaptations across the late Pleistocene. Based on the studies reviewed above, drastic differences are not expected between the faunal assemblages generated by Neandertals and anatomically modern humans when it comes to planning or organizing hunting activities. Where there are differences, explanations related to environmentally driven changes in prey structure, human hunting pressures, or changes in site use or occupation intensity are evaluated, as opposed to the assumption that there was a cognitive difference between 29 Middle and Upper Paleolithic hominins. Faunal studies are more instructive in understanding other kinds of changes that occurred during the Paleolithic, such as responses to local environmental shifts or rising human population densities, either of which can stress prey populations. The Klissoura Cave 1 occupation spans the late Pleistocene and early Holocene, and includes Middle Paleolithic through Mesolithic cultural layers. Understanding variation within as well as between these occupations are central to this study. Theoretical models from evolutionary ecology are important for formulating predictions pertaining to changes in hunting and processing strategies through the occupation of the site. APPLICATIONS FROM EVOLUTIONARY ECOLOGY Human dietary change may reflect a variety of internal and external factors, from climatic shifts to population growth, changes in landscape use and site occupation intensity, or cultural preferences. Evolutionary ecology is a useful paradigm for framing questions about shifts in game use (see review in Bird and O'Connell 2006; Lupo 2007). Foraging models in particular are based on the idea that organisms maximize their fitness by making decisions that optimize their foraging efficiency and reproductive success within a given set of dynamic environmental circumstances (Pianka 2000). The paradigm is associated with a series of theories and behavioral models (sometimes called optimal foraging models) that explain foraging decisions: these include the diet breadth or prey choice model (Pianka 2000; Stephens and Krebs 1986), patch choice models (MacArthur and Pianka 1966), and central place foraging models (Orians and Pearson 1979; Schoener 1979). Other models are also employed by evolutionary ecologists (e.g., costly signaling theory and tolerated theft) but are 30 not discussed further as the three models mentioned previously are more readily applied to the Paleolithic archaeological record. Many criticisms have been leveled at evolutionary ecology and its application to human foragers in the ethnographic and archaeological records (e.g. Dwyer 1986; Joseph 2000; Martin 1983; Pyke 1984). The critiques offer some valuable and often nuanced points; the major misgivings seem to focus on 1) a discomfort with applying mathematical models to human behavior, because they are too simplistic or mechanistic and 2) practical matters such as defining the currency used to test optimality models. Some allies of evolutionary ecology also point out that due to the nature of the archaeological record, foraging models tend to focus on the faunal portion of the diet and ignore the vegetal component (Broughton and Grayson 1993; Grayson and Cannon 1999). Hill et al. (1987) also point out some of the difficulties of testing optimal foraging models. Specifically, an ethnoarchaeologist cannot coerce foragers to exploit low-ranked resources, so it is impossible to quantify their return rate. Because of this, it is unclear if certain items are excluded from the diet because of their low nutritional return, or for other, possibly cultural reasons. These criticisms are addressed extensively elsewhere (Broughton and O'Connell 1999; Smith 1991, 2000; Smith and Winterhalder 1985; Winterhalder 2002). The existence of potential problems with some aspects of optimal foraging models does not render them useless to anthropologists. The thought of humans conforming to simple mathematical models is may be unsettling, as such models may seem to deny us free will or cultural preference. More is learned, however, from deviations from expectations put forth by evolutionary ecology and optimal foraging models than from cases that fit predictions. The second criticism addresses the currency and goals used to test optimality 31 models. Typically, evolutionary ecologists focus on kilocalories as the unit of measurement of cost and gain in foraging pursuits, but other nutritional factors, such as complete protein or fat yields, may actually be more important. Further, goals may change depending on who is doing the foraging, and if they are responding to specific circumstances such as provisioning offspring. In general, these issues are addressed in the ethnographic literature (e.g. Bird and Bliege Bird 1997; Hill and Hawkes 1983; Hill, et al. 1987; Kelly 1995; Smith 1991), but they are difficult to test in the archaeological record because there are so many more unknowns. Kilocalories, and in most cases prey mass (see below) are used as proxy measures of nutritional value in most studies, and it is acknowledged that they are only estimations. As far as paying more attention to the meat component of archaeological diets, this is true in the same way that technological studies focus more on stone than organic tools in most instances. If organic remains, such as plant foods or perishable technologies preserve at a site, they should absolutely be included in considerations of diet breadth or technological adaptations. But the absence of such evidence does not mean that we cannot use models from evolutionary ecology to help understand exploitation of faunal resources. Prey Choice Model The diet breadth, or prey choice model, is particularly helpful for understanding dietary change among foragers (Pianka 2000; Stephens and Krebs 1986). The prey choice model assumes that resources are encountered randomly and sequentially on a landscape. Prey items are ranked according to their return rate, which is a measure of the caloric or nutritional value of a resource per unit of time invested in obtaining it, and includes both search and post-encounter handling costs (Pianka 2000; Stephens and Krebs 1986). Foragers 32 attempt to maximize their net caloric return by preferring prey with the highest net returns. Following the prey choice model, foragers are always expected to pursue high-ranked prey whenever they are encountered on the landscape. Further, the decision to pursue low-ranked species is irrespective of their abundance in the environment; rather foragers only add lowranked species to the diet as high-ranked resources become scarce (Emlen 1966; MacArthur and Pianka 1966; Pianka 2000; Stephens and Krebs 1986). Other variables can influence prey selection among humans, such as composition of the foraging party (Bird and Bliege Bird 2000; Hill, et al. 1987; Hurtado, et al. 1985; Jochim 1988; Lupo and Schmitt 2002, 2005), prestige (Hawkes 1990, 1991; Hawkes and Bliege Bird 2002; Smith 2004), or cultural taboos or preferences for certain resources. Since the archaeological record represents a palimpsest of foraging events, it is assumed for the purposes of this study that the remains at Klissoura Cave 1 are an average of the prey items exploited by foraging groups of variable composition (e.g., men and women) and that any temporal trend reflects changes in the foraging strategies of the entire group. Many ethnographic studies have employed prey choice models in order to examine human hunting decisions (Hames and Vickers 1982; Hawkes, et al. 1982; Hawkes and O'Connell 1981, 1985; Hill and Hawkes 1983; Hill, et al. 1987; Smith 1991; Winterhalder 1981). Hawkes and O’Connell (1981) provide an early application of the prey choice model to human foragers, in this case the gathering of tree and grass seeds by Alyawara women in Australia. By taking into account collection and processing costs, they determine that workloads for many foragers are actually considerably higher than was initially reported by Lee (1979) in his work with the !Kung San, who were presented as the model case for the “affluent forager.” Indeed, including processing costs in the calculation of foraging 33 efficiency drastically decreases the return rate of mongongo nuts, a key dietary resource of the !Kung (Hawkes and O'Connell 1981). Prey choice models are utilized in a series of papers that focus on the Aché, a group of semi-settled hunter-gatherers in eastern Paraguay (Hawkes, et al. 1982; Hill and Hawkes 1983; Hill, et al. 1987). In this case animals, particularly large-bodied prey, tend to be ranked higher than plant resources, especially plant foods that require processing (Hawkes, et al. 1982; Hill, et al. 1987). Further, Hawkes, Hill and colleagues find that the resources during the study period yielded a return higher than the average return rate for the total Aché diet, which agrees with optimal foraging predictions (Hawkes, et al. 1982; Hill and Hawkes 1983). The availability of shotguns does not cause certain low-ranked resources, such as small monkeys and birds, to be included in the diet; rather, guns increase the return rates of all prey species, so already low-ranked resources maintained their relative position (Hill and Hawkes 1983). Hill et al. (1987) take into account the desirability of specific nutrients, such as protein, in the diet as opposed to just calories, which is often a criticism of prey choice models (see above). Winterhalder (1981) applies the prey choice model to the Cree foragers of northern Ontario, who rely heavily on hunting for their subsistence. The Cree provide an interesting test of the prey choice model because the resources that they exploit, such as moose and caribou, waterfowl and fish, all require different technologies to procure. Therefore, the hunting technologies (shotgun, 0.22 rifle, fishing nets) carried on a foraging trip determine which resources are pursued, and the Cree often bring diverse weapons even if they set out to hunt one species (Winterhalder 1981). Further, the introduction of technologies in the mid1900s such as snowmobiles and onboard motors for canoes have increased the distances and 34 prey types that the Cree seek to procure. Winterhalder’s (1981) study of the Cree provides one of the best examples of how changing hunting and travel technologies impact forager decisions. The introduction of motorized travel and guns within a short period brought about a dramatic change, but prehistoric technological changes such as the appearance of composite tools and projectile weapons, the bow and arrow, nets and snares, and horseback riding would likewise have altered the efficiency with which humans exploited game. One of the key components of the prey choice model is determining which species are low- and high-ranked in an environment. Prey ranking typically is thought to be tied closely to body size, with large taxa assigned a higher rank (see Hawkes, et al. 1982; Kelly 1995; Simms 1987; Winterhalder 1981), as long as they are not unreasonably difficult to procure or time-consuming to process (Byers and Ugan 2005; Jones 2004; Smith 1991). Prey ranking among large game species can also be influenced by the age of animals in a population, because adult animals tend to be larger and have a higher proportion of body fat (Broughton 2002; Munro 2004; Speth and Clark 2006; Stiner 1994). As a general rule, small-bodied animals have typically been categorized as low-return because they provide a smaller nutritional package than large game species (Broughton 1994a, 1994b, 1999). However, this categorization of small-bodied prey as low-return has not traditionally taken into account differences in handling costs. Stiner and colleagues (Munro 2004; Stiner 2001; Stiner and Munro 2002; Stiner, et al. 1999; Stiner, et al. 2000) have shown that capture costs, and therefore return rates, vary a great deal among small game animals, based on fundamental differences in their flight behavior. This means that some small game could be much higher-ranked than others. For example small, relatively sessile species (e.g., limpets and tortoises) have much lower 35 handling costs than small, fast-moving species (e.g. hares, birds and fish), even though their caloric (and meat weight) values may be similar. Efficient exploitation of small, fast-moving species require a considerable technological investment, such as snares or nets, and/or cooperation of large groups of people (Cannon 2000; Jones 2006; Lupo and Schmitt 2002; Madsen and Kirkman 1988; Madsen and Schmitt 1998; Schmitt, et al. 2004; but see Ugan 2005b for a discussion of small game in different habitats). Because of these differences in search and handling costs, a small slow-moving animal has a higher return rate than a fastmoving animal of a similar size, and should be preferentially exploited by human hunters (Stiner 2001; Stiner, et al. 2000). This is borne out by ethnographic data (Kelly 1995; Kuhn and Stiner 2001: Table 5.1). Some applications of prey choice models, and other optimal foraging models discussed below, pertain to the question of resource depression, which occurs when prey capture rates decline in response to harvesting pressures (Charnov, et al. 1976). Many archaeological studies that employ prey choice models also address resource depression (e.g. Broughton 1994a; Broughton 1994b, 1997, 1999, 2002; Butler 2000, 2001; Butler and Campbell 2004; Cannon 2000; Grayson and Delpech 1998; Hill 2007; Jones 2004, 2006; Manne and Bicho 2009; Munro 2004; Nagaoka 2001, 2002a, 2002b; Speth 2004; Speth and Clark 2006; Stiner 2001, 2005, 2009; Stiner and Munro 2002, 2011; Stiner, et al. 1999; Stiner, et al. 2000; Ugan 2005a). This is because resources at a site are fairly easily ranked and change through time is assessed by the addition or removal of prey species in the diet. Independent data on prey size diminution may also be used to check these interpretations. Often, resource depression is linked to changes in local environments or climatic conditions, but may also result from increased human population densities or resource intensification. 36 The archaeological examples cited above are far too numerous to discuss in detail in this presentation. Suffice it to say, their application ranges from the Paleolithic through historic periods on nearly all continents. The works listed above that relate more directly to Paleolithic prey selection are discussed further in later chapters. Patch Choice Model The patch choice model is similar to the prey choice model, except it assumes that resources are encountered in discrete patches, as opposed to evenly distributed across the landscape. The model was developed by MacArthur and Pianka (1966) to determine which patches an organism will utilize and has subsequently been applied to human foragers. Patches are encountered by foragers randomly and sequentially on the landscape and are ranked based on their return rates (Kelly 1995). Upon entering a patch, the forager's return rate drops as resources are depleted. The patch choice model is often used in conjunction with the marginal value theorem (MVT), which assumes that once the return rate from the patch is less than the average rate of the surrounding environment (i.e., outside of the patch) a forager will leave the patch and travel to the next one (Charnov 1976; Charnov, et al. 1976). The MVT adds travel time between patches and handling time within a patch, which is not taken into account with patch choice alone. Application of the model requires more information than the simple prey choice model described above. Several ethnographic studies have used patch choice to understand recent human foraging patterns (Beckerman 1983; Hawkes, et al. 1982; O'Connell and Hawkes 1984; Smith 1991; Sosis 2002; Winterhalder 1981). The result from these studies are mixed, and seem to suggest that foragers do not exclusively target the most profitable patches on the 37 landscape. Rather, patch exploitation varies on a daily or seasonal basis in response to variations in patch profitability. A benefit of the patch choice model is that patches can be defined quite broadly, referring to the distribution of a single species (e.g., blackberry bushes), an ecosystem (e.g., marine, terrestrial, montane, deciduous forest), foraging strategy (e.g., fishing or hunting large game) (Lupo 2007). Many archaeological applications of the patch choice model have drawn on ethnographic studies that look at prey types that inhabit a specific ecosystem, for example grassland or marine environments (Broughton 1999; Cannon and Meltzer 2008; Jones 2007, 2009; Nagaoka 2002a, 2002b). These studies find changes in resource exploitation as the result of increased “patchiness” of environments caused by climatic change (Jones 2007, 2009), as well as shifts to more technologically difficult to exploit environments, such as offshore areas (Nagaoka 2002a, 2002b), or the use of distant patches as a result of resource stress (Broughton 1999). Another archaeological application of the patch choice model seeks to understand processing intensity of individual animal carcasses. The model is used in much the same way as examining patches on a landscape, but the scale of analysis is different. Carcasses are analyzed as individual patches and decisions are made about how intensively to process each patch (carcass) before moving onto the next patch (kill) (Burger, et al. 2005; Nagaoka 2005). Animal tissues vary in their nutritional return based both on their inherent caloric value and the effort needed for extraction. In general, meat and organs have the highest returns, followed by bone marrow, and finally bone grease (see values in Binford 1978; Kooyman 1990; Lupo 2006; Madrigal and Holt 2002 for a few examples). In exploiting the patch (carcass), foragers initially should utilize the highest return parts, meat and organs, followed 38 by bone marrow, and only move on to processing labor-intensive bone grease (see Chapter 6 for a description of the process) if the processing costs are lower than the combined travel and handling costs of seeking another prey item. Presumably, this occurs during times of resource stress brought on by seasonal factors or long-term resource depression. Burger et al. (2005) apply the MVT to predict foraging intensity across animal carcasses. Using values from Binford (1978) and Madrigal and Holt (2002), the authors construct food gain curves to predict the point at which certain low-return ungulate elements should be processed for bone marrow. Burger et al. (2005:1151) note the nutritional differences between lean protein and fatty bone marrow, which is easier to metabolize than protein, and suggest that foragers will always seek some amount of marrow when processing an animal. Foragers are expected to process marrow more intensively during times of resource stress and when meat is most lean, though this can also vary depending on the seasonal condition of an animal. The authors propose that a “stop element approach” can be taken with archaeological assemblages, where different rates of marrow processing for lowutility elements are indicative changing butchery strategies, possibly in response to resource stress (Burger, et al. 2005). The authors also note that increased rates of fragmentation may indicate resource stress, either associated with marrow processing or bone grease rendering. Obviously post-depositional factors that affect fragmentation would also have to be taken into account in archaeological cases. Nagaoka (2005) uses the patch choice model to examine increases in the processing intensity of moa at the Shag River Mouth site in New Zealand, which dates to 1250-1450 CE. The author uses the MVT to predict that moa should be processed more intensively when resources were stressed in later time periods, which is well-established at the site (Nagaoka 39 2001, 2002a, 2002b). Following Wolverton (2002), she looks at the percent of unbroken moa elements and the amount of bone marrow each element yields in order to determine if elements were processed more intensively through time. Nagaoka (2005) determines that, whereas large marrow-rich moa long bones were consistently broken throughout the sequence, phalanges with small amounts of marrow are broken more frequently in later layers. This trend indicates a rise in the intensity of carcass processing with time. The author also looks for evidence of bone grease rendering by looking at changes in NISP:MNE ratios (following Wolverton 2002). She finds no evidence for grease extraction, except on moa tibiotarsi (Nagaoka 2005). Many other archaeological studies examine carcass processing as a means of intensifying animal resources, either in response to human or environmental pressures (e.g., Adouze 1987; Adouze and Enloe 1991; Binford 1993; Brink 1997; Broughton 1999; David and Enloe 1993; Egeland and Byerly 2005; Hill 2007; Manne and Bicho 2009; Munro and Bar-Oz 2005; Potter 1995; Stiner 2003b; Weniger 1987). Though these authors do not necessarily present their results in terms of patch choice models, they follow the basic tenets of patch choice and MVT, as they address changes in the frequencies of bone marrow or grease processing through time in response to resource depletion. Central Place Foraging Models Central place foraging (CPF) models were developed by ecologists to understand transport decisions by species that bring food resources to a home base, or central place, in order to eat without interference, provision offspring, or store resources (Orians and Pearson 1979:156; Schoener 1979). A central goal of the models is to understand how resource 40 patches are chosen, how prey is selected, and how load size is determined (Orians and Pearson 1979:156). The models predict that as distance from a central place increases, the size of the load also increases in order to make travel costs worthwhile (Orians and Pearson 1979; Schoener 1979:913-914). This prediction seems to be valid for many non-human predators (see Stephens and Krebs 1986:194-197). CPF models were readily adopted by anthropologists, as movement of resources to a home base is a typical human behavior. Anthropological applications of CPF models are wide-ranging, from the transport of raw materials for stone tool production (Beck, et al. 2002), to the movement of plant resources (Barlow and Metcalfe 1996; Bettinger, et al. 1997; Jones and Madsen 1989), to the transport of aquatic (Bird 1997; Bird and Bliege Bird 1997) and terrestrial animal species (Bunn, et al. 1988; Egeland and Byerly 2005; Lupo 2006; O'Connell, et al. 1988, 1990; O'Connell and Marshall 1989; Zeanah 1999, 2004). Each of these studies attempts to assess the distance at which it becomes worthwhile for human foragers to field process a given resource, and the amount of processing that yields optimal economic gain. The formal model introduced by Metcalf and Barlow (1992:343) explores the relationship between the round-trip travel costs of visiting a resource patch, the amount of time spent procuring and field processing the resource, and the utility of the resource load. The model predicts an inverse relationship between an increase in load utility from field processing and the distance traveled before field processing becomes efficient (Metcalfe and Barlow 1992:347). Therefore, field processing that results in a large increase in load utility is more beneficial at shorter travel distances, and field processing resulting in small increases in load utility only become worthwhile after longer distances are traveled. Metcalfe and Barlow’s (1992) model can be most directly applied to simple, “structured” resources such as 41 seeds or mollusks that must be processed in a specific order (i.e., the removal of the outer shell before the inner shell when processing some kinds of nuts). However, they recognize challenges associated with utilizing more complex, “unstructured” resources such as animal carcasses that can be processed in an almost infinite number of ways (e.g., after skinning, the hind limbs, forelimbs, head parts or axial skeleton of an animal could be butchered next). Cannon (2003) expands on Metcalfe and Barlow’s (1992) model by presenting a “central place forager prey choice model” that takes into account differences of prey size, transport distances, and processing costs. The author argues that though intensive field processing may increase load utility, return rates are always higher for resources taken close to home, regardless of processing costs (i.e., increased load efficiency does not fully compensate for travel costs) (Cannon 2003:9). He also points out that while in general largebodied prey provides a higher return than small-bodied prey, there comes a point if large game is locally unavailable that the search and transport time for such resources becomes so costly that local small prey actually provides a higher return rate (Cannon 2003:12). Applications of CPF models or those using similar rational have been fruitful in explaining butchery and transport decisions that surround the processing and movement of vertebrate remains from kill sites to a home base (Binford 1978; Bunn, et al. 1988; Egeland and Byerly 2005; Lupo 2006; O'Connell, et al. 1988, 1990; O'Connell and Marshall 1989). Of the three models, application of this one requires the most information a priori. Binford’s (1978) groundbreaking ethnoarchaeological work with the Alaskan Nunamiut established a baseline set of data that included utility indices for meat, marrow, and grease from different skeletal parts of dall sheep (Ovis dalli) and caribou (Rangifer tarandus). The combined values of the indices from the three kinds of tissue (meat, marrow, grease) constitute the 42 general utility index (GUI), which helps to predict the decisions the Nunamiut make concerning the butchery, transport and storage of carcasses (Binford 1978:72). The GUI is further derived into the modified general utility index (MGUI) that takes into account lowutility elements that may be transported along with high-utility portions (Binford 1978:74). Based on the MGUI and frequency of various skeletal parts in an assemblage, Binford (1978:81) introduced a set of utility curves that predict different body part transport strategies. His study, which is discussed further in Chapter 6, was an important first step in quantifying carcass transport decisions by human hunters. This study has influenced almost all subsequent ethnographic and archaeological investigations that employ CPF models. Many ethnographic studies of body part transport focus on the Hadza of Tanzania (Bunn, et al. 1988; Lupo 2006; O'Connell, et al. 1988, 1990). Bunn et al. (1988) provide an account of transport patterns, but do not track the relative utility of different body parts. They find that appendicular elements tend to be transported more frequently to base camps than do axial elements (Bunn, et al. 1988:450), and that much fluidity exists in defining locations as “butchery sites” or “home base sites.” This is because Hadza hunters often process and deposit bones at “snack sites” between the kill location and home base, frequently re-using both kinds of sites for various activities (i.e., using old “home base sites” as “snack sites”) (Bunn, et al. 1988:437). Thus, they caution archaeologists attempting to interpret transport patterns that sites may have multiple usages over time. O’Connell et al. (1988; 1990) explain Hadza transport strategies in terms of carcass size and which elements should be selected for transport. Using a scalogram (Gutman matrix) analysis to construct a ranking system for different body parts, they determine that though there is some variation among species, vertebrae, scapulae, pelvises and upper limb bones tend to be preferentially transported from 43 kill sites to base camps (O'Connell, et al. 1988:138). In order to determine the number of bones transported for each carcass, transport costs were calculated based on the average weight of a carcass per human carrier, multiplied by the distance in minutes from the butchering site to the home base (O'Connell, et al. 1988:132). In general, O’Connell et al. (1990:311) find an inverse relationship between the number of elements moved and transport costs. Additionally, they find that variation in transport exists between species of similar sizes, particularly if the animals have different body forms (O'Connell, et al. 1990:310). The significance of the studies done by O’Connell et al. (1988; 1990) are in dispelling ideas that appendicular elements are typically transported preferentially over axial elements; rather, there is variation depending on the prey species under consideration and prevailing technology (e.g., the Hadza have cooking pots suitable for boiling). Further, they caution that lumping different species with similar body sizes, while a common archaeological practice due to identification constraints, can obscure species-specific transport decisions. Lupo (2006) employs CPF models to evaluate how processing and transport costs of certain anatomical portions of impala and zebra affect decisions made by Hadza foragers. She measures utility using the food utility index, which is the total body part weight minus the dried bone weight (after Metcalfe and Jones 1988). Lupo distinguishes between FUI(t), the total food utility index equivalent to the FUI of Metcalfe and Jones (1988) and FUI(r), the remnant animal products, calculated by subtracting the weight of the meat removed in the initial filleting and the weight of marrow removed from FUI(t) (Lupo 2006:36). Lupo finds that the kind of element (e.g., axial vs. appendicular) is less predictive of transport decisions than FUI(r), and there are differences between impala and zebras. Hadza butchering strategies for impala are fairly straightforward: foragers discard elements that can quickly be 44 depleted of meat and marrow (Lupo 2006:46). However, zebras tend to be processed only for meat in the field, with the separated meat and bone subsequently transported to the residential site, the latter for marrow and grease processing (Lupo 2006:47). Lupo (2006:48) postulates that this is because of the specific structure of equid bone, which makes it difficult to process in the field without heat. Her study refines observations made by O’Connell et al. (1988; 1990) and Bunn et al. (1988) and provides a quantitative argument for understanding different transport decisions made by the Hadza. These ethnographic studies are increasingly sophisticated in their use of CPF models or similar rationale. They provide insight into interpreting carcass transport decisions in the archaeological record, while also cautioning against using modern examples as direct analogs. The greater range of unknowns in archaeological cases must also be acknowledged. The ethnoarchaeological studies also illustrate the range of carcass transport decisions made by modern foragers depending on the prey species, number of participants in the hunt or transport effort, distance between the kill and habitation sites, available technology, and local climatic conditions. These ethnographic examples are instructive for archaeologists seeking to evaluate different factors past foragers may have taken into account when making decisions about carcass transport. Following the ethnographic work discussed above, many recent archaeological studies have adopted CPF models to evaluate body part transport decisions of past foragers (Broughton 1999; Cannon 2003; Faith 2007; Marín Arroyo 2009; Nagaoka 2002b; 2005, see also Speth 1991; Speth and Scott 1989 for discussions of the long-distance transport of large game to horticulturalist sites). Broughton’s (1999) application of CPF models to the Emeryville Shellmound faunas from the San Francisco Bay (2600-700 BP) set the stage for 45 many subsequent studies of large game body part transport. Average food utility index (FUI) values for artiodactyls are established for each stratigraphic layer in order to determine the economic utility of elements transported to the site. Relative skeletal abundance (RSA) is calculated by dividing the number of identified specimens for each anatomical portion by the number of times the part occurs in the body (Broughton 1999:58). Mean FUI values are calculated by multiplying RSA and FUI values for anatomical portions (from Metcalfe and Jones 1988) and dividing by the total RSA for the archaeological unit. Broughton (1999) finds an initial decline in mean FUI of artiodactyl remains across the lower layers of Emeryville Shellmound, then a subsequent increase in FUI in the later layers. The author interprets this as lowered foraging returns following the depression of local artiodactyl populations, and a decrease in body part selectivity as local resources were exploited more intensively (Broughton 1999:59). Then, in the upper layers, fewer low-utility elements were transported to the site, indicating that foragers traveled further from the site to procure large game. Nagaoka (2002b; 2005) applies Broughton’s (1999) methodology to forager decisions surrounding the movement of moa and seal anatomical parts to the Shag River Mouth site. She finds an overall increase in high-utility moa body parts through the sequence, indicating moa hunting further from the site. A closer look at the data that separates out the highest and lowest-utility elements indicates that there was a ubiquitous increase in high-utility elements, and an initial increase followed by a decrease in the lowest-utility parts (Nagaoka 2002b, 2005). The author concludes that the lowest-utility elements were initially “riders” transported to the site still connected to high utility elements, but as foragers began to process carcasses more efficiently these elements were removed before transport in later periods. In 46 examining the transport of seal remains, Nagaoka (2002b) determines that in earlier periods higher-utility elements are found at the site, and in later periods both low and high-utility elements appear. She proposes the possible explanation that seals were hunted at distant rookeries in later periods, but transport costs were different because of the use of canoes. Faith (2007) expands on Broughton’s method by using it and the Shannon evenness index (following Faith and Gordon 2007) to evaluate the representation of reindeer elements at Grotte XVI, France during the Middle Paleolithic through Magdalenian. The author builds on a study by Grayson and Delpech (2003; Grayson, et al. 2001), who noted an increase in the frequency of reindeer in the later layers of Grotte XVI as a result of cooling summer temperatures. Rather than investigating resource intensification, Faith (2007) tests the hypothesis that foragers became less selective about reindeer body part transport in response to increased abundance near Grotte XVI. He finds a decrease through time in the mean food utility index of elements in the assemblage, in addition to an increase in skeletal evenness of elements transported to the site. Faith’s (2007) analyses suggest that there was less selective transport of reindeer parts to the site as prey availability increased in the region during the late Upper Paleolithic (see Grayson and Delpech 2003; Grayson, et al. 2001). Following the basic premise of Broughton (1999), Cannon (2003) examines changes in the anatomical representation of ungulate remains from multiple sites in the Mimbres Valley, New Mexico between about 400 and 1450 CE. The author calculates the mean FUI for artiodactyl remains from different cultural phases similarly to Broughton (1999), then compares the means of each phase though an analysis of variance (Cannon 2003). The author finds that average FUI values are highest in the earliest and latest phases (Early Pithouse and Classic Mimbres/Terminal Classic Mimbres) and are lower in the intervening periods, 47 indicating that low utility elements were selectively excluded at the beginning and end of the sequence. It must be noted, however, that the sample sizes for these periods are extremely small (n<16), a surprising contrast to the rigorous statistical tests applied. Cannon (2003) attributes the initial decline in FUI values to increased hunting party size and the later increase to foraging trips further from the analyzed sites. Marín Arroyo (2009) examines central place foraging to understand site catchment in the Cantabrian region of Spain during the Magdalenian (14,000-10,000 BP), with a focus on red deer and ibex hunting. Following Cannon’s (2003) model, the author introduces time and distance components to the model, determining that smaller-bodied ibex are an energetically efficient hunting option if their habitat is 45% closer to a base camp than red deer habitats (Marín Arroyo 2009:31). She also calculates that red deer are completely butchered on a 7.5 hour hunting trip, resulting in the transport of high-utility elements to the site. Finally, Marín Arroyo (2009:31) determines that it is rarely efficient to process red deer bone marrow in the field, unless it was consumed as an on-site snack. The author applies these assumptions and topographic data to nineteen Magdalenian sites in Cantabria and finds a significant relationship between topography and patch distance, and frequencies of highland- versus lowland-dwelling species in the zooarchaeological assemblages (Marín Arroyo 2009). The examples discussed above illustrate the potential utility of applying central place foraging models to archaeological sites from diverse geographical areas and time periods. The underlying assumptions allow for a quantitative evaluation of change in body part transport strategies through an archaeological sequence. Obviously factors beyond travel time can affect transport strategies (e.g., size of the foraging party, transport technologies 48 such as boats or vehicles), but establishing a trend or lack thereof is a first stage in evaluating forager transport strategies. Predictions of Hominin-Prey Relationships Central to the use of models from evolutionary ecology that are outlined above is the assumption that foragers seek to optimize their behavior if it will make them more reproductively fit. Therefore, it follows that human foragers attempt to maximize the return they get from their prey by exploiting the highest-ranked species or age groups available, maximizing the gains from transport of resources depending on their distance from a site, and through carcass processing. Substantial shifts in any of these areas of predatory behavior is of interest, as it may be indicative of environmental pressures, resource depression or other causes of socioeconomic change. This dissertation seeks to apply the diet breadth or prey choice model, and the patch choice and central place foraging models to the long sequence of zooarchaeological remains from Klissoura Cave 1 in southern Greece. The prey choice model is applied to changing proportions of low- and high-ranked prey (Chapter 5). The patch choice and central place foraging models are employed for understanding changes in butchery patterns (i.e., marrow processing) and body part transport decisions, respectively (Chapter 6). Data on environmental change are used to qualify the interpretations of changes in prey use over time. In addition to environmental impacts on the faunas available to foragers occupying Klissoura Cave 1, evidence from elsewhere in the Mediterranean indicates that human population densities were increasing through the Middle and Upper Paleolithic (Lahr and Foley 2003), and into the Mesolithic. Such population growth would likely place increasing 49 pressure on the animal populations exploited by human hunters. In this context, it is predicted that resource depression, if present, should be apparent in the dietary spectrum at Klissoura Cave 1 over time. Expanding diets may be reflected by 1) an increase in low-ranked prey types, including difficult-to-catch small animals such as hares and birds, as well as more young ungulates, 2) an increase in high-utility ungulate body parts as hunters forage further from the site, and 3) more intensive processing of carcasses, including opening long bones for marrow and possibly even bone grease rendering. The following chapters explore these lines of evidence to determine the extent to which the depression of food resources in southern Greece resulted in changes in animal exploitation, and to what extent environmental factors played a role in dietary change. THE SAMPLE The research presented in this dissertation is based on the vertebrate faunal assemblages from Klissoura Cave 1, a stratified rock shelter near Prosymni in Peloponnese, Greece. The site contains a large Middle Paleolithic component that spans Marine Oxygen Isotope Stage (MIS) 5a at about 80,000 BP through the late Middle Paleolithic (ca. 53,600 cal BP), Early Upper Paleolithic (Uluzzian), Aurignacian, Epigravettian and Mesolithic, the latter two periods being less well represented (Koumouzelis, Ginter, et al. 2001; Kuhn, et al. 2010; Stiner, et al. 2010). The long, relatively intact cultural sequence of Klissoura Cave 1 is particularly important for diachronic comparisons of changes in hunting strategies and diet in southern Greece during the late Pleistocene. The sample includes nearly 25,000 identified remains from a long series of stratigraphically intact layers (mixed deposits were excluded from the final analysis, see Chapter 2). To date, the analysis presents one of the largest, most 50 complete Paleolithic faunal sequences from Greece. Though one site cannot be taken as fully representative of hominin behavior in an area, the data from Klissoura Cave 1 provide a critical starting point for understanding changes in subsistence strategies in southern Greece during the late Pleistocene. The robust sample provides evidence for dietary trends that relate to local climatic fluctuations, as well as larger demographic changes that occurred as hominin populations grew throughout Europe during the Paleolithic and later periods. The Klissoura Cave 1 data are compared to published faunal data sets from contemporary faunal series in Greece and the Mediterranean Basin to more fully understand late Pleistocene subsistence strategies and dietary change. DISSERTATION STRUCTURE This dissertation is organized into eight chapters, beginning with an introduction to the problem and the theoretical paradigm through which the data are interpreted. Chapter 2 provides an overview of Paleolithic research in Greece and a discussion of the excavations (history and procedures) at Klissoura Cave 1. Background information about the environmental conditions for the duration of the occupation of Klissoura Cave 1, with a look at the Mediterranean region in general, and the local conditions in Greece specifically is provided in Chapter 3. Chapter 4 describes the analytical methods employed in this study and explores the non-human taphonomic processes that may have affected the character of the faunas from Klissoura Cave 1. A discussion of non-human taphonomic factors, including the potential of in situ density-mediated attrition, is critical to understanding the formation of a faunal assemblage and distinguishing trends in human faunal exploitation from natural processes. 51 Chapters 5 and 6 focus on human hunting patterns and bone modification at Klissoura Cave 1. Chapter 5 discusses prey choice and species abundance by looking at the frequency of different taxa exploited at the site. Prey species are considered individually, as well as by the categories of body size and flight pattern discussed above. Using diversity indices, changes in species and prey group exploitation patterns are evaluated for trends. Alternative indicators of changes in faunal exploitation patterns are presented in Chapter 6, focusing on body part transport decisions, prey butchery patterns and age structure analysis. Chapters 5 and 6 draw heavily on expectations predicted by evolutionary ecology, using the prey choice, central place foraging, and patch choice models to evaluate intensification efforts that occurred at Klissoura Cave 1 during the late Pleistocene. Chapter 7 examines inter-site variation of bone burning and fragmentation at Klissoura Cave 1 to the extent that available excavation data permit. A series of intact hearth structures, including dozens of clay-lined heart features and a possible habitation structure in the Aurignacian layers, suggest important spatial variation in site use; this analysis will focus on the richest, most intact layers. Finally, Chapter 8 offers a summary and synthesis of the faunal results from Klissoura Cave 1, and a discussion of the overarching trends in faunal exploitation patterns from the Middle Paleolithic through Mesolithic. The results from Klissoura Cave 1 are then examined in the context of the archaeological record from Eastern Europe and the greater Mediterranean region, with a special focus on resource intensification in light of environmental variability and possible trends in human demography. 52 CHAPTER 2: KLISSOURA CAVE 1 AND THE GREEK PALEOLITHIC INTRODUCTION Paleolithic research in Greece began rather late, if compared to parts of Western Europe and Southwest Asia. This is often attributed to the historical interest in Classical archaeology of the region, and the divide between classical and prehistoric studies in universities (Runnels 1995b, 2001). Ultimately, this lack of attention early on could be argued to be a positive development as many Greek Paleolithic sites have had the benefit of being excavated using modern methods and technologies. Only a few small-scale excavations or surveys were conducted prior to 1950, and unfortunately artifacts associated with these early investigations were lost during the war years (Runnels 1995b, 2001). Starting in the late 1950s, interest increased in the Greek Paleolithic. Surface finds in the Argolid (NW Peloponnese) led to a series of unsuccessful cave excavations in the region (e.g., Bialor and Jameson 1962). At the same time, Paleolithic surveys and excavations were underway by a team in Thessaly (central Greece) (Milojčić, et al. 1965). The Middle Pleistocene Petralona skull was subsequently discovered in the mid 1960s (Hennig, et al. 1981; Ikeya 1980; Kurtén and Poulianos 1977). The first comprehensive survey of northern Greece was conducted in the early 1960s, which led to excavations of the early Middle Paleolithic open-air site of Kokkinopilos, and the late Pleistocene rock shelters of Asprochaliko and Kastritsa (Dakaris, et al.; Higgs and Vita-Finzi 1966; Higgs, et al. 1967). Shortly thereafter, as part of a larger Neolithic and Bronze Age project on the Ionian islands, Grava, an Upper Paleolithic rock shelter on Corfu, was excavated (Sordinas 1969). A salvage 53 operation was conducted at Kephalari Cave, a Middle and Upper Paleolithic site in the Argolid, in the early 1970s. The first large-scale systematic excavation at a Greek Paleolithic site also began at this time at Franchthi Cave, which spans the Upper Paleolithic through Neolithic. The site was excavated from 1967 until 1979 and is located in the Argolid (Jacobsen 1981; Jacobsen and Farrand 1987). These projects ushered in the modern era of Greek Paleolithic archaeology, which has witnessed a series of surveys and excavations primarily in Epirus (northwestern Greece) and Peloponnese. Consequently, most of what we currently know about the Greek Middle and Upper Paleolithic comes from northern Greece and the western half of Peloponnese (though some "Lower Paleolithic," i.e. early Middle Paleolithic, tools were reportedly discovered on Crete, see Strasser, et al. 2010). Some early work on the Greek Mesolithic was conducted by Adalbert Markovits in the 1920s, but his collections were thought to be missing until recently (Galanidou 2003). The excavations at Franchthi Cave exposed an extensive Mesolithic occupation, including multiple human burials and cremations (see below), which brought the Greek Mesolithic to the attention of many prehistorians. Subsequently, a small handful of sites with Mesolithic components were excavated and collections from old excavations are currently undergoing analysis. A perennial problem with Paleolithic archaeology in Greece is that the results of many excavations are only partially, or never, published. The recent interest in Greek prehistoric archaeology gained additional traction in the late 1980s, and has resulted in a dramatic influx of information available to a much wider academic audience. New excavations and analyses of previously excavated materials have yielded a growing body of 54 faunal and lithic data, in addition to the publication of some organic materials and human skeletal remains. In this chapter, I review the results of archaeological studies of the Paleolithic and Mesolithic of Greece, focusing especially on recent work and mainly on cave sites. I will emphasize the late Pleistocene and early Holocene, beginning at about MIS 5a (80,000 BP), which marks the earliest occupations at Klissoura Cave 1. Open-air sites are excluded from this presentation, unless they preserve faunal material or have radiometric dates (see Ammerman, et al. 1999; Darlas 1999; Dousougli 1999; Papagianni 2000; Runnels 2001; Runnels, et al. 2003; Runnels and van Andel 1993 for a discussion of some open-air Greek Paleolithic sites). This chapter is organized geographically, starting with the sites in the north, then moving to the Peloponnesian sites, and concluding with an in-depth description of Klissoura Cave 1, the focus of this study. Sites included in the text appear in Figure 2.1; see Figure 2.2 for ranges of dates for sites discussed below. All radiocarbon dates are calibrated. NORTHERN GREECE Asprochaliko Asprochaliko, a Middle to Upper Paleolithic rock shelter in the Louros Valley, was part of the early excavation efforts in northern Greece during the 1960s (Higgs and VitaFinzi 1966). Excavations occurred inside the rock shelter and on a slope outside of the shelter, though the stratigraphy of the latter is questionable (Bailey, et al. 1983). The sequence contains an Upper Paleolithic and two Middle Paleolithic components, and spans about 100,000 to 26,000 BP based on thermoluminescence (TL) and radiocarbon dating (Bailey, et al. 1983; Bailey, et al. 1992; Huxtable, et al. 1992). 55 Figure 2.1 Map of Greece with Paleolithic and Mesolithic sites mentioned in the text. TL dates place the basal Mousterian (Middle Paleolithic) at Asprochaliko between about 102,000 and 96,000 BP (Huxtable, et al. 1992). Tools in this older layer are “typical” Mousterian and the assemblage is dominated by side scrapers and notches (Gowlett and Carter 1997:445). The later Mousterian contains the same raw materials found in the older 56 layer but is technologically distinct (Papaconstantinou and Vassilopoulou 1997:461; Bailey et al. 1983 calls it the "micro-mousterian"). Pseudo-Levallois points are found in the later Mousterian layers. The technology used to manufacture such points is a combination of the Levallois technique and a discoid method; the authors subsequently call it the Asprochaliko method as it is not found elsewhere (Papaconstantinou and Vassilopoulou 1997:463). “Asprochaliko flakes” manufactured by this method are considerably smaller than flakes produced using the typical Levallois technique. The authors argue that it is difficult to place the Greek Mousterian within a larger Eastern European chronology because of the variability found in the region (Papaconstantinou and Vassilopoulou 1997:477). The Upper Paleolithic lithics from Asprochaliko are presented by Bailey et al. (1983) and Adam (1989; 1997). The Upper Paleolithic minimally dates to 26,350 +/- 1,900 BP, and is defined as possibly Gravettian (Bailey, et al. 1983). The lithic sample is fairly small and lacks evidence of production stages on site; no blades or blade cores are present. Bladelet tools and endscrapers dominate the assemblage (Adam 1989, 1997). The authors note that there is no Early Upper Paleolithic industry comparable to the Aurignacian or Uluzzian (Bailey, et al. 1983:33), and a sterile layer of sediments indicates a hiatus between the Middle and Upper Paleolithic components. A few worked bone specimens were recorded (Bailey, et al. 1983). The faunal assemblage contains 1,414 identifiable specimens, but only about forty percent of these came from areas of known provenience (Bailey, et al. 1983:34). Despite this qualification, the authors assert that the faunal composition changes through the sequence, though it is always dominated by ungulates. The basal Mousterian contains abundant red deer, and rhinoceros was recorded in the earliest levels (Bailey, et al. 1983:34). In the 57 Figure 2.2 Calibrated radiocarbon dates and ranges of dates for sites mentioned in the text. Cultural periods are listed above their respective dates. Where radiocarbon dates are unavailable, placement of cultural periods are estimated based on dated industries in Greece. 58 subsequent Mousterian, there is a fairly even representation of red deer and fallow deer. By the Upper Paleolithic, red deer is still present, but ibex replaces fallow deer (Bailey, et al. 1983:34). Theopetra Cave Theopetra Cave in Thessaly contains a long prehistoric sequence, spanning the Middle Paleolithic through the end of the Neolithic. The Middle Paleolithic layers were initially thought to date to about 46,050 to 34,050 BP based on radiocarbon dating (Facorellis, et al. 2001; Kyparissi-Apostolika 1999). Recent TL dates now place the Middle Paleolithic layers at 124,000 +/- 16,000 and 129,000 +/- 13,000 (Valladas, et al. 2007:306). This means that the earlier Middle Paleolithic deposits at Theopetra Cave are among the oldest in Greece. Because they are much older than the Klissoura Cave 1 assemblages, they are discussed only briefly here. Three Middle Paleolithic lithic phases were recorded at Theopetra Cave (Panagopoulou 2000). The oldest (phase 1) is defined as a Quina industry, followed by phase 2, which has Levallois blank production and a Mousterian toolkit. Lithic phase 3 was initially defined as a terminal Middle Paleolithic or a Middle to Upper Paleolithic transitional industry (Panagopoulou 2000), but recent dating indicates that the these materials actually come from mixed deposits (Valladas, et al. 2007:307). A stratigraphically higher TL date yielded an age of 59,000 +/- 6,000 BP, though little cultural material was noted in this layer (Valladas, et al. 2007). The stratigraphy of Theopetra Cave is difficult to understand as it includes a series of hiatuses and cryogenic events (Karkanas 2001). The recently published TL dates provided by Valladas et al. (2007) will undoubtedly lead to a reinterpretation of the stratigraphy. Many 59 long hiatuses are apparent at the site, most of which relate to the early Middle Paleolithic layers. Early Upper Paleolithic deposits are mixed with Middle Paleolithic layers (Adam 1999a). There is a hiatus in the sequence during the last glacial maximum, and the subsequent Upper Paleolithic/Epipaleolithic deposits range between 17,850 +/- 700 and 13,000 +/- 200 BP (Kyparissi-Apostolika 1999). The Mesolithic layers are largely intact and date to between 11,400 +/- 1,250 and 8,815 +/- 215 BP (Kyparissi-Apostolika 1999). Lithic raw materials during the Upper Paleolithic and Mesolithic layers are primarily radiolarite, though some non-local flints were also noted in the Upper Paleolithic (Adam 1999a). The Upper Paleolithic industries contain a low ratio of tools to debitage, and few cores (Adam 1999a). The most common tool types are unilaterally backed bladelets, retouched bladelets, and only a few blades, endscrapers and truncations are present (Adam 1999a:267). The author suggests based on the amount of debitage that many of the tools were made at the site. In the Mesolithic layers there is no evidence for the production of laminar debitage (Adam 1999a:267). Tools include truncations, notches and flakes with alternate retouch. In general the results on the Upper Paleolithic and Mesolithic industries remain preliminary. To date, slightly fewer than 4,500 faunal remains from the Paleolithic and Mesolithic layers at Theopetra Cave have been analyzed and published, of which 436 are identifiable (Newton 1999). The author notes high incidences of post-depositional breakage of the materials, with only a few green (fresh) bone breaks (Newton 1999:116). Human modifications on the bones include some burning damage and a few cut marks, and rare carnivore damage is also present. A range of ungulate species is recorded, including ibex, equids, red deer, fallow deer, and roe deer. Ibex and equids are only found in the Paleolithic 60 layers (Newton 1999). Carnivores comprise about a fifth of the assemblage and include large and small forms such as bear, hyena, wild cat, wolf, fox, badger, and marten. Small mammals such as hare and beaver are found at Theopetra Cave, as are avian taxa including corvids, rock dove, partridge, a few aquatic birds, and some remains of small owls and diurnal raptors (Newton 1999). A generalized “reptile” category occurs mainly in the oldest Paleolithic layers, but it is unclear which species are included and how they may relate to human activities in the cave. Newton (1999) attempts to define trends through four major stratigraphic sequences, but the samples are rather small and prohibit definitive conclusions. Two human skeletons were discovered at Theopetra Cave, one from the late Upper Paleolithic, which yielded a date of 16,500 +/- 500 BP on bone, and the other from the Mesolithic layers within a pit that contains charcoal dating to 10,420 +/- 180 BP (the skeleton itself was dated to 8,975 +/- 325 BP) (Stravopodi, et al. 1999:270). The late Upper Paleolithic human remains were disturbed at some point in the past and only a few long bone elements and the calvarium were recovered; they are thought to have belonged to an adult male. The Mesolithic human remains are almost completely intact (Stravopodi, et al. 1999). These remains are from a young woman between the ages of 18 and 20, who suffered from slight nutritional deficiencies, possibly a lack of iron (Stravopodi, et al. 1999). A final notable feature from Theopetra Cave is a series of four left footprints, probably from two children between the ages of two and four years old (Manolis, et al. 2000). The prints are from the Middle Paleolithic layers that date in excess of 120,000 BP (Valladas, et al. 2007). Based on experimental studies using stereophotography and photogrammetry, the authors conclude that two of the prints were probably made by an individual wearing some kind of foot covering, while the third was not. The fourth print is 61 somewhat obscured and might indicate an overlap of prints from the two different individuals (Manolis, et al. 2000). Kastritsa Kastritsa is an Upper Paleolithic rock shelter that was discovered and excavated as part of the work conducted in northern Greece during the 1960s (Higgs, et al. 1967). Five stratigraphic layers were uncovered spanning 23,930-19,000 BP (AMS dates on charcoal) (Galanidou and Tzedakis 2001; Galanidou, et al. 2000), though most of the archaeological materials are in the upper three layers that post-date 22,280 BP. The lithic industries were not assigned to a specific cultural period, though they encompass a range defined elsewhere as Gravettian (Adam 2007). Most of the tools in the lowest layer (Stratum 9), which is a beach deposit, are endscrapers; there are few bladelet tools and no evidence of core exploitation (Adam 1997). Laminar blanks were produced in all layers. This period probably represents brief visits to the site. Starting in the overlying archaeological layer (Stratum 7), and especially in Stratum 5 the shelter was used much more intensively. Shouldered points and backed bladelets that were hafted appear at this time (Adam 1997). Stratum 5 contains stone lined hearths and a stone tool industry dominated by flakes; microgravettes are also present (Galanidou 1997, 1999). A small fraction of the raw materials in this layer were transported up to 50 km from the Voidomatis River area (Adam 2007). Refitting studies of the lithics indicate no preserved spatial variation in site use in Stratum 5, rather it is interpreted as a “horizontal palimpsest” of multiple occupations in a short time period (Galanidou 1997, 1999). Stratum 3 contains more exotic raw materials, also from the Voidomatis River (Adam 2007). The technologies in this layer are slightly different; there is an increase in cores and 62 the microburin technique appears. Backed bladelets are common (Adam 2007). In the uppermost layer, the microburin technique persists (Adam 1997). Overall the stone tool types found at Kastritsa are especially diverse, even as compared to other sites in Epirus and those in Peloponnese (Elefanti 2003). The Kastritsa assemblages include an abundance of worked bone and antler, as well as ornaments, in Stratum 5 and above (Adam and Kotjabopoulou 1997; Bailey, et al. 1983; Kotjabopoulou and Adam 2004). Antler points are more common than bone points (Bailey, et al. 1983). Seven perforated red deer canines were recovered, one of which is decorated with incised grooves, in addition to three grooved long bone fragments (Kotjabopoulou and Adam 2004). Despite the fact that Kastritsa is over 60 km from the coast, about two dozen perforated Cyclope and Dentalium shells were found. Finally, two perforated serpentinite “beads” were excavated in Stratum 5 (Kotjabopoulou and Adam 2004). The faunas of Kastritsa were extensively studied by Kotjabopoulou (2001). A total of 4,623 specimens were identified to body class or species, most of which come from the upper three stratigraphic layers. The assemblages are dominated by red deer, followed by birds, wild ass, caprines, and carnivores. Wild pig, roe deer and lagomorphs were also recorded in small numbers (Kotjabopoulou 2001). Water birds (family Anatidae, which includes ducks, geese and swans) occur in the assemblages, though their proportions are unclear (Bailey, et al. 1983; Kotjabopoulou 2001). No temporal trends related to species representation are apparent at Kastritsa. High-utility bovid and cervid body parts are the most common at the site, indicating the preferential transport of these elements (Kotjabopoulou 2001:77). Conversely, equid anatomical profiles indicate a dominance of heads and hooves. This is interpreted as either being the result of scavenging or of butchering at the site and moving 63 higher utility parts elsewhere (Kotjabopoulou 2001:77). However, the author notes that diagenesis may have affected the assemblage (Kotjabopoulou 2001:92) which may explain some relationships between utility indices and survivorship. Using data from cervid tooth eruption, Kotjabopoulou (2001:88) determines that the site was occupied during warm months. Klithi Klithi, in Epirus, is one of the most intensively studied and thoroughly published Paleolithic sites in Greece (e.g., Adam 1989; Bailey 1997, volumes 1 and 2; Bailey 1999; Elefanti 2003; Galanidou 1999). Twenty-two AMS dates on bone collagen, charcoal and burned bone range from 20,250 +/- 1,200 to 16,400 +/- 550 BP, though most dates fall between 19,700 and 16,225 BP (Bailey and Woodward 1997; Gowlett and Carter 1997). Spatial and temporal resolution at Klithi is low, partly because isolated trenches were excavated in multiple areas across the site, and because stratigraphic layers could not be correlated between the trenches (Wenban-Smith 1997). Consequently, analyses of the artifacts treat all layers of the site as a single unit. Local, low-quality raw materials dominate the lithic assemblages at Klithi (Roubet 1997). Backed bladelets are the most common tool type at the site (Adam 1997, 1999b; Elefanti 2003; Roubet 1997, 1999; Roubet and Lenoir 1997). The microburin technique is found throughout the site, in addition to the “transversal Klithian fracture (TKF)” which is an alternate method of removing the ends of backed bladelets (Roubet 1999; Roubet and Lenoir 1997:157-161). Endscrapers are also common, in addition to notches and denticulates, piercers and burins, side scrapers, and composite tools (Elefanti 2003; Roubet 1997). Use- 64 wear studies indicate that some backed bladelets were used as barbs on projectiles, for working hides, or cutting/scraping bone or wood (Moss 1997). A large number of organic artifacts (osseous tools and debitage) were recovered at Klithi (Adam and Kotjabopoulou 1997; Kotjabopoulou and Adam 2004). Needles are common (n = 21) as are awls (n = 20). Spatulas and bone points were also recorded (Adam and Kotjabopoulou 1997:246; Kotjabopoulou and Adam 2004). The analysts recorded two antler tools and eight perforated red deer canines, which is notable because only eleven cervid specimens were recorded in the faunal assemblage (Adam and Kotjabopoulou 1997; Gamble 1997; Gamble 1999). Nearly two hundred marine gastropods were recovered, most of which are perforated. Cyclope is the dominant species, followed by Homalopoma sanguineum, and a few occurrences of other species including Dentalium (Adam and Kotjabopoulou 1997:251; Kotjabopoulou and Adam 2004:40). Over 140,000 faunal specimens were identified at Klithi. The remains are described as well-preserved but highly fragmented (Gamble 1997:210). Gamble (1997; 1999) notes the extreme dominance of ibex and chamois at the site; these species have an NISP count over 12,000, the next closest taxon or class-specific identification is “indeterminate rodent” (NISP = 316), followed by birds (NISP = 202). Indeed, occupants of Klithi focused on ibex hunting in the classic sense described by Straus (1987) (Gamble 1997; Gamble 1999; PhocaCosmetatou 2003a). Analysis of anatomical representation indicates a bias against low-utility parts (Gamble 1997; Gamble 1999). Butchery damage is common in the assemblage; occurring on 13.4% of the remains (Gamble 1997:219). Age profiles based on bone fusion data suggest that the majority of ibex in the death assemblage are adults (4-8 years), whereas tooth eruption and wear data indicate that juvenile and young adult (0-4 years) and old (8+ 65 years) groups are the most well-represented. The author proposes that the hunters at Klithi followed a gourmet transport curve when prime-age adults were hunted, and a bulk utility curve when old or young individuals were killed (following Binford 1978). The author does not, however, address the effects attritional processes may have had on the assemblage. Based on the occurrence of fetal or neonate remains and bone fusion patterns, the site was probably occupied in spring, summer and autumn (Gamble 1997). Boïla Boïla rockshelter is a late Upper Paleolithic site in the Voidomatis Gorge. The deposits are fairly shallow and four archaeological layers or sub-layers were noted. Unit IV is undated and contains Epipaleolithic or Mesolithic tool types. All other layers contain stone tools similar to the Balkan Late Epigravettian (Kotjabopoulou, et al. 1999:202). Unit IIIb ranges between 13,000 +/- 1,350 and 11,850 +/- 600 BP based on dated charcoal and burned bone, respectively. Unit IIIa yielded a date of 14,800 +/- 750 BP and Unit II 16,550 +/- 600 BP, both on burned bone (Kotjabopoulou, et al. 1997:429). These dates should probably be regarded with some caution, as dates on burned bone are often too young, though the dated carbon in Unit IIIb is fairly close to the date on burned bone in that layer, and the stone tool industries match well with what is expected during this time period in Greece. Diachronic change is apparent in the stone tool industries at Boïla. In the lower layers (Units II-IIIb), cores are larger and there are more unilaterally backed bladelets. In Unit IV, the microburin technique appears, along with Sauveterre points and microgravettes, and core size is much smaller, suggesting a more intensive use of raw materials (Kotjabopoulou, et al. 1999:206). In all layers the lithic assemblages indicate that the site was primarily used as a 66 hunting camp where tool maintenance occurred (Kotjabopoulou, et al. 1999:206). Units IIIIIb at Boïla correlate technologically with Stratum 1 at Kastritsa and lithic phases IV-VI at Franchthi Cave (see below). A little over fifty ornaments were recovered from Boïla (Kotjabopoulou and Adam 2004). These were made from marine and freshwater mollusk shells, most of which are perforated Cyclope. Only one modified red deer canine was recorded, and one steatite bead was recovered (Kotjabopoulou and Adam 2004). The faunal assemblage at Boïla is small and highly fragmented, with 1,203 identified bones and teeth (Kotjabopoulou 2001:223). Ungulates dominate the remains, specifically ibex and chamois, followed by red deer. Birds, fish, lagomorphs, and carnivores are also present in the assemblage in lower frequencies (Kotjabopoulou 2001:223). According to Kotjabopoulou (2001) humans at Boïla either had similar body part transport strategies for cervids or caprines, or caprines were hunted at a distance from the site and cervids were taken close to the shelter (depending on if unidentified ungulates are added to the calculations) (Kotjabopoulou 2001:123-124). The analyst offers no analysis of taphonomy or potential attritional processes affecting the faunas. The lack of this information combined with the small sample sizes indicates that her conclusions should be viewed with some caution. Young ungulates are common in the assemblage, and the faunal remains indicate that the site was probably used between late spring and mid-summer (Kotjabopoulou 2001). Grava Grava is an undated rockshelter on the southern part of Corfu that yielded diagnostically Upper Paleolithic stone tools (Sordinas 1969). Corfu was connected to the 67 mainland until after the last glacial maximum (Van Andel and Shackleton 1982), so Paleolithic populations could have moved freely onto the “island” without crossing water. The stone tools have recently been re-analyzed by Adam (2007) who found many similarities to the industries from Strata 3 and 1 at Kastritsa, which coincide temporally with the Gravettian. In general, blades are more common than bladelets in the assemblage. Tool types include unilaterally backed bladelets, endscrapers made from retouched blades, and burins. The author notes that both the tools and the technological styles are quite similar to those observed at Kastritsa (Adam 2007:155-156). Worked bone was also recovered at Grava, which includes a bone point and an incised bone (Sordinas 1969). A perforated red deer canine was also found, along with one ochre-stained pebble (Adam 2007). The faunal assemblage was not studied in depth, though Sordinas (1969) notes that it was abundant. Aurochs are common, and other ungulate species include red deer, fallow deer, roe deer, variously-sized equids, and wild pig. Snails and small mammals such as hare and rodents were recovered, as well as fox, hyena, badger, and mustelids. Sordinas (1969:399) notes “the great number of large birds which (he) was not able to identify.” Based on other Paleolithic Greek faunas, I suspect that at least some of these remains were from great bustard (Otis tarda). Sarakenos Cave In central Greece, on the shores of the now dried Lake Kopais, Sarakenos Cave preserves a sequence that spans the Paleolithic through the Late Helladic period (Sampson, et al. 2009). Though the authors discuss the Paleolithic and Mesolithic periods, they mostly focus on the rich Neolithic layers. The Paleolithic occupations of the site were fairly 68 ephemeral, and few faunal remains were recovered (Sampson, et al. 2009:228). Some potential Early Upper Paleolithic or Middle-to-Upper Paleolithic transitional tools were recovered from an undated layer of Sarakenos Cave. These include Levallois points and flakes, and the authors compare them to transitional industries from Temnata Cave in Bulgaria (Sampson, et al. 2009:214). Sampson et al. (2009) also note side scrapers, denticulated-notched pieces and a blade thinned by the Kostenki technique, but the authors do not provide measurements or further comparisons in support of these conclusions. Late Paleolithic stone tools are tentatively assigned to the Epigravettian period, and include a retouched truncation made from an obsidian blade (Sampson, et al. 2009:214). A radiocarbon date on charcoal from this layer dates to 14,750 +/- 700 BP. Numerous hearth features were uncovered in the Lower Mesolithic layers, which yielded a tight series of radiocarbon dates ranging from 10,315 to 10,435 BP (Sampson, et al. 2009:213-214). The Late Mesolithic deposit dates to between 8,895 and 9,020 BP (Sampson, et al. 2009:214). Sarakenos Cave was more intensively occupied during the Lower Mesolithic, less during the Late Mesolithic (Sampson, et al. 2009:228). Further, during the Late Mesolithic, occupants used local raw materials for the manufacture of stone tools, which Sampson et al. (2009:228) suggests was a result of local groups becoming isolated. The authors note that at this time big game hunting was replaced by the exploitation of birds and plant foods, but they provide no quantitative evidence for this assertion. Sidari Sidari on Corfu contains Mesolithic and later horizons. Discovered in the 1960 the Mesolithic component of the site is a shell midden with a radiocarbon date of 8,750 +/- 800 69 BP (Sordinas 1969, 2003). Lithics from the site are mostly of flint; the source of the most common type is unknown, though the author postulates that perhaps they are nonlocal (Sordinas 2003:91). Non-geometric microlith technology is common, probably destined for use in composite tools, and retouched pieces mostly include points, though scrapers, burins, and other tools occur in low numbers. Backed bladelets are almost entirely absent from the assemblage (Sordinas 1969, 2003). The author notes that the microlithis from Sidari are unlike those found on the Greek mainland. Sordinas (1969:405) notes that small game is present in Mesolithic level of the site, but the majority of the animal remains are from Cardium edule (cockle shells). No quantitative values are provided for the shells or small game. Cave of Cyclope The Cave of Cyclope is located on the southwestern part of the island of Youra in the Aegean. The site contains rich Neolithic and Mesolithic deposits (Sampson 1998). The Lower and Late Mesolithic layers preserve abundant faunal remains, hearth features, land snails, marine shells, fishhooks and lithic materials, as well as a human skull from an older female in the lowest stratigraphic layer (Sampson 1998). Three radiocarbon dates from the Cave of Cyclope place the Lower Mesolithic occupation between 10,335 and 10,530 BP, while dates from the Late Mesolithic range from 9,080 to 9,810 BP (Sampson, et al. 2003:125). The Mesolithic lithic assemblage from the Cave of Cyclope is fairly small (n = 179). Raw material types include nine different kinds of flint, chalcedony, quartz, and a few pieces of obsidian (Sampson, et al. 2003). Cores are found in the assemblage, as well as 70 endscrapers, retouched flakes and notched tools; a burin and two backed pieces were also recorded (Sampson, et al. 2003:127). Evidence of tool manufacture at the site is apparent for all raw material types except obsidian. The obsidian was chemically sourced to the island of Melos, which is also the source of obsidian found in the Mesolithic layers of Franchthi Cave in the Argolid (see below). Obsidian artifacts were not manufactured at the Cave of Cyclope; pieces include eight tools, four blades, two flakes, and only one chip (Sampson, et al. 2003:127). The authors note that tools made of obsidian are different forms than those made of other materials, and include crescents, a backed blade, a truncation, and a trapeze (Sampson, et al. 2003:127). A large osseous tool assemblage was recovered from the site (n = 78 from all Mesolithic layers) and includes fish hooks, bi-points (or gorges), and bone points (MoundreaAgrafioti 2003). Most of the tools were recovered in Late Mesolithic layers. Bi-points, or gorges, are the most frequent tool type in the Mesolithic layers. The author compares them to Paleolithic points: splinters that taper toward one or both ends, constructed of small or large animal long bones or antler (Moundrea-Agrafioti 2003:138). Bi-points are thought to have been used for fishing, along with hooks, meaning that there are two classes of bone fishing artifacts found at the site. Hooks are the second most common bone implement at the Cave of Cyclope and range between 1-4 cm, though two are 5-6 cm long. Some of the hooks are broken but many are intact (Moundrea-Agrafioti 2003). Fish hooks were constructed mainly from ungulate long bone diaphyses, or the bones of small animals or birds. One is made of deer antler, a species that probably was not found on the island (Moundrea-Agrafioti 2003:134). 71 An abundance of faunal remains, including mammals (NISP = 13,350), birds (NISP > 850), and especially fish (NISP > 17,000), were recovered from the Mesolithic layers at the Cave of Cyclope. The faunas in the Lower and Late Mesolithic layers differ from one another. Caprines, mostly goat but also some sheep, dominate the mammalian assemblage of both of the Mesolithic occupations, and suids are present in small amounts in the Lower Mesolithic. Based on biometric and mortality data, Trantalidou (2003:170) postulates that goats in the Lower Mesolithic were in a stage of proto-domestication, while those in the Late Mesolithic were domesticates brought from the mainland. Land snails are also common in the Lower Mesolithic layers, but the author does not consider their dietary contributions. Avian remains include Cory’s shearwater, Manx shearwater, cormorant, shag, sparrow hawk, quail, great bustard, turtle dove, little owl, and raven (Trantalidou 2003:159). The author only discusses Manx shearwater and great bustard remains in detail, though their abundances relative to the other faunal remains are not discussed. It is unclear if their frequencies changed through the sequence. Trantalidou (2003:144) suggests that fish dominate the Lower Mesolithic, but published data on the fish remains (Mylona 2003; Powell 2003) indicate that fish are much more common in the Late Mesolithic. The cranial and vertebral remains of fish were studied separately, the former presented by Powell (2003) and the latter by Mylona (2003). The conclusions of the two studies vary slightly, but their overall findings are the same. At least 29 species of fish were exploited at the Cave of Cyclope; the most common families include sea bream, groupers, mackerels and tuna, and mullets (Mylona 2003:184; Powell 2003:173). Both studies indicate that a wider range of species were identified in the Late Mesolithic 72 layers (Mylona 2003; Powell 2003), which may or may not be the result of larger sample sizes in these levels. SOUTHERN GREECE Recent work on the Mani Peninsula in southern Peloponnese has yielded several Paleolithic cave sites. Some of these sites are in advanced phases of excavation and publication (Kalamakia, Lakonis, Apidima) and can be discussed at some length below. Other sites on the Mani Peninsula have only undergone survey or initial testing and therefore cannot be covered in detail here (see Darlas and Psathi 2008 for some recently discovered Upper Paleolithic sites in the region). Kalamakia Cave Kalamakia Cave is a karstic limestone cave on the coast of southern Peloponnese that preserves Middle Paleolithic deposits. Seven stratigraphic layers were excavated and archaeological materials are found in layers III, IV and V (layers with lower numbers are stratigraphically older) (Darlas and de Lumley 1999, 2004). The excavators believe that the lowest layer was deposited in MIS 5e, despite a marine shell in layer II that was dated to 109,000 +/- 14,000 BP using the uranium disequilibrium method (Darlas and de Lumley 1999, 2004). The earliest archaeological evidence probably dates to MIS 5c, and a radiocarbon date from layer IV provides a minimum age of 40,050 years (Darlas 2007). Initial studies of the Kalamakia lithics indicated that there was a slight difference between the lower and upper archaeological layers. The artifacts from the lower levels were defined as typical Mousterian, with frequent high-quality Levallois flakes and numerous 73 flakes with facetted striking platforms (Darlas and de Lumley 1999). Conversely, artifacts from the upper level belong to a more “Mousteroid” industry because they are of moderate quality and do not belong to a standard Mousterian type. There is also a lack of facetted platforms and low frequency of Levallois flakes (Darlas and de Lumley 1999). Subsequent studies of the lithics have stressed that tools in the assemblage are typical Mousterian made with Levallois debitage (Darlas 2007; Darlas and de Lumley 2004). Raw materials include flint from 12-20 km from the site, and andesite, quartz and quartzite (Darlas 2007). Despite the abundance of Levallois debitage in the assemblage, few points were recovered (Darlas 2007; Darlas and de Lumley 2004), which may indicate that the site was used for tool manufacture. A unique feature of the Kalamakia Middle Paleolithic assemblage is the presence of a few shells of Callista chione (a smooth-shelled large clam). The shells were probably not consumed, rather they were retouched and possibly used as scrapers (Darlas 2007; Darlas and de Lumley 2004). This phenomenon is also reported at Grotta dei Moscerini in Italy (Stiner 1994:187-188 and references therein). Mammalian faunal remains are in the process of being analyzed and published. A wide range of ungulates are recorded at Kalamakia, dominated by wild goat, chamois and fallow deer, but also including extinct elephant and rhinoceros, wild pig, red deer, roe deer, and a large bovid (possibly aurochs or bison) (Darlas 2007; Darlas and de Lumley 1999; Gardeisen, et al. 1999). At this point no trends are apparent in the faunas (Darlas 2007). Ample carnivore remains in some of the layers indicate that the site alternated between a human occupation site and carnivore den (Darlas and de Lumley 2004). Large and small carnivores are found at the site, including brown bear, leopard, wolf, Eurasian lynx, wild cat, 74 fox and mustelids (Darlas 2007; Darlas and de Lumley 1999, 2004; Gardeisen, et al. 1999). Microfauna are abundant in the Kalamakia stratigraphic series and include 59 species of small mammal, amphibians, reptiles, and moderate amounts of tortoise. Much of the microfauna is thought to be refuse from raptors or owls that nested in the site (Roger and Darlas 2008a). A range of avian fauna was recorded in layers IV and V. Avian NISP counts are low (NISP = 95) but represent twenty-one species (Roger and Darlas 2008b). The most common species in the assemblage is quail, followed by rock partridge and pigeon, though several species of diurnal raptor and owl were also recorded (Roger and Darlas 2008b). No human modifications, such as cut marks or burning damage, are apparent on the bird remains and only one incidence of raptor digestion was recorded. The authors suggest that the most widely represented avian taxa, such as the partridges, quails and pigeons (which also appear at other Greek Paleolithic sites) were collected by humans, while smaller birds represent the remains of raptor meals (Roger and Darlas 2008b). Hominin remains were identified in layer IV and include several Neandertal teeth, a cranial fragment, fibula, and lumbar vertebrae (Darlas 2007). Lakonis I Lakonis I, a limestone cave in the southern Peloponnese, is reported to contain stratified Middle (Units IV to Ib) through “Initial” Upper Paleolithic (Unit Ia) deposits. TL and U-series dating of beach rock and speleothems indicates that the earliest Middle Paleolithic was deposited between 130,000 and 120,000 BP (Elefanti, et al. 2008; Panagopoulou, et al. 2002-2004). A series of AMS dates on charcoal place the late Middle Paleolithic and earliest Upper Paleolithic between 46,600 +/- 6,050 and 38,600 +/- 2,450 BP, 75 though the authors note that the range of dates are statistically indistinguishable between the two layers (Elefanti, et al. 2008; Panagopoulou, et al. 2002-2004). Raw materials at Lakonis I are mostly local, found within about 5-10 km of the site (Panagopoulou, et al. 2002-2004:331). The Middle Paleolithic layers are defined by the production of blanks using the Levallois laminar method as well as non-Levallois methods. The most common tool types are side scrapers and retouched Levallois and Mousterian points; bifacial scrapers, knives and points are also found (Panagopoulou, et al. 20022004:336). Most points were produced using châpeau de gendarme platform preparation (Panagopoulou, et al. 2002-2004:332). The uppermost archaeological deposits (Unit Ia) contain a mix of Middle and Upper Paleolithic tool types and technological styles (Elefanti, et al. 2008; Panagopoulou, et al. 2002-2004), which may indicate some mixing in certain parts of the site. However, articulated faunal remains, refitted lithics, and micromorphology are taken to indicate that the layer is stratigraphically sound (Elefanti, et al. 2008; Panagopoulou, et al. 2002-2004). Raw materials in the Initial Upper Paleolithic layers are mostly local, and flint is much more common than in underlying deposits (Elefanti, et al. 2008:89). Levallois and prismatic blank production techniques are found in the layer. Bladelet manufacture is common, using two different chaînes opératoires (Elefanti, et al. 2008:92). Stone tool types include Mousterian points and notches/denticulates, which are considered Middle Paleolithic, though manufactured as part of prismatic core production, similar to the Central European Bohunician (Elefanti, et al. 2008:92). The most common tools in Unit Ia are Upper Paleolithic in nature: retouched flakes, blades and bladelets, as well as burins, endscrapers, and a few Aurignacian scrapers and blades (Elefanti, et al. 2008:92). Panagopoulou et al. 76 (2002-2004) argue that the IUP industry is indicative of an in situ transition from Mousterian to Upper Paleolithic technologies. The faunal remains from Lakonis I are still undergoing analysis, and preliminary publications stress the extreme fragmentation of the assemblage (Panagopoulou, et al. 20022004). Panagopoulou et al. (2002-2004:339) note that as of 2001, 4.4 kg of faunal remains were recovered from Lakonis I. Of this, 818 specimens are bone or dental specimens larger than 2 cm. In this subset 73 specimens were identified (Panagopoulou, et al. 2002-2004:339). The remains exhibit a high frequency of burning damage, with 67% (NISP = 31) of the identifiable remains burned (Panagopoulou, et al. 2002-2004:340). The small size of the assemblage makes stratigraphic divisions impossible, though the authors do attempt to discuss anatomical representation. Identified taxa include fallow deer, red deer, roe deer and pig, as well possibly aurochs and rhinoceros (Panagopoulou, et al. 2002-2004:340). No carnivore remains were noted, and small game was not mentioned in the assemblage. Based on a few fallow deer teeth, both adult and juvenile animals are present at the site (Panagopoulou, et al. 2002-2004:341). One of the most notable features of the Lakonis I project is the recovery of a Neandertal lower left third molar from layer identified as Initial Upper Paleolithic (Unit Ia) (Harvati, et al. 2003; Panagopoulou, et al. 2002-2004). This would be one of only a few sites in Europe where human remains were found with a transitional stone tool industry, and in this case it suggests that Neandertals were the makers of the IUP industry at Lakonis I (Harvati, et al. 2009). Strontium levels in the tooth were analyzed using laser-ablation PIMMS (Richards, et al. 2008). This analysis indicates that the enamel in the tooth formed when the individual lived in an area with older, more radiogenic volcanic bedrock, as 77 opposed to the limestone cave in which it was found. The authors conclude that this Neandertal lived as a child in an area at least 20 km away from Lakonis I (Richards, et al. 2008, but see Nowell and Horstwood 2009, who suggest that a different method of correcting for isobaric inferences for Sr may actually indicate that the tooth formed locally). Apidima Apidima is a locality on the Mani Peninsula that includes four caves (Α, Β, Γ, Γ) that contain Pleistocene deposits. Lithic artifacts and faunal remains were recovered, but the site is better known for its human remains (Pitsios 1985, 1995, 1999). Two Middle Pleistocene skulls were recovered Cave A, and an Upper Paleolithic burial of a 17-23 year-old female was found in Cave Γ (gamma) (Ligoni and Papagrigorakis 1995). The skeleton in Cave Γ was found laying on one side in the fetal position, with her head on a slab of stone, and the burial was associated with a disk-shaped stone, animal bones, and perforated shell ornaments (Cyclope neritea) (Karali-Giannakopoulou 1995:6-7; Pitsios 1999). The burial is thought to date to about 30,000 BP based on associated lithics and the stratigraphic layer in which it was found (Darlas 1995; Pitsios 1999). No reliable radiometric dates are available (but see Lyritzis and Maniatis 1989, 1995). A preliminary paper on the small lithic assemblage from Cave Γ suggests an Aurignacian affiliation for the stone tools, though some Mousterian characteristics are also apparent (Darlas 1995). Tsoukala (1999) presents a paleontological discussion of the faunal remains at Apidima. In Cave Γ, she notes a wide range of small and large carnivores, including leopard, lynx, wild cat, badger, and stone marten, as well as ibex, red and fallow deer, Megaloceros, and Elephantidae. Based on the presence of juvenile carnivores, the 78 author postulates that the cave was occasionally used as a den. Small animals, such as rodents, birds, lagomorphs, and turtle bones are also present (Tsoukala 1999). Damage from burning and butchery marks are present. Unfortunately, no quantitative values indicating species frequency or taphonomic factors are available. In an analysis of the fauna of the four caves, Lax (1995:133) determines that the faunal accumulations of Cave Γ (NISP = 7,383) are largely natural, despite occasional burning (7.3%) and cut or chop marks on at least two specimens. Kephalari Cave Kephalari Cave was exceptionally well excavated (including wet sieving of the sediments) but never fully published. It lies in a network of caves above the small town of Kephalari, about 6 km southwest of Argos (Felsch 1973; Reisch 1976). It was excavated from 1972 to 1975 by the German Archaeological institute at Athens (Felsch 1973; Reisch 1976). The site contains a late Middle Paleolithic component, Upper Paleolithic layers, and disturbed Neolithic deposits. Unfortunately, no radiocarbon dates are available for the site, but diagnostic lithics indicate that it contains some kind of early Upper Paleolithic, Aurignacian, Gravettian and Epigravettian components (Hahn 1984; Reisch 1976). Occupations associated with the late glacial and the Last Glacial Maximum (LGM), particularly the Gravettian, tend to be absent in other Peloponnesian sites (Adam 2007; Karkanas 2010; Koumouzelis, Ginter, et al. 2001). If Kephalari Cave turns out to contain LGM deposits, it has a lot of potential for adding to our understanding of the Paleolithic in southern Greece. The lithics are currently under analysis by Dr. Gilbert Marshall. Worked 79 bone is fairly abundant in the Kephalari Cave assemblage, and shell ornaments are common, though no quantitative values are currently available. A small portion of the Upper Paleolithic faunal remains were published in a brief account by Reisch (1976). He notes a range of small animals in the assemblages, including hedgehogs, bats, squirrel, and voles. Other small game species that were potentially of economic importance at other Paleolithic sites in southern Greece include hare, tortoise, partridge and fish (Reisch 1976). Small carnivores occur in the assemblages, including fox, lynx, and wild cat, as well as some large hyena or leopard-sized carnivores. Ungulate species identified include wild ass, wild pig, red deer, wild cattle, and wild goat (Reisch 1976). The Kephalari Cave fauna is currently under study by the author and preliminary analysis confirms the species list proposed by Reisch (1976). This new analysis also indicates that large ungulates were much more common in the older layers, and hares, fish, and especially partridge are found in large numbers in the later Paleolithic layers. Franchthi Cave Franchthi Cave, on the coast in the southeastern Argolid, is probably the most wellknown prehistoric site in Greece. It was excavated as part of a long-term interdisciplinary study and resulted in an extensive publication series (e.g., Farrand 2000; Hansen 1991; Jacobsen and Farrand 1987; Perlès 1987, 1991; Shackleton 1990; Stroulia 2010; Van Andel and Sutton 1988; Wilkinson and Duhon 1991, in addition to many articles and book contributions). The site preserves Upper Paleolithic through Neolithic deposits, with a possible Middle Paleolithic component below the modern water table that could not be excavated (Farrand 2000). The Campanian ignimbrite tephra, a well-known volcanic event 80 that occurred in Italy about 39,280 +/- 110 BP (40Ar/39Ar) (De Vivo, et al. 2001) is recorded at the base of the Upper Paleolithic sequence; an Aurignacian occupation post-dates this event, possibly by a few thousand years. Between about 33,000 BP and 25,000 BP, and after 15,000 BP, there was heavy use of the cave, punctuated by a few hiatuses (Farrand 2000). The site is well known for its Neolithic component, but the Paleolithic and Mesolithic are the focus of the review here. Perlès (1987; 1991; 1999) divides the Upper Paleolithic and Mesolithic into nine lithic phases based on stone tool abundance and technologies. Lithic phase I contains a small sample but is typologically similar to the Aurignacian (Perlès 1987, 1999). A hiatus separates lithic phases I and II; radiocarbon dates place the latter at 22,400 +/- 750 or 21,950 +/- 3,000 BP (Jacobsen and Farrand 1987). Lithic phase II mostly contains single and double backed bladelets, though endscrapers and notches are also present (Perlès 1987, 1999). Lithic phase III is stylistically similar to the preceding phases, with a higher proportion of single-backed bladelets (Perlès 1987, 1999). A major hiatus spanning the LGM divides lithic phases III and IV. Lithic phase IV dates to 14,850 +/- 800 BP (Jacobsen and Farrand 1987) and contains backed bladelets, though there is no apparent reduction sequence for the manufacture of the bladelets. The microburin technique appears at this time, as well as “La Mouillah points” in the assemblage (Perlès 1987, 1999). In general, there is an increase in typological diversity in lithic phase IV. During lithic phase V, which dates to 13,200 +/- 300 BP (Jacobsen and Farrand 1987), the microburin technique was not used, and the first geometrics appear (Perlès 1987, 1999). Double-backed bladelets are common, and there is further diversification of tools, including notches, denticulates and endscrapers. Perlès (1987) interprets this to be a different 81 lithic tradition than those below. Four radiocarbon dates are available for lithic phase VI, ranging between 12,600 +/- 1,450 and 12,100 +/- 750 BP (Jacobsen and Farrand 1987). Production techniques in this phase are the same as those in lithic phase V, but there are some typological changes. New geometrics appear in phase VI, and there is a further increase in diversity of tool types. Also at this time obsidian from the island of Melos appears, indicating seafaring in the Aegean (Perlès 1987). There is a small hiatus between lithic phases VI and VII, which marks the beginning of the Lower Mesolithic at Franchthi Cave. Seven radiocarbon dates were taken from the Lower Mesolithic layers and span 10,725 +/- 475 and 10,275 +/- 375 BP (Jacobsen and Farrand 1987). Stone tools in the Mesolithic differ drastically from those in the late Upper Paleolithic in type and decline in quality in the later periods (Perlès 1991). Backed bladelets and geometrics decrease in frequency in lithic phase VII. Most tools in the Lower Mesolithic are laterally retouched and include notches, denticulates and endscrapers (Perlès 1991, 1999). Lithic phase VIII represents the Middle Mesolithic and has five radiocarbon dates that range between 10,000 +/- 400 and 9,550 +/- 300 BP, in addition to one questionable early date (Jacobsen and Farrand 1987). There is a general paucity of lithic remains but an increase in microliths at this time, which the author relates to tuna fishing (Perlès 1991). The Final Mesolithic, lithic phase IX, dates to 9,215 +/- 215 BP (Jacobsen and Farrand 1987). The lithic assemblage is quite similar to that during the Lower Mesolithic, with crude tools such as notches and denticulates, with some arrowheads (Perlès 1991, 1999). Bone tools are rare in the Franchthi assemblage. They are found in the late Paleolithic layers (lithic phase VI) and in the Lower Mesolithic (lithic phase VII), though no extensive publications are available on osseous tools at the site. The ornament assemblage from 82 Franchthi Cave is large, with about 10,000 shell ornaments from the Upper Paleolithic and Mesolithic deposits (Perlès and Vanhaeren 2010). The most common shells in the assemblage are from Cyclope neritea, though Columbella rustica and Dentalium sp. are also found (Perlès and Vanhaeren 2010). Interestingly, the shell ornaments at Franchthi Cave were likely heat treated, possibly for aesthetic reasons. Perles and Vanhaeren (2010) note that Cyclope neritea shells (both perforated and unperforated) display higher incidences of burning, by several orders of magnitude, than other organic remains in the assemblage, including animal bones and land snail shells, possibly for aesthetic purposes. Faunal remains at Franchthi were initially studied by S. Payne (1975; 1982), who determined that changes in the relative proportions of large mammals from about 25,000 BP onward were related primarily to environmental but also economic change. He documented an increase in small animals, such as birds, mollusks, and especially fish by the Mesolithic (Payne 1975:128). The fish component in the Franchthi sequence shifted fairly dramatically from the exploitation of brackish-water species such as eel, sea bream and gray mullet in the Late Upper Paleolithic and Lower Mesolithic (ca. 13,020-10,230 BP) to open-water taxa such as large bluefin tuna, grouper and barracuda in the Late Mesolithic (10,230-8,890 BP) (Rose 1995). In this latter phase, the quantity of fish also increases in the faunal assemblage. Marine mollusk exploitation shifted during this period, from shallow-dwelling Patella, Monodonta and Gibbula at the end of the Upper Paleolithic to brackish-water Cyclope nerita (for ornament-making) during the Lower Mesolithic and to Cerithium, a slightly deeper water-dwelling gastropod, in the Late Mesolithic (Shackleton 1990). Subsequent ongoing analyses of a large sample of the Franchthi faunas (NISP > 17,000) are presented by Stiner and Munro (2011). They identify eight major faunal zones 83 that relate in part to changes in the local environment and resource intensification that may have accompanied human population growth. These faunal zones do not altogether correspond to lithic phases described above, so cultural designations are indicated. Upper Paleolithic layers include Zone VIII (Aurignacian), which contains a diverse set of ungulates dominated by red deer, and Zone VII (some Aurignacian/Gravettoid), which displays an increase in wild ass and indicates a more open environment. The Epipaleolithic includes Zone VI (Epigravettian) corresponds to an increase in small game exploitation, including land snails, shellfish and coastal fishing. Zone V (Epigravettian) is marked by increased cultural inputs, a further decline in ungulate diversity, and shift from tortoises to pond turtles. Zone IV (Final Paleolithic) is dominated by red deer, and a notable disappearance of opendwelling taxa such as ground birds, aurochs and wild ass. The Mesolithic includes three faunal zones: Zone III (Lower Mesolithic) has an increase in burned bone and a corresponding increase in seed exploitation (Hansen 1991), Zone II (Mesolithic) contains barracuda and tuna from open/deep water, and Zone I (Late Mesolithic) is dominated by open or deepwater fished tuna (Stiner and Munro 2011). For the most part, shifts in large terrestrial fauna (ungulates) correlate with changes in environmental conditions . Two, consecutive trends of increasing dietary breadth are indicated by the small game data, the first within an exclusively terrestrial context, and the second as marine habitats came into use through the end of the Mesolithic. The intensity of the human occupations at this site increased in tandem with intensified use of animal and plants (Stiner and Munro 2011). Extensive botanical remains were recovered from Franchthi Cave, allowing for local environmental reconstructions (see Chapter 3) as well as an evaluation of the contribution of vegetation to the diet of foragers at the site (Hansen 1991). In the late Upper Paleolithic, from 84 about 15,350 to 13,020 BP, edible species found in the site include pistachios, almonds, pears, legumes (bitter vetch, lentils, vetch/vetchling), wild oats and barley (Hansen 1991:160). The late Upper Paleolithic/Lower Mesolithic transitional zone (cf. 10,495-10,230 BP) contains mostly pistachio, almond, pair and oats, with an increase in wild oats and barley (Hansen 1991:161). The first part of the Late Mesolithic (10,230-9,645 BP) yielded few botanical remains, but the period that follows until about 8,705 BP is similar in botanical content to the late Upper Paleolithic/Mesolithic zone, with the addition of coriander (Hansen 1991:161). These data suggest the use of a wide suite of plants, thought it is difficult to ascertain the quantitative role that each played in the diet. Multiple burials and cremations were recovered from the Lower Mesolithic deposits. Few Mesolithic human remains have been found in Greece, and Franchthi Cave is invaluable in understanding mortuary practices from this period in the region. Seven fragments of human bone were recovered in the Paleolithic layers, but the provenience for these remains is not mentioned (Cullen 1995:274). At least eight burials were recovered in the Lower Mesolithic layers: an infant, two young adults (male and female), three females and two males, ranging in age from 20 to 45 years (Cullen 1995:275). At least two of the Lower Mesolithic burials were also cremated. Only one possible intact burial was found in the Late Mesolithic, a small collection of infant remains (Cullen 1995:272). Additionally, dozens of isolated human bones and teeth were found throughout the Mesolithic layers, representing 725 individuals (nineteen is likely based on spatial distribution across the site) (Cullen 1995:278). The author suggests that human bone scattered across the site is the result of postdepositional disturbances, not as part of a funerary process (Cullen 1995:280). 85 THE LATE MIDDLE PALEOLITHIC THROUGH MESOLITHIC ARCHAEOLOGICAL SEQUENCE IN GREECE One of the most striking features of the Middle Paleolithic through Mesolithic record in Greece is the rarity of sites with a long or continuous occupation sequence that preserves both lithic and faunal materials. Klithi, for example, has a large faunal assemblage, but the site seems to represent a fairly short occupation span. Theopetra Cave, on the other hand, contains a long stratigraphic sequence, but it has many long hiatuses and an overall lack of faunal materials. Franchthi Cave, despite its many depositional hiatuses before the late Upper Paleolithic, preserves by far one of the longest sequences of occupation with abundant lithic, faunal and botanical remains. Of the sites discussed above, only Theopetra Cave and Sarakenos Cave contain Middle Paleolithic, Upper Paleolithic and Mesolithic deposits; the former is riddled with gaps and mixed deposits, and the extent of the Middle Paleolithic layers of the latter are limited. Therefore, we are left to patch together the Late Pleistocene and Early Holocene landscape in Greece from barely over a dozen archaeological sites. Mousterian deposits are found at several Greek sites, including Asprochaliko, Theopetra, Lakonis I, and Kalamakia. At all of these sites there appear to be technological differences between the earlier and later Middle Paleolithic layers, though the patterns of variation are not similar among the sites (e.g., the “micro-Mousterian” at Asprochaliko is not found elsewhere). Middle to Upper Paleolithic transitional industries are absent in Greece, with the possible exception of Lakonis, where a potentially stratigraphically sound layer contains a mosaic of Middle and Upper Paleolithic production methods and tool types. It was once thought that Theopetra Cave contained a transitional industry, but TL dates indicate that the subject layer is, in fact, mixed. Aurignacian deposits are found at many Greek sites 86 (Franchthi, Apidima, and possibly Kephalari), but only those at Franchthi are published. Deposits dating to the same interval as the Eastern European Gravettian occur in Greece, but the lithic industries are often called “Gravettoid” or “Gravettian-like”; cultural manifestations of this time period seem to take on a regional variation (e.g., Asprochaliko, Kephalari, Kastritsa and Grava) (see also Adam 2007). The LGM is largely absent in the Greek archaeological record. It may be represented at Kephalari Cave, but this is not certain as the deposits are undated. Some radiocarbon ranges span this period, such as at Kastritsa, but at most sites the LGM corresponds to a hiatus (Franchthi, Theopetra). It is unclear if Greece was depopulated during the LGM, if the sites did not preserve, or if people were concentrated in coastal areas that are now underwater. Late Upper Paleolithic sites are comparatively common, and often contain Epigravettian components (e.g., Theopetra, Klithi, Boïla, Sarakenos Cave, Kephalari and Franchthi). Mesolithic deposits are found at Theopetra, Boïla, Sarakenos Cave, the Cave of Cyclope and Franchthi Cave; the latter two contain extensive archaeological materials. Another feature of the Greek Paleolithic is that lithic raw material sources appear to be primarily local and often of low quality. This may seem to be the case because no comprehensive provenance studies have been conducted, and because many raw materials used in the Paleolithic are not conducive to sourcing. The situation changes at the end of the Upper Paleolithic and in the early Mesolithic when non-local materials, particularly obsidian from Melos, are found at sites in Peloponnese (Franchthi Cave), as well as in the northern Aegean (Cave of Cyclope). Organic artifacts, including bone points, fishhooks and ornaments, appear during the Upper Paleolithic in Greece and proliferate in the Mesolithic at 87 some sites. A few questionable early pieces were recovered, as well as the retouched shell scrapers at Kalamakia, but these are fairly rare in the Middle Paleolithic. Faunal remains are difficult to compare among the sites discussed above, mostly because of preservation issues or an altogether lack of published quantitative data. The large game component of Paleolithic Greek sites may be largely reflective of local environments rather than deliberate variation in prey choice by human hunters. The dominance of ibex and chamois in the faunas of Klithi, for example, correlate with the mountainous region in which the site is located. At sites such as Apidima, Asprochaliko, Kalamakia, Lakonis, megafauna that later go extinct (Pleistocene rhinoceros and elephants) occur in the lower layers. Carnivores are common in many but not all of the faunal assemblages discussed above (e.g., Kalamakia, Theopetra, Apidima). Without rigorous taphonomic studies of these assemblages, it is unclear the extent to which carnivores shaped at least some of these archaeofaunas (see Chapter 4). Diachronic change in faunal assemblages is also difficult to ascertain from the sites discussed above with the exception of Franchthi Cave, because of small samples and the limited nature of temporal sequences. One robust trend comes from Greek coastal sites dating to the Late Pleistocene and Early Holocene (e.g., Franchthi Cave, Cave of Cyclope and Sidari), where there is an increase in use of marine resource. Contemporary terrestrial Mesolithic assemblages are essentially absent in the Greek record (the faunal sample from this time period is exceptionally small at Theopetra cave). It is unclear at this point if this represents a general movement of human populations to coastal areas, or if it is a bias in the archaeological record in both the location of excavated Mesolithic sites and the preservation of terrestrial faunas at inland sites that do not contain Mesolithic layers. 88 While terrestrial mobility may have declined towards the end of the Pleistocene, there are indication of increased mobility on marine waters via innovations in seafaring, particularly during the Mesolithic period. This is evidenced by tuna fishing, and the appearance Melos obsidian on the mainland and on other Aegean islands. The inclusion of fish and marine mollusks at coastal sites indicates a widening of the diet breadth of Early Holocene foragers in this region. Klissoura Cave 1 has the potential to fill many of the gaps in the regional record discussed above. The site is 12 km inland, and contains Middle Paleolithic through Mesolithic deposits. It also preserves an Early Upper Paleolithic layer and thick Aurignacian deposits unlike any found in Greece. There are some depositional hiatuses, including during the LGM, but there are ample, well-preserved lithic, faunal and botanical remains in almost all of the cultural layers represented at the site. I now turn to a discussion of Klissoura Gorge and its caves, specifically Klissoura Cave 1. KLISSOURA GORGE AND KLISSOURA CAVE 1 Klissoura Gorge is part of a valley that stretches for approximately 3 km, connecting the Berbati Valley and Argive Plain in northeastern Peloponnese (Figures 2.1, 2.3). The gorge narrows to 500 m wide in the area that contains several karstic cave sites of the same name. The valley formed in Triassic limestone and is drained by the Berbatiotis River (Koumouzelis, Ginter, et al. 2001). Thirty-five caves and rock shelters were systematically surveyed in the gorge, six of which contain surface archaeological materials, while a handful of others preserve stratified sequences (Koumouzelis, et al. 2004). The gorge was initially surveyed and recorded during the Berbati-Limnes Archaeological survey in 1988-1990 89 Figure 2.3 Photograph of Klissoura Gorge, facing southwest. (Runnels 1995a). A joint project between the Ephory for Paleoanthropology and Speleology in Athens and the Institute of Archaeology at the Jagiellonian University in Kraków tested Klissoura Caves 4 and 7 in 1993, while systematic excavations began at Klissoura Cave 1 in 1994 (Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001; Koumouzelis, et al. 1996). Today, the gorge lies roughly 12 km inland; in the past the closest marine shorelines were slightly further away (Van Andel and Shackleton 1982) (Figure 2.4). Three 1x1 meter test trenches were excavated in Klissoura Cave 7, yielding three stratigraphic layers with a few hundred lithic artifacts in each layer (Koumouzelis, et al. 2004). A radiocarbon date on calcium carbonate yielded a date of 14,850 +/- 700 BP, but the authors contend that the true age is slightly younger (Koumouzelis, et al. 2004:35). Surface finds included some tools and flakes attributable to the Mesolithic and Upper Paleolithic, and 90 Figure 2.4 Comparison of modern landmass (grey fill) to LGM shorelines (black line and white fill, indicating land inundated today) and paleolakes (black fill). Adapted from Petit-Maire et al. (2005). the authors propose that stone tools found in the excavated layers correlate stylistically with lithic phases V and VI at Franchthi Cave, which are late Upper Paleolithic (Perlès 1987). At Klissoura Cave 4, two 1x1 meter test trenches were excavated, one of which contained primarily mixed sediments. The second trench includes several layers of mixed Neolithic and Paleolithic artifacts, overlying several intact late Upper Paleolithic deposits (Koumouzelis, et al. 2004). The authors postulate that these layers also correspond to the late UP at Franchthi Cave, specifically lithic phases V and IV (Perlès 1987). A bone sample from 91 one of the lower layers of Klissoura Cave 4 dates to 13,600 +/- 750 BP (Koumouzelis, et al. 2004:43). The initial excavations at Klissoura Cave 1 (Figure 2.5) established the presence of Holocene layers containing Mesolithic flake industries and Late Pleistocene layers that include Aurignacian-like industries. Subsequent excavations revealed a deep series of deposits that include thick Upper and Middle Paleolithic components (Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001; Sitlivy, et al. 2007). The site was excavated in 1x1 meter units (Figure 2.6) in 5 cm arbitrary levels, and all sediments were screened. Excavations at the site began in 1994 and concluded in 2006. Figure 2.5 Photograph of Klissoura Cave 1, facing northwest. 92 Figure 2.6 Plan map of Klissoura Cave 1. Adapted from Karkanas (2010). Stratigraphy and Dating of Klissoura Cave 1 In total, about a six meter deep trench was excavated from Klissoura Cave 1, with twenty-one stratigraphic layers and various sub-groups in many of the layers. Table 2.1 lists the excavated layers along with their cultural designations as defined by stone tool typologies, radiocarbon date ranges, and depositional histories (Kaczanowska, et al. 2010; Kuhn, et al. 2010; Sitlivy, et al. 2007). Certain stratigraphic units were grouped together to improve sample sizes for the faunal study. Middle Paleolithic layers are grouped by sedimentary units that formed in similar or related depositional environments (Sitlivy, et al. 2007; Karkanas personal communication 2009), while the Upper Paleolithic and later layers correspond to changes in material culture (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001). Koumouzelis, Ginter, et al. (2001) list a first series of radiocarbon dates from the Upper Paleolithic and later layers of the excavation. 93 However, these dates were taken mainly on ash-derived carbonate fraction in the sediments, which likely underwent some recrystallization (Koumouzelis, Kozlowski, et al. 2001; Kuhn, et al. 2010). Therefore, most of these initial dates probably represent minima and are not presented here. A new series of AMS radiocarbon dates were recently established for Klissoura Cave 1, based almost entirely on carbon, typically wood charcoal (Kuhn, et al. 2010). Pre-treatment techniques for cleaning samples obtained from the early Upper Paleolithic and late Middle Paleolithic layers include the ABOX technique. The ABOX "super-clean" pre-treatment method was developed by Bird et al. (1999) and employs wet oxidation and step heating of samples that are subsequently processed in a special vacuum line (Pigati, et al. 2007). This method is well-equipped to deal with samples that date to Table 2.1 Stratigraphic layers for Klissoura Cave 1 and corresponding cultural units based on stone tool types. Layer groupings are used throughout the dissertation for analyzing faunal remains. From Kaczanowksa et al. 2010, Karkanas 2010, personal communication 2009, Kuhn et al. 2010. 94 beyond the range of traditional radiocarbon dating, and samples that contain a small amount of carbon. Kuhn et al. (2010) present the full list of dates taken at Klissoura Cave 1; an abridged version of this is provided in Table 2.2. Certain layers are excluded (specifically 6, 6a, 7, and VI) because analysis of the artifacts shows these layers to be mixed. I also exclude dates from areas with ambiguous contacts between layers; for example, it is not clear if a date from layer “IV/V” is lower Aurignacian or Early Upper Paleolithic. Four radiocarbon dates come from the Middle Paleolithic layers, ranging from roughly 53,600 to 62,300 BP (Table 2.2, note the large error ranges). The three oldest dates are considered minimum ages for the two Middle Paleolithic layers from which they were derived. Radiocarbon dates from the Early Upper Paleolithic (Uluzzian) layer V remain problematic, which is unfortunate because this layer represents a rare industry in Greece. Most of the dates are either anomalously young, infinite, or of questionable provenience (Table 2.2). Kuhn et al. (2010:41) propose two scenarios for the age of the layer. The first is that the older date (43,841 +/-764 BP) is correct, as are two similar dates from the underlying mixed layer VI, which would indicate a hiatus of 6,000-7,000 years separating the EUP and lower Aurignacian layers. The second possibility is that the ~40,000 BP dates are intrusive from the late Middle Paleolithic, and the EUP layer V was deposited sometime between 33,000 and 40,000 BP. Recent analysis of a micro-tephra from the interface between layer IV and V indicates that it is the Campanian Ignimbrite (Karkanas, personal communication 2011), which is also found at Franchthi Cave and dates to 39,280 +/- 110 BP using 40Ar/39Ar (De Vivo, et al. 2001). This is consistent with the first scenario proposed by Kuhn et al. 95 Table 2.2 Radiocarbon dates from Klissoura Cave 1. From Kuhn et al. (2010). 96 (2010), where there is a large hiatus between the EUP and Aurignacian. It seems likely, therefore, that the Uluzzian of layer V is greater than 39,000 BP. Some anomalously young dates for the rest of the Upper Paleolithic layers are noted in Table 2.2, but for the most part the dates from these layers appear sound. The Aurignacian at Klissoura Cave 1 was probably deposited between 35,000 and 37,500 BP (Kuhn, et al. 2010:40). The Mediterranean-backed bladelet industry (III’), which is the uppermost of two distinctly non-Aurignacian layers (see below), yielded a date of 35,381 +/- 690 BP (Table 2.2), as well as an anomalously young date. No dates in the new series address the ages of the Epigravettian (IIa-d) and Mesolithic (3-5a) layers, but some are presented in Koumouzelis, Kozlowski et al. (2001). These are radiocarbon dates on sediment carbon fraction that yielded ages of 17,150 +/- 600 and 10,350 +/- 800, respectively. Stratigraphy The depositional history of Klissoura Cave 1 is fairly well understood. Geoarchaeological analyses and micromorphology of the Middle Paleolithic layers are still being studied (but see Sitlivy, et al. 2007 for preliminary results), but the analysis of the Upper Paleolithic layers are completed (Karkanas 2010; Karkanas, et al. 2004). The Middle Paleolithic sediments contain a much higher geogenic input, as opposed to the heavily anthropogenic signature of the Upper Paleolithic layers (Sitlivy, et al. 2007). The lowest sedimentary layer at Klissoura Cave 1 is XXI, which is only found toward the eastern wall of the shelter (Sitlivy, et al. 2007). There is a large disconformity between this layer and the overlying Middle Paleolithic deposits, suggesting a major environmental event at this time. The faunal component from XXI is also different from overlying layers 97 and is composed primarily of microfauna and small bird remains. Because the fauna was determined during analysis to be non-anthropogenic, it is not included in this study. The overlying layers include XX through XVI (Table 2.1). Sediments in these layers are finegrained, with a high degree of post-depositional chemical alteration, especially in layer XX (Sitlivy, et al. 2007:4). The lower deposits slope toward the entrance of the site, and the authors note that layer XVIII is especially rich in lithic and bone artifacts. Sitlivy et al. (2007) describe a difference between the stratigraphically higher (VII-XIV) and lower (XVXX) Middle Paleolithic sediments: the higher layers are composed of stone-rich, reddish clay, and the lower layers contain fine-grained grey silty and ashy sediments. The authors believe that the difference between the upper and lower Middle Paleolithic layers is explained by a collapse of the back chamber of the cave during the formation of layer XV. This collapse created a chimney that allowed rainwater, terra rossa sediments, and roof fall to wash into the cave (Sitlivy, et al. 2007:4). The stratigraphically higher Middle Paleolithic layers (VII-XV) are composed of brownish-red fine-grained sediments with stony deposits at the back of the shelter, indicating a higher input of clay and clastic materials (Sitlivy, et al. 2007:3). Layers XI-XIV contain alternating clay-rich layers and thin, anthropogenic burned layers. There is an erosional depression at the front of the shelter that truncates layers XI and XII (Sitlivy, et al. 2007:3). Layer IX contains very few faunal remains, but this layer along with layer X contain burned remains cemented by surface water. An erosional event formed a trough in layer X that was subsequently filled by either natural sediments or anthropogenic materials (Sitlivy, et al. 2007:3-4). Layers VI through VIII contain clay-rich sediments and an increasing proportion of anthropogenic inputs, including burned lenses. It was determined by the analysis of the 98 lithic and ornament assemblages that layers VI and VII are mixed in some parts of the site, containing both Middle and Upper Paleolithic materials (Sitlivy, et al. 2007; Stiner 2010; Stiner, et al. 2010); consequently, layers VI and VII were excluded from the faunal analysis. Based on geoarchaeological and micromorphological analyses, Karkanas (2010) divides the Upper Paleolithic and later cultural layers in Klissoura Cave 1 into depositional facies that are grouped into a stratigraphic sequence. An in-depth description of the facies is presented in Karkanas (2010). In general, the Upper Paleolithic deposits are defined by an extremely high anthropogenic input. Anthropogenic inputs decrease somewhat in the Epigravettian (IIa-d) and Mesolithic (3-5a) layers, which are dominated by clastic sediments (Karkanas 2010:32). The Early Upper Paleolithic or Uluzzian layer V is significant because it is one of few EUP deposits found in Greece. It is composed of dark gray clayey silt with reworked and in situ burned features in some areas. Although layer V is diffuse in some parts of the excavation, it is intact in the middle and southern parts of the site (Karkanas 2010:30). The deposits in layer V represent a similar sedimentary environment as those that formed layer IV, though based on recent dating evidence discussed above, there is reason to suspect that a hiatus of several thousand years separates the formation of the two layers (Kuhn, et al. 2010). Layer IV, the lower Aurignacian, contains many burned features and deposits, including numerous clay-lined hearths (see below) (Karkanas 2010; Karkanas, et al. 2004; Pawlikowski, et al. 2000). The sediments are gravelly silts and are light in color due to a high wood ash content. Clay-lined hearth structures were not found in the southern units of the excavation, near the entrance of the shelter; rather the sediments in this area are dominated 99 by homogenous reddish-brown clay (Karkanas 2010:30). A linear or clustered arrangement of large stones representing a possible structure are found in this area. The layer III series is filled with burned deposits from reworked hearths, intermingled with undisturbed deposits. This layer also contains clay-lined hearth structures, along with some gravel-rich clusters (Karkanas 2010:30). Many of the stratigraphic layers included in this sequence are horizontally discontinuous, and were not necessarily deposited simultaneously. Karkanas (2010) does note, however, that units within the layer III series can be grouped into a coherent stratigraphic order: IIIe-g are the oldest and are overlain in succession by IIIb-d, III", III', and finally III. This is relevant to the discussion of stone tool industries discussed below, as the bottom-most two layer groups (IIIe-g and IIIb-d) are Aurignacian deposits, and III’ and III” are assigned to distinct non-Aurignacian Mediterranean blade or backed bladelet industries (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001). The contact between layer III’ and the Epigravettian layers is locally diffuse. Some freeze-thaw action is apparent in certain parts of layer III’ but not in any other layer of Klissoura Cave 1 (Karkanas 2010:34). An anthropogenic ditch or intrusion of mixed deposits (6-7a) cuts into layers III’ and III” in certain parts of the site. This layer is often referred to in earlier publications (e.g., Karkanas, et al. 2004; Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001; Starkovich 2009) as part of the Upper Paleolithic sequence, but it is now known to be mixed based on the lithic results and is excluded from this study. The Epigravettian layers (IIa-d) are composed mostly of brown, clay-rich deposits, though in certain areas there are reworked burned remains and sorted, laminated sediments (Karkanas 2010:28). Clusters and lines of limestone rocks representing deflated surfaces (i.e. 100 hiatuses) are common. Some bioturbation from insects was noted in the Epigravettian layers (Karkanas 2010:34). The youngest layers include the modern surface and disturbed Bronze Age deposits (1-2), through the partly intact Mesolithic deposits (3-5a). The uppermost parts of these layers have numerous recent pits and evidence of animal trampling. The Mesolithic layers (35a) are composed of clay-rich deposits with lenses of loose ash and shell and laminated geogenic layers (Karkanas 2010:28). Lithic Assemblages Tens of thousands of stone tools or lithic fragments were recovered from the Middle Paleolithic at Klissoura Cave 1, and over 130,000 from the Upper Paleolithic and later layers (Kaczanowska, et al. 2010; Sitlivy, et al. 2007). Most of the assemblages are composed of chips or debitage, followed by flakes and tools (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001; Sitlivy, et al. 2007). The site does not preserve an uninterrupted sequence during the span of occupation (Kaczanowska, et al. 2010); rather there are several hiatuses in the sequence, as discussed previously. The Middle Paleolithic lithic assemblages from Klissoura Cave 1 are still under analysis, but preliminary results from a large sample (n = 37,922) are presented in Sitlivy et al. (2007, see also Koumouzelis, Kozlowski et al. 2001). Proportions of raw material types are similar throughout the sequence, with a dominance of radiolarite followed by flint. In terms of rare raw material types, limestone and quartz are more common in the lowermost and middle part of the MP sequence, and chalcedony occurs in the uppermost layers (Sitlivy, et al. 2007:6). 101 The authors find many other similarities in the Middle Paleolithic lithic assemblages of Klissoura Cave 1, and a high degree of homogeneity throughout. In general, Middle Paleolithic tools and debitage fragments are quite small in all layers, though blanks are slightly larger in the lowest units (Sitlivy, et al. 2007:10). The earliest phases of raw material testing and removal of the cortex are absent at the site, suggesting that the occupants brought pre-tested raw materials to the cave for tool production (Sitlivy, et al. 2007:9). A variety of reduction strategies are found in all layers, including unidirectional or discoidal and centripetal methods, and the frequency of Levallois technologies is low throughout (Koumouzelis, Kozlowski, et al. 2001; Sitlivy, et al. 2007). The tool assemblages are primarily Mousterian, though some Upper Paleolithic elements are apparent in the younger layers, indicating some mixing in certain parts of the site. Some subtle differences are also apparent among the Klissoura Cave 1 Middle Paleolithic lithic assemblages. The proportion of blades and bladelets is much higher in the lowest layers (XVIII, XIX and XX) (Sitlivy, et al. 2007:8). The authors note that volumetric production methods typically associated with Upper Paleolithic industries were used to make blades in the lower layers, along with other reduction strategies, though Upper Paleolithic tools are not present in the older Mousterian layers (Sitlivy, et al. 2007:12). By far the most common tools in the Middle Paleolithic assemblages are side scrapers, though their frequencies vary slightly with time. They occur in the smallest numbers in the latest (VII and VIIa) and earliest (XVIII and XIX) Middle Paleolithic layers, and they peak in the middle of the sequence (XIII) (Sitlivy, et al. 2007:11). Mousterian points are documented in all layers except VIIa, and they are the most common in the lower layers (XVIII and XIX). Two different styles of points are recorded, elongated points and short massive types. Most of the 102 variability in point form is found in the lowest layers (XVIII, XIX and XX) (Sitlivy, et al. 2007:12). Upper Paleolithic tool types, including endscrapers, burins, perforators and splintered pieces are found in the youngest Middle Paleolithic layers (VII, VIIa and VIII), and completely disappear by layer X (Sitlivy, et al. 2007:12). Although the contents of layers VII, and especially VIII, are essentially Middle Paleolithic, the possibility of low levels of mixing from the early Upper Paleolithic cannot be excluded. Sitlvy et al. (2007) conclude that Klissoura Cave 1 is different from other Greek Middle Paleolithic sequences. Specifically, the authors note that there is no “microMousterian” overlying Levallois/Mousterian layers such as that found at Asprochaliko (Bailey, et al. 1983). Also, though there are small numbers of bifacial tools in the lowest MP layers, these are absent in the later layers at Klissoura Cave 1, suggesting a different evolution of stone tool technologies in southern and northern Greece (Sitlivy, et al. 2007:14). The Early Upper Paleolithic layer (V) in Klissoura Cave 1 is similar to the EUP industries of Italy and has thus been defined as Uluzzian (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001). This is significant as it is the first published Uluzzian occupation in Greece and represents a rare and exceptionally old Early Upper Paleolithic industry in the country. The sample is small compared to the later layers (n = 4,228). Flakes are common in the assemblage, as are blades and tools. Radiolarite is the most common raw material, followed by local flint (Kaczanowska, et al. 2010). Diagnostic artifact types in this layer include arched backed blades and convex truncations along with splintered pieces; endscrapers, retouched blades, and notched denticulated tools also occur but are scarce (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001). 103 A thick series of Aurignacian deposits was excavated at the site, with assemblages defined by micro-retouched bladelets and carenate cores or endscrapers (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001). Based only on their stratigraphic position, they are divided by Karkanas (2010) into the lower (IV), middle (IIIe-g) and upper (IIIb-d) Aurignacian units (though Kaczanowska, et al. 2010 divide the latter two as IIId-g and IIIa-c; this is not problematic in terms of the fauna as almost all analyzed material came from IIIc, IIIe, or IIIg). In the Aurignacian series, as with the EUP, the most common raw material type is radiolarite, followed by flint. In general, the composition of tools types in the series is similar, with end scrapers being the most common, followed by denticulated-notched pieces, retouched flakes, burins, side scrapers, and blades/retouched bladelets, among others (Kaczanowska, et al. 2010). The bulk of the Klissoura Cave 1 lithics are found in the lower Aurignacian layer IV (n = 63,837), a quarter of which were recovered from within a rocklined structure (see below) (Kaczanowska, et al. 2010). About half of the recorded Upper Paleolithic hearths were also found in this layer (n = 55). Aurignacian layer IIId-g contains a large sample as well (n = 28,625), whereas layer IIIa-c yielded the smallest sample of the series (n = 5,623). Thirteen lumps of ochre were recovered in the middle Aurignacian layers, and two from the upper (Kaczanowska, et al. 2010), though their sizes are not mentioned. Above the Aurignacian layers there are two technically distinct but unnamed Mediterranean Upper Paleolithic backed bladelet or blade industries. The older of the two layers (III”) is referred to in this presentation as the non-Aurignacian Upper Paleolithic, and the younger (III’) is called the Mediterranean backed bladelet industry. The non-Aurignacian Upper Paleolithic is a relatively small lithic assemblage (n = 2,935). Flakes are much more common than blades in this layer, and tools include endscrapers, denticulates and notched 104 tools, retouched flakes and side-scrapers (Kaczanowska, et al. 2010). Layer III” is different from the underlying Aurignacian because of the importance of flakes in blank production, few cores for bladelet production, and few carenate cores or endscrapers. The Mediterranean backed bladelet industry layer III’ is distinct because of the importance of backed bladelets, microretouched bladelets, and truncations (Kaczanowska, et al. 2010). The lithic sample from this layer is moderately sized (n = 5,943), and most of the pieces are radiolarite followed by flint. The most common large tool type is again endscrapers, though backed bladelets and microliths occur in much higher frequencies in layer III’ than in other layers at the site, followed by retouched flakes (Kaczanowska, et al. 2010). Kaczanowska et al. (2010) suggest that this may represent one of the oldest backed bladelet industries in the Mediterranean Basin, and they compare the tools in layer III’ to those found in the Asprochaliko sequence. After a stratigraphic hiatus representing the LGM, there is a comparatively limited Epigravettian deposit (IIa-d, n = 6,312) containing an industry that is defined by backed blades with ventral retouch, para-geometric forms and shouldered points (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001). Most of the lithic materials are made of radiolarite, followed by flint. Backed tools, microlithics and truncations are common, followed by endscrapers, and blades with ventral retouch. The most recent major stratigraphic layer at Klissoura Cave 1 is a Mesolithic (3-5a) deposit with 6,812 lithics (Kaczanowska, et al. 2010). As with the earlier layers, radiolarite is common followed by flint, though two obsidian artifacts are also found (Kaczanowska, et al. 2010; Koumouzelis, et al. 2003). Given that obsidian from the island of Melos appears during the early Mesolithic at Franchthi Cave, it is possible, but not certain, that the obsidian 105 from Klissoura Cave 1 is also from Melos. In the Mesolithic layers there is a much lower proportion of chips and lithic debris as compared to tools, so it seems likely that many tools found in the assemblage was manufactured off-site (Kaczanowska, et al. 2010). Tools include endscrapers, side scrapers, retouched flakes and blades, microlithis, and geometric forms, among types rarer in the assemblage. The authors note that the Klissoura Cave 1 Mesolithic is unlike the industries found at Franchthi Cave (Kaczanowska, et al. 2010; Koumouzelis, et al. 2003). Botanical Remains Phytoliths and macrobotanical remains were studied in the Upper Paleolithic and later layers of Klissoura Cave 1; analyses of the Middle Paleolithic layers are ongoing. The contribution of these studies to local paleoenvironmental reconstructions are discussed in Chapter 3. Phytolith samples were collected from the Aurignacian and later layers, from the tops of clay-lined hearth structures (see below) and other hearths, and from the surrounding matrix (Albert 2010). Importantly, few phytoliths were found in hearth samples in any of the layers. Albert (2010) suggests that this finding indicates that the leaves and stems of monocotyledonous taxa and wood or bark were the most common fuel sources, as wood and bark contain far fewer phytoliths than grasses, particularly in certain Mediterranean species (Albert and Weiner 2001). Phytoliths from grasses with edible seeds were found outside of hearth areas in the Aurignacian series, as well as in the two layers that contain backed bladelet industries (III’ and III”). Sedges, which are high in starch and protein content and also useful materials for fiber-working, were found especially within the layer III series, but 106 also in the lower Aurignacian (IV) and Epigravettian (IIa-d) (Albert 2010). No phytolith samples are available for the Mesolithic (3-5a). Water-flotation was undertaken on samples from hearths and sediments from selected excavation units in order to collect wood charcoal (see Ntinou 2010 for a full description of methodology). In general, large pieces wood charcoal are rare in hearths, occurring in only twenty-nine of over a hundred of hearths, though microscopic charcoal is found throughout the layers in all parts of the site (Karkanas 2010). Further, there were no higher incidences of charcoal in clay-lined or stone-delineated hearth structures. Ntinou (2010) postulates that the lack of macroscopic charcoal is due to the intense use of hearths, with complete combustion of plant materials or cleaning events after each hearth was used. The most common plant taxa used as fuel at the site include maple, deciduous oak, and Prunus sp., and deciduous oak. The proportions of these fuel sources vary among hearths within and between layers (Ntinou 2010). Features Among the most unique features at Klissoura Cave 1 are the dozens of clay-lined hearths in the Aurignacian and overlying Upper Paleolithic deposits (Karkanas, et al. 2004; Kot 2009; Pawlikowski, et al. 2000). These features represent some of the earliest recorded uses of modeled clay in the world (Pawlikowski, et al. 2000). Over a hundred well-preserved hearths were encountered in the archaeological sequence of Klissoura Cave 1, in all layers. However, over fifty clay-lined hearths are confined to the layers that contain the nonAurignacian backed bladelet industries (III’ and III”) and the underlying Aurignacian series (IIIb-d, IIIe-g, IV) (Karkanas, et al. 2004; Pawlikowski, et al. 2000). The earliest clay-lined 107 features at Klissoura Cave 1 date to between 35,000 and 37,500 BP, which predates the Gravettian clay figurines from Dolni Vestonice (Vandiver, et al. 1989) by over five thousand years. During excavation, these orange, basin-shaped features were easily recognized and are obvious in profile (see images in Karkanas 2010; Karkanas, et al. 2004). Later clay-lined hearths are often superimposed on earlier ones. Most of the clay-lined and unlined hearths range between 30 and 40 cm in diameter, and a few of the clay-lined features have rims (Karkanas, et al. 2004:515). Generally, the clay-lined features are covered by ashy lenses filled with charcoal fragments, lumps of soil, and burned lithics and bone. Analyses of micromorphology samples from the hearths and from an incidental mudflow in the cave indicate that the hearths did not form in-situ via a natural process (Karkanas, et al. 2004:518). Rather, the clay was brought to the site, probably from the floodplain in the base of the gorge, and then wetted and pooled into the basin-like shape. Fourier transform infra-red spectrometry (FTIR) and differential thermal analysis (DTA) performed on samples from the clay-lined hearths indicates that they were heated between 400 and 600 °C (Karkanas, et al. 2004, but see also Pawlikowski et al. 2000, who suggest that at least one hearth was fired above this temperature based on the degradation of the mineral kaolinite). It is not known if the clay hearths were hardened with fire in anticipation of their first use or instead hardened as the result of use. Kot (2009) presents the results of experimental studies on the formation of the claylined hearth structures at Klissoura Cave 1 and proposes five possible uses for the features: long-term ember storage, exposed cooking or roasting meat and plant foods, water boiling, surfaces for holding food during burial with hot ashes or embers, or easily cleaned surfaces for reuse as hearths (see also Meignen, et al. 2001). Drying or curing meat and skin are also 108 possibilities for clay-lined hearth use (Ntinou 2010). Taking the time to create such structures indicates longer term or more intensive site use, which is attested to by the influx of artifacts and anthropogenic sediments into the site during these periods of occupation. Karkanas et al. (2004) suggest that the small clay-lined structures were satellite hearths around a larger flat (simple) hearth. They also note that starch and seed grass phytoliths were found in a sample taken from the basin of one of the clay hearths, so it is possible that they were used on some occasions for cooking or parching plant materials. In addition to the clay-lined structures, five hearths are rock-lined and four are associated with stones (Ntinou 2010). Nearly all of these structures are found in the lower Aurignacian (IV), though one is from the Early Upper Paleolithic (V), and one other from the middle Aurignacian (IIIe-g). Another interesting feature, a circular man-made rock alignment, was noted in the lower Aurignacian (IV) (Koumouzelis, Ginter, et al. 2001; Stiner 2010; Stiner, et al. 2010). This structure is apparent in plan maps (see Chapter 7) and is defined by a large circle of limestone rocks, dark staining inside the circle, a lack of hearths, and a concentration of ornaments and lithic debris within the structure. The structure takes up four excavation units, giving it a maximum diameter of about 1.5-2 meters (Koumouzelis, Ginter, et al. 2001). Ornaments Shell ornaments are common in the Upper Paleolithic Klissoura 1 sequence, with over 1,500 recovered (Stiner 2010). Ornaments are found in all UP and later layers, though their frequencies change dramatically. The majority of the ornaments are from the lower Aurignacian (IV), followed by the middle Aurignacian (IIIe-g). Ornament frequencies are quite low in the Epigravettian and Mesolithic layers, partially due to small sample sizes for 109 all cultural materials (Stiner 2010). The author suggests that declines in mollusk species richness after the middle Aurignacian (IIIe-g) likely corresponds to the expansion of lagoon formations in the western Argolid. Similar to other Paleolithic sites in Greece, Cyclope sp. is the most common species in the assemblages overall, followed by Columbella rustica. Stiner (2010) notes that the sample size for the EUP layer is small, but that Dentalium is particularly common, not unlike that situation in some Italian Uluzzian sites (e.g., Palma di Cesnola 1966). Foragers occupying Klissoura Cave 1 collected mainly vacant shells from the marine littoral, along with exploited fossil deposits and some freshwater sources. They disproportionately selected the shells of predatory and scavenging carnivorous taxa for ornaments, possibly because of their scarcity or aesthetic value, a phenomenon that is found elsewhere in the Mediterranean (Stiner 2003a, 2010). Ornaments at the site display a narrow size range: 1.3-1.4 cm in the Upper Paleolithic, and 1.4-1.7 cm in the later layers. Naturally red Clanculus and striped Theodoxus shells were enhanced with red ochre (Stiner 2010). Unlike the situation at Franchthi Cave, burning frequencies on ornaments at Klissoura Cave 1 are probably from random, accidental burning. Cord wear is documented on many of the ornaments, indicating they arrived at the site already strung as beads or fastened to organic materials. A final observation about the assemblage from the lower Aurignacian (IV) is that the ornaments seem to cluster in the squares delineated by the possible rock-lined structure discussed above. Stiner (2010) suggests that many or most of the ornaments may have been attached to an organic hide or mat, which resulted in dark staining of the sediments observed during excavation. 110 The presence of the rock-lined structure, numerous hearths, and the nature of the lithic and faunal remains indicates that a range of domestic activities occurred at the site. This implies that Klissoura Cave 1 was probably a base camp or short-term residential site during much of its history. SUMMARY AND GOALS Considering all of the lines of evidence discussed above, Klissoura Cave 1 is an extremely important site for understanding behavioral change and shifting foraging strategies over the course of the Middle Paleolithic through Mesolithic in southern Greece. The length of the sequence, with the exception of a few hiatuses, and intact stratigraphy fills many gaps in our understanding of the Greek Paleolithic, in particular the Early Upper Paleolithic (Uluzzian layer V) and Aurignacian. In addition, advanced excavation methods and specialized analyses (specifically macrobotanical remains, phytolith analyses, and micromorphology) provide many avenues of understanding not available to previously excavated sites. The goals of the larger Klissoura Cave 1 project are to better understand the Greek Paleolithic through Mesolithic as a whole, examining how stone tool types, use of ornaments as symbols, and hunting patterns fit with other sites in Greece, and how the site fills gaps in our understanding of Greece before the Neolithic. Additionally, though it is less than a day’s walk from the Gulf of Argos, it is the only well-published inland site on Peloponnese. Thus, intensification and dietary expansion that is recorded at the end of the Upper Paleolithic at other sites (e.g., Franchthi Cave) may take on a different dimension at Klissoura Cave 1 in the absence of marine resources. 111 The well-preserved faunal remains from the Middle Paleolithic through Mesolithic layers of Klissoura Cave 1 are the focus of this dissertation. In the coming chapters, I examine dietary change in the Klissoura Cave 1 sequence from the perspective of animal exploitation, using zooarchaeological and taphonomic methods and drawing on concepts of evolutionary ecology discussed in Chapter 1. Changes in environmental conditions are evaluated (Chapter 3) before examining shifts in prey choice, to distinguish as best as possible between natural and human processes affecting prey availability. Evidence for nonhuman taphonomic factors in the assemblages is examined (Chapter 4) in addition to questions about human decisions surrounding the partitioning, transport and butchery of large and small-bodied prey (Chapters 5 and 6). Spatial variation in modified faunal remains is also considered within the layers that contain large samples (Chapter 7). The coming chapters are meant to highlight the uniqueness of Klissoura Cave 1 as a fully terrestrial site on one of three major southern European peninsulas, while considering the locality as part of a larger Paleolithic landscape within Greece. Ultimately, the goals of this dissertation are to understand diachronic trends at a site that spans much of the late Pleistocene and early Holocene, and use this sequence to augment gaps in our understanding of the evolution of human subsistence strategies during the Paleolithic in this part of the Mediterranean Basin. 112 CHAPTER 3: LATE PLEISTOCENE CLIMATE AND ENVIRONMENT IN THE EASTERN MEDITERRANEAN INTRODUCTION Climatic variability was one of the key pressures acting on human groups worldwide during the late Pleistocene. Changing climates led to shifts in environmental regimes, which affected the availability of various plant and animal species as biotic communities expanded and contracted regionally. Clearly, human groups dealt with such change successfully overall as populations continued to grow during the Middle and Upper Paleolithic, though there were declines and extinctions on a local level. This variability would have provided an environmentally unpredictable backdrop for Neandertal populations in Europe and western Asia, and for the movement of anatomically modern humans into Europe from the Middle East after 50,000 years ago (Huntley and Allen 2003). The occupation of Klissoura Cave 1 spans much of the late Pleistocene, from marine oxygen isotope stage (MIS) 5a through MIS 1 (about 80,000 until 10,000 BP, see Table 2.1) (Karkanas, personal communication 4/23/10; Kuhn, et al. 2010). This period is marked by a high degree of climatic variability with oscillations between glacial and interglacial conditions, as well as smaller stadials and interstadials (Martinson, et al. 1987). The everchanging climate of this time period witnessed a rapid turnover of plant communities, ranging from open steppe environments during the glacial periods, to an expansion of forests during interglacials (Huntley and Allen 2003; Macklin, et al. 2002; Van Andel and Tzedakis 1996), which in turn affected the composition of animal communities. Such changes are 113 often reflected by the prey spectra in archaeological sites that correspond with the availability of local resources (Huntley and Allen 2003). A central goal of this chapter is to tease out possible evidence for human population growth and behavior from environmental factors based on data on the hunting patterns of Pleistocene hominins. As such, some reconstruction of the local and regional environment is crucial. In this chapter, larger-scale climatic changes that affected the Mediterranean Basin from MIS stage 5a through 1 are discussed, followed by more localized changes that occurred in Greece, specifically around Klissoura Cave 1 for the latter part of the occupation of the site. The general availability of animal resources is predicted in light of changing plant communities in the area. The environmental reconstructions are based almost exclusively on sediment cores taken from lakes or the Mediterranean Sea, qualified by data on the macrobotanical remains from Franchthi Cave (Hansen 1978, 1991) in the southern Argolid and Klissoura Cave 1 (Albert 2010; Karkanas 2010; Ntinou 2010). In general, nonarchaeological contexts were preferred for environmental reconstructions because they offer a more complete record without an anthropogenic signature. However, the data from Klissoura Cave 1 are included because they provide the most local reconstructions available, and the proximity of Franchthi Cave to Klissoura Cave 1 (about 40 km) justifies the inclusion of Hansen’s work on plant communities and exploitation. All radiocarbon dates presented here are calibrated, unless otherwise noted. ENVIRONMENTAL HISTORY OF THE MEDITERRANEAN: MIS 5a THROUGH 1 Climatic fluctuations in the northern hemisphere during the Pleistocene are primarily understood from long marine cores from the Indian Ocean (Martinson, et al. 1987) and North 114 Atlantic (Bond, et al. 1993; Bond, et al. 1992; Bond and Lotti 1995), and ice cores from Greenland (Dansgaard, et al. 1993; Grootes, et al. 1993). The use of these records as a baseline for other Pleistocene environmental reconstructions is fairly standard (see below). One of the most significant contributions to come out of this research are the marine oxygen isotope stages (MIS) that chronicle changing ocean temperatures (Martinson, et al. 1987), which in turn reflect larger atmospheric changes. Martinson et al. (1987:1) provide a synthesis of the MIS dating scheme by combining radiocarbon and uranium-thorium dates with stacked oxygen-isotope chronologies and four different “orbital tuning” approaches, where they correlate climatic events with orbital forcing. They place the following date ranges for the MIS stages: 1 (0 – 12,050 BP), 2 (12,050 – 24,110 BP), 3 (24,110 – 58,960 BP), 4 (58,960 – 73,910), 5 (73,910 – 79,250 BP, which marks the beginning of MIS 5a) (Martinson, et al. 1987:20). Figure 3.1 is adapted from Martinson et al. (1987:19) and shows the fluctuations of δ18O values; positive values indicate generally warmer ocean conditions, negative values generally cooler conditions. Aside from the major fluctuations, smaller events are also noted and dated (Martinson, et al. 1987:20). The case will be made below that the larger-scale fluctuations are an appropriate baseline for analyzing climatic change in the Mediterranean Basin, though the smaller fluctuations are less useful for understanding local climatic shifts. Establishing a link between marine (Bond, et al. 1993; Bond, et al. 1992; Bond and Lotti 1995) and the classic Greenland ice core records (Dansgaard, et al. 1993; Grootes, et al. 1993) is important in understanding global climate shifts during the Pleistocene. These 115 Figure 3.1 Marine oxygen isotope chronology, adapted from Martinson et al. (1987). Bar at the top indicates period that were warm (w), cold (c) and mild (m), compared to surrounding periods. studies typically record fluctuations on the millennial scale; as will be discussed below, higher resolution is often found in local climate records. Ice cores from Greenland indicate that climatic instability over the course of the last hundred thousand years was extreme, with the exception of the comparatively stable Holocene (Dansgaard, et al. 1993:220; Grootes, et al. 1993:553). In the ice cores, a series of temperature events, called Dansgaard-Oeschger (DO) oscillations were identified (Dansgaard, et al. 1993). Bond et al. (1992:247) examined marine cores from the North Atlantic and identified a series of six so-called Heinrich events over the course of the last 70,000 years. These events are marked by drops in sea surface 116 temperature and salinity, a decrease in the flux of planktonic foraminifera, and a discharge of icebergs (Bond, et al. 1992:248). Bond et al. (1993:146) establish a link between the D-O oscillations and Heinrich events, which punctuate the end of D-O oscillations between 20,000 and 80,000 BP. This confirms the relationship between ice sheet melting and changing atmospheric and ocean temperatures (Bond, et al. 1993:145), though some additional ice rafting cycles seem to appear outside of the D-O oscillations (Bond and Lotti 1995). A few Mediterranean-wide climatic studies are published for the late Pleistocene. Considering several lines of evidence, including ice cover, sea-level, and pollen data from ice, marine and pollen cores, Van Andel and Tzedakis (1996) present vegetation maps for several different stages of the late Pleistocene. They determine that MIS 4 was cold, with an expansion of open vegetation, whereas MIS 3 was variable with many climatic oscillations (Van Andel and Tzedakis 1996). Macklin et al. (2002) examine aggradation in alluvial systems in Greece, Libya and Spain. They determine that alluviation occurred when steppe vegetation expanded during cool, dry stadials (Macklin, et al. 2002:1634). They found that MIS 4 was quite cold, with aggradation at sites in their study (Macklin, et al. 2002:1638). Conversely, MIS stages 3 and 2 are marked by multiple periods of alluviation, indicating a high degree of climatic variability (Macklin, et al. 2002:1639). Both of these studies offer fairly coarse reconstructions, though they correspond well with broader trends in late Pleistocene climates. Harrison and Digerfeldt (1993) compiled lake-level data for the Mediterranean Basin since the last glacial maximum (LGM), with the assumption that high lake levels are indicative of wetter climatic conditions, at least on a regional basis. They find that the last high stand of lakes throughout the region occurred at the LGM (Harrison and 117 Digerfeldt 1993:241). Subsequent drying began first in the western Mediterranean at about 16,000-15,000 BP (uncal) and later in the eastern Mediterranean between 13,000 and 11,000 BP (uncal). Conditions became wetter by 12,000 BP (uncal) in the west and gradually ameliorated in the east until about 6,000 BP (uncal) (Harrison and Digerfeldt 1993:241). Many studies in the Mediterranean Basin have attempted to link local signatures with the marine and ice core sequences from the Northern Hemisphere. A 48,000 year core from the Alborán Sea south of Spain was analyzed and correlated to other Iberian sequences, as well as North Atlantic ice cores (Cacho, et al. 1999; Fletcher and Sánchez Goñi 2008; Sánchez Goñi, et al. 2002). Heinrich events 1-5 are evidenced in the sequence by drops in sea surface temperatures and fluctuations in planktonic foraminifera, similar to those found in the North Atlantic (Cacho, et al. 1999:703). Dansgaard-Oeschger oscillations are also apparent in the Spanish record, though they are less intense than the observed Heinrich events (Cacho, et al. 1999:703). Sanchez Goni and colleagues (Fletcher and Sánchez Goñi 2008; Sánchez Goñi, et al. 2002:104) correlate vegetation changes and sea surface temperatures from the Alborán core and determine that D-O stadials were dry and cold, while D-O interstadials were humid and mild. Heinrich events were the driest and coldest periods. They then relate the Alborán record to an Atlantic sequence off the coast of Portugal, and find broad agreement between the two cores in the timing of events, though differences are apparent in the amount of moisture or temperature gradients at the two sites during various Heinrich events (Sánchez Goñi, et al. 2002:103). In Italy, the 102,000 year sequence from Lago Grande di Monticchio was correlated with the marine sequence, as well as the Greenland ice cores (Allen, et al. 1999; Allen and Huntley 2000; Allen, et al. 2000; Watts, et al. 1996). The Monticchio record matches the 118 marine oxygen isotope stages well, with a difference of less than 2,000 years for the major boundaries between stages (Allen and Huntley 2000:124; Allen, et al. 2000:106). The sequence also records many fluctuations found in the North Atlantic, including D-O oscillations during MIS 2 and 3 (Watts, et al. 1996:150) and Heinrich events 1-6, though there are some variations. Allen and colleagues (Allen, et al. 1999:742; Allen and Huntley 2000:124) find a divergence between the Monticchio and North Atlantic records before 65,000 BP, with the former displaying a more detailed record with additional fluctuations, and Watts et al. (1996:151) find that Heinrich events 1-6 in the sequence manifest slightly later than is indicated in the ice cores. Additionally, some seasonal trends were apparent during interstadials in the high-resolution Italian record that do not appear in the ice core records (Allen, et al. 2000:108). The Monticchio record also correlates well with other pollen cores from southern Europe, including Padul (Spain), Lac du Bouchet (France), Valle di Castiglione (Italy) and Tenaghi Philippon (Greece), though the Monticchio core has higher resolution and displays some fluctuations not apparent in the other sequences (Allen, et al. 2000:124). At Soreq Cave in Israel, changes in the proportion of δ18O and δ13C in speleothems reflect climatic fluctuations over the last 60,000 years (Bar-Matthews, et al. 1997; BarMatthews, et al. 1999). The authors find cold peaks at 46,000, 25,000, 19,000, 16,600 and between 13,200 and 11,400 BP, which correspond to Heinrich event 5 (H5), H2, the LGM, H1 and the Younger Dryas, respectively (Bar-Matthews, et al. 1999:89; Bond, et al. 1992). Heinrich events 3 and 4 are not represented in the Soreq Cave sequence, and there is a cold event recorded at 35,000 BP that is not found in the Greenland ice core record (Bar- 119 Matthews, et al. 1999:89; Bond, et al. 1992). Warm peaks were found at 54,000 and 36,000 BP. Climatic reconstructions from the Mediterranean Basin tend to compare quite well with worldwide MIS shifts (Allen and Huntley 2000; Allen, et al. 2000; Cacho, et al. 1999; Fletcher and Sánchez Goñi 2008; Sánchez Goñi, et al. 2002). In some cases, individual Heinrich events and D-O oscillations are also identified (Allen, et al. 1999; Allen and Huntley 2000; Bar-Matthews, et al. 1997; Bar-Matthews, et al. 1999; Cacho, et al. 1999; Fletcher and Sánchez Goñi 2008; Sánchez Goñi, et al. 2002; Watts, et al. 1996). However, many of these events manifest themselves differently in terms of length or severity in the local records, and in many cases certain Heinrich events are entirely absent from Mediterranean cores (Allen, et al. 1999; Allen and Huntley 2000; Bar-Matthews, et al. 1997; Bar-Matthews, et al. 1999; Sánchez Goñi, et al. 2002; Watts, et al. 1996). For this reason, the general character (warm, mild, cold) of the different MIS stages from 5a onward are included on the axes of many figures in this dissertation. Though the larger curve fits quite well with the Greek data discussed below, it must be noted that glacial conditions of MIS 4 may have begun slightly later and MIS stages 2 and 1 began slightly earlier in Greece than in the Indian Ocean record (Martinson, et al. 1987, see also Figure 3.2). On the other hand, smaller Heinrich events and D-O oscillations seem to take on a local character across the Mediterranean, including Greece (see below). Therefore, this presentation uses the MIS shifts as a general guide for Mediterranean-wide and Greek climatic changes, but local environmental studies are used to establish finer-scale shifts. 120 Figure 3.2 Simplified schematic of Greek paleoclimate studies mentioned in the text. Dark bars = warm, wet (forest expansion), grey bars = mild (mixed forest-steppe), white bars = dry, open steppe. Core locations from studies mentioned in this figure are mapped in figure 3.3. Bottom axis corresponds to MIS stages in Figure 3.1. 121 ENVIRONMENTAL HISTORY OF GREECE: MIS 5a THROUGH 1 Northern and Central Greece Climatic changes in Greece since MIS 5a generally mirror those from the larger Mediterranean Basin. In some cases, certain climatic events are expressed differently in the eastern Mediterranean than to the west (see above). Many environmental reconstructions are available for Greece, particularly in the northern and central part of the country (see Figure 3.3 for core locations mentioned in this chapter). Of these studies, most focus on the LGM and later (Kotthoff, et al. 2008; Lawson, et al. 2005; Lawson, et al. 2004; Okuda, et al. 1999; Turner and Sánchez Goñi 1997; Tzedakis 1993, 1994, 1999; Tzedakis, et al. 2004; Tzedakis, et al. 2002; Wijmstra 1969; Willis 1997), but some longer cores were analyzed as well (Tzedakis 1993, 1994, 1999; Tzedakis, et al. 2004; Tzedakis, et al. 2002; Wijmstra 1969). A synthesis of all of the pollen data from Greece, some taken from pollen cores, others from archaeological sites, is presented in Figure 3.2. This figure indicates periods of expanded forest communities with increased moisture or changes in the seasonal availability of moisture, intermediate conditions with mixed forest-steppe environments, and dry periods of open steppe vegetation. Changes in the vegetation regimes will be used to interpret the availability of herbivore species below. Tzedakis et al. (2002) present pollen data from a core taken from the Ioannina Basin (core I-284). In general, the area was likely a refugium for forest communities throughout the Pleistocene, but contractions and expansions of these forests correspond fairly well with environmental changes found elsewhere in Greece. During MIS 5a and into MIS 4 (between 83,000 and 68,000 BP), forest communities were widespread in the area (Tzedakis, et al. 2002:2044). An earlier long core analyzed by Tzedakis (1993:437; 1994:414), also from the 122 Figure 3.3 Map of Greece indicating locations of pollen cores discussed in the text, and in Figure 3.2. Ioannina Basin (core I-249), indicates that MIS 5a was marked by Quercus (oak) forests with Ulmus (elm) and Carpinus (hornbeam). These conditions are also found in central Greece, where a pollen core from Lake Kopais (core K93) has a high incidence of arboreal pollen in MIS 5a and into MIS 4, from 84,000 to 74,000 BP (Tzedakis 1999:428). A core from the 123 Tenaghi Philippon in northeastern Greece also found expanded Quercus ilex (holm oak) and Pinus cf. nigra (European black pine) forests during MIS 5a (Wijmstra 1969:524). Forest communities contracted in the Ioannina Basin between 68,000 and 59,000 BP (MIS 4), and open vegetation with scattered woodlands appeared for the duration of this period (Tzedakis, et al. 2002:2045). The resolution for the other Ioannina core (Tzedakis 1993, 1994) is exceptionally coarse during MIS 4-2. Transitional parkland, grassland steppe, and desert steppe communities were identified in this sequence, but no temporal control was established, so this core will not be included any further in the discussion of these periods. The study of Lake Kopais also has somewhat lower resolution in MIS 4 between 74,000 and 59,000 BP, but there were probably oscillating conditions with an initial decrease of forest communities, followed by the expansion of forests, and finally the appearance of more open vegetation (Tzedakis 1999:431). In the Tenaghi Philippon sequence, the vegetation shifted from mixed European black pine forest to open Artemisia and Chenopodiaceae steppe during MIS 4, but there are no dates for this change (Wijmstra 1969:524). The majority of MIS 3 in the Ioannina Basin between 59,000 and 26,000 BP was represented by intermediate forest communities, with some short intervals of forest expansion and contraction (Tzedakis, et al. 2002:2045). Data from the Tenaghi Philippon sequence, which dates to the latter half of MIS 3, indicates shifts between mixed European black pine forest and dry, open Artemisia and Chenopodiaceae (Wijmstra 1969:525). This period in the Lake Kopais core also shows a continuation of oscillating forest and open conditions, with expansions and contractions of oak, Pinus (pine), Juniperus (juniper) and Abies (fir) (Tzedakis 1999:431). Lawson et al. (2004) reanalyze the later portion of core I284 from Ioannina and incorporate recent radiocarbon dates into the analysis. They look at 124 pollen and microfauna as well as sedimentological studies to reconstruct terminal Pleistocene/Holocene environments in the region. The earliest part of the sequence dates to the full glacial, starting at the end of MIS 3, from about 25,150 to 18,650 BP. The pollen is mostly from herbaceous taxa indicative of an arid environment, and the authors propose that there was slightly moister Gramineae (grass) dominated steppe interspersed with shrubby Artemisia and Chenopodiaceae semi-desert (Lawson, et al. 2004:1614). Low frequencies of oak and pine pollen also persist in the sequence, so there were probably small stands of deciduous trees in some places on the landscape. A fairly robust pollen record is available for the late glacial and Holocene, between 26,000 and 11,500 BP, in northern and central Greece. In the Ioannina Basin, more open vegetation with scattered woodlands appear late in MIS 3 and then persist until the early Holocene (Tzedakis, et al. 2002:2045). Lawson et al. (2004) finds that the late glacial pollen record of MIS 2 in the same region, from 18,650 to 11,280 BP, indicates warmer, wetter conditions than the preceding full glacial. This period is transitional between the open glacial landscape and interglacial forests, with steppe-grassland and a few trees evidenced by increasing frequencies of oak, Artemisia and Chenopodiaceae, and corresponding decreases in pine and Gramineae (Lawson, et al. 2004:1615). The Tenaghi Philippon sequence shows a gradual amelioration of the climate with an increase in forested conditions, though periods of open steppe communities continue to be interspersed between mixed or forested conditions (Wijmstra 1969:525). Lawson et al. (2005) analyze a series of cores from northern Greece, the longest of which is Nisi Fen, which spans from the late glacial through the Holocene. They find that from 20,000 to 12,000 BP, the area was mostly open, but had some small stands of Betula (birch), pine, juniper, and variable amounts of deciduous oak (Lawson, et al. 125 2005:883). To the west, Willis (1997) looked at pollen cores from Gramousti Lake and Rezina Marsh, and Turner and Sánchez Goñi (1997) examined cores from Lake Ziros and Lake Tseravinas, all of which are in Epirus. Vegetation during the glacial maximum, between 18,350 and 17,900 BP (and possibly earlier), was steppe grassland with Artemisia and Chenopodiaceae, as well as grasses and herbaceous taxa (Turner and Sánchez Goñi 1997:580). After this there was an expansion of oak woodland, resulting in a mixed foreststeppe from 17,900 to 13,020 BP (Turner and Sánchez Goñi 1997:581). The mixed foreststeppe persisted until 11,405, though pine largely replaced oak in the woodlands (Turner and Sánchez Goñi 1997:581). During the late glacial, from 15,350 to 11,205 BP, Willis (1997:404-407) finds open steppe environments, with a slight expansion of oak at about 13,020 in the same region. Kotthoff et al. (2008) examine pollen from a marine core (SL 152) in the Aegean Sea that spans the last 20,000 years. During the late glacial, from 20,800 to 14,600 BP, steppe vegetation dominates the sequence, with Artemisia and Chenopodiaceae persisting, and an increase of Ephedra (Kotthoff, et al. 2008:1022). Following this, from 14,600 to 12,700 BP, there is an expansion of broadleaf arboreal vegetation, though steppe communities persist (Kotthoff, et al. 2008:1025). During the Younger Dryas, from 12,700 to 11,700 BP, Artemisia becomes more common, followed by increases of Chenopodiaceae and Ephedra, indicating dry, open conditions (Kotthoff, et al. 2008:1025). In the Lake Kopais core, MIS 2 is represented by full glacial conditions, with high incidences of Gramineae, Artemisia and Chenopodiaceae pollen and a low frequency of arboreal pollen, indicating dry, open conditions (Tzedakis 1999:428). Okuda et al. (1999) also analyze a core from Lake Kopais that encompasses that late Pleistocene and Holocene. In a pollen zone that dates to earlier 126 than 14,750 BP (though there is no maximum date), the authors find an open steppe environment with Chenopodiaceae and Artemisia, which is thought to indicate glacial conditions (Okuda, et al. 1999:453). From 14,750 to 12,400, there was an expansion of open oak and pine forests (Okuda, et al. 1999:453), with increases of pine and Ephedra pollen. In northern and central Greece the Holocene was marked by forest expansion. Tzedakis et al. (2002:2044) observe forest expansion in the Ioannina Basin between 11,500 and 5,000 BP. Other studies also find forest expansion during the Holocene in this region, with an increase in oak, Ostrya (hop-hornbeam), and elm pollen, with the addition of Olea (olive) and fir (Tzedakis 1993, 1994). Lawson et al. (2004:1615) find changes in forest composition during the early Holocene in the Ioannina Basin, from 11,280 to 6,295 BP. Deciduous oak forests transitioned to mixed woodlands at this time, with increased quantities of Ostrya carpinifolia/Carpinus orientalis (European hop-hornbeam/Oriental hornbeam), Fraxinus ornus (Southern European flowering ash), and C. betulus (European hornbeam). Wijmstra (1969:525) also found a shift to forests at around 11,600 BP in the Tenaghi Philippon core. In the early Holocene, from 12,000 to 5,000 BP, the record at Nisi Fen indicates an expansion of forests, in particular deciduous oak and pine, though other taxa, including Ostrya-type, Tilia cordata-type (small-leaved lime), European hornbeam, Southern European flowering ash, and Pistacia (pistachio) also become more common (Lawson, et al. 2005:883). During the transition from the late glacial to the early postglacial from 11,205 to 10,495 BP, Willis (1997:408) finds a decrease in open herbaceous species, such as grasses and sedges, at the expense of expanding oak and pine forests in the Gramousti and Rezina cores. Following this, arboreal taxa in the record become increasingly diverse, though a mosaic of trees, shrubs and herbs persisted between 10,495 and 8,330 BP (Willis 1997:409). 127 Turner and Sánchez Goñi (1997:582) also find a continued spread of open forests during the early Holocene in Epirus. Kotthoff et al. (2008:1025) find increases in non-steppe herb pollen and a minor increase in non-saccate arboreal pollen (such as oak) starting in the early Holocene, about 11,700 BP. Precipitation increases after 10,200 BP led to the expansion of Holocene forests (Kotthoff, et al. 2008:1026). The Holocene in central Greece is also marked by expanded forests after 12,400 BP, with deciduous oak, Corylus (hazel) and hornbeam/hop-hornbeam (Okuda, et al. 1999:453). Southern Greece The climatic record for southern Greece is sparse compared to the north. No reconstructions exist for MIS 4 or 5, though there are some fairly complete records after about 45,000 BP (Albert 2010; Geraga, et al. 2005; Hansen 1978, 1991; Karkanas 2010; Ntinou 2010). Geraga et al. (2005) identify a series of stadials and interstadials from MIS 3 and later using planktonic foraminifera, pollen, and microfauna δ18O values from a sediment core taken from the Cretan Basin in the southern Aegean Sea. Pollen indicative of wet, forested conditions, including Quercus ilex/coccifera (holm/Kemes oak), hazel, hornbeam and elm is found between 42,000 and 35,000 BP, with a slight contraction of the forests between 41,000 and 40,000 BP (Geraga, et al. 2005:321) (Figure 3.2). The pollen record is interrupted between 35,000 and 25,000 BP. From 25,000 to 24,000 the vegetation was mixed, followed by an expansion of shrubby Artemisia and Ephedra between 24,000 and 11,000 BP, indicative of a steppe environment (Geraga, et al. 2005:321). After 11,000 BP there was an expansion of forests and modern Mediterranean taxa, such as pistachio, olive 128 and Phillyrea, that lasted until 8,000 BP, when a more mixed environment appeared until 7,000 BP (Geraga, et al. 2005:324). Hansen (1978; 1991) presents the macrobotanical remains from Franchthi Cave in southeastern Peloponnese (Figure 3.3). Because Franchthi is an archaeological site with several fairly large depositional hiatuses (see Chapter 2), the data from Hansen (1978; 1991) are fairly coarse-grained and patchy (Figure 3.2). The earliest archaeological sequence dates to between 39,280 and 20,000 BP (De Vivo, et al. 2001; Farrand 2000), and is marked by a cold, dry climate with herbaceous plants, including Lithospermum arvense (corn gromwell, now called Buglossoides arvensis), Anchusa sp. and Alkanna cf. orientalis, comprising a steppe environment (Hansen 1991:105). Between 15,000 and 10,500, there was an expansion of open woodlands, evidenced by the appearance of pistachios, almonds, pears, bitter vetch, lentils, vetch, wild oats, and wild barley (Hansen 1991:160), as well as arboreal species indicative of increased moisture and higher temperatures (Hansen 1991:161). Other macrobotanical data from Franchthi post-date the occupation of Klissoura Cave 1. A reconstruction of the immediate environment of the Klissoura Gorge (Prosymni) during the later part of MIS 3 and MIS 2 is possible based on evidence from Klissoura Cave 1. Albert (2010) and Ntinou (2010) analyzed phytoliths and charcoal, respectively, from the Upper Paleolithic and later components of the site. In general, they found that the local environment during the Aurignacian occupations of the cave during the later part of MIS 3 was milder than that of the post-LGM Epigravettian (Figure 3.2). No data were available for the Mesolithic, and preliminary geoarchaeological interpretations of the Middle Paleolithic simply indicate that the environment was more humid at that time than it was during later periods (Sitlivy, et al. 2007:4). 129 Ntinou (2010) identified plants from the Prunus genus (which includes almonds, Webb’s almond, and blackthorn) throughout the Upper Paleolithic and Epigravettian layers, indicating dry, open parkland in the vicinity of the cave. Additionally, during the Early Upper Paleolithic and Aurignacian, deciduous oak, elm, Acer (maple), and hornbeam/hophornbeam were identified (Ntinou 2010:55). These are mesophilous taxa, but thin growth rings observed on some of the samples indicate that the trees were probably growing with the minimum amount of water necessary for survival (Ntinou 2010:56). During the Upper Paleolithic, Klissoura Gorge was probably a mosaic of open parkland and open woodland, with Prunus on the hillsides and stands of mesophilous trees on the valley floor (Ntinou 2010:56). A diverse array of thermophilous species, including olive, evergreen oak and pistachio are found, particularly in Aurignacian layer IV. Ntinou (2010:56) argues that these warm-loving species are indicative of a refugium in northern Peloponnese during an interstadial of MIS 3. Together, the mesophilous and thermophilous species point to mild winter temperatures and just enough precipitation to support the growth of temperate trees (Ntinou 2010:56). Based on phytolith analysis, Albert (2010:78) finds an abundance of C3 grasses and sedges in Aurignacian layer IV, which indicates that the environment was more humid at this time than in later periods of occupation. The overlying Aurignacian layers (IIIe and IIIg) are dominated by monocots and a low diversity of C3 grasses, also indicative of greater humidity (Albert 2010:81). During this period, there was an influx of inflorescence phytoliths, meaning that grass plants were brought into the site during the flowering season (Albert 2010:81). Other components of the layer III series, III and III’ (which is not actually considered to be an Aurignacian layer, see Chapter 2), contain C3 grasses and reeds, such as 130 Arundo donax (giant cane) (Albert 2010:81). These plants may have grown near the perennial river that ran through the gorge. C3 grasses are generally indicative of more temperate environments, and Arundo donax requires fairly moist soils, often growing along rivers and in marshy areas (Turland, et al. 1993:164; Tutin, et al. 1980:253). These findings support Ntinou’s (2010) assessment that the Upper Paleolithic climate in the vicinity of Klissoura Gorge was fairly mild. The only phytoliths found in Layer 6 are from C3 grasses, but these samples come from a disturbed area designated as a pit so it is unclear if they are associated with the Upper Paleolithic or later periods. As discussed above, environmental conditions changed throughout the Mediterranean during the LGM, but this period is not represented in the Klissoura Cave 1 sequence (Karkanas 2010; Ntinou 2010). The LGM is also missing from Franchthi Cave (Farrand 2000; Hansen 1991), so this may represent a depositional hiatus in southern Greece, and possibly also a movement of human populations out of the region at this time. After the LGM, the vegetation of the gorge reflects a drying trend and a more open environment (Albert 2010; Ntinou 2010). Prunus remained dominant in the Epigravettian layers (Ntinou 2010:55). Phytoliths from C4 grasses, which are indicative of a drier, warmer environment, are found in addition to C3 grasses and reeds in the Epigravettian layers (Albert 2010:82). These lines of evidence point to dry, open parkland in the area around Klissoura Cave 1 in this later period, in contrast to a more temperate Upper Paleolithic environment (Albert 2010; Ntinou 2010). Geoarchaeological lines of evidence offer a fairly low-resolution picture of climatic conditions during the Klissoura Cave 1 occupation. Karkanas (2010:32) notes that the uppermost layers (Mesolithic and Epigravettian) have more in common depositionally with 131 the late Middle Paleolithic than with the intervening Aurignacian. This is because sedimentation during the Aurignacian can primarily be attributed to anthropogenic processes, whereas natural processes, particularly the introduction of clastic sediments into the site, caused the majority of sedimentation in the earlier and later periods (Karkanas 2010:32). Based on this, Karkanas (2010:32) suggests two possible scenarios for the accumulation of clastic materials at Klissoura Cave 1 during the Mesolithic and Epigravettian: either precipitation rates were higher following the LGM, which caused more runoff and slope wash into the cave, or a cooling and drying trend led to a more open environment in the gorge, leaving the hillsides destabilized and more vulnerable to erosion during occasional storms. Karkanas (2010) points out that the latter scenario is supported by the botanical results of Albert (2010) and Ntinou (2010). The older Upper Paleolithic layers, which include from top to bottom a Mediterranean backed bladelet industry (III’), another non-Aurignacian Upper Paleolithic industry (III”), and multiple Aurignacian layers, have loose, dusty-looking sediments that lack recrystallized or cemented wood ash, indicating that they formed under periods of relatively low humidity (Karkanas 2010:33), though this may reflect conditions inside of the shelter. Despite decreased humidity at this time, there was clearly enough moisture to sustain temperate tree populations in the valleys (Ntinou 2010) and a more stable landscape (Karkanas 2010:34). To date, botanical, phytolith and geoarchaeological analyses of the Middle Paleolithic at Klissoura Cave 1 are ongoing. General observations of the stratigraphy and formation of the Middle Paleolithic sediments indicate that they occurred under more humid conditions than those in the Upper Paleolithic (Sitlivy, et al. 2007:4). Karkanas (2010:32) notes that the late Middle Paleolithic layers may have formed under similar conditions to the Mesolithic 132 and Epigravettian, so as with the later layers, it is unclear if during the Middle Paleolithic the environment was wetter, causing an influx of clastic sediments, or if it was cooler and drier, with a decrease in vegetation and subsequent destabilization of the landscape. Once any botanical and phytolith samples from the Middle Paleolithic are analyzed, this picture will hopefully become clearer. At this point, however, we must rely on non-local climatic reconstructions from northern Greece to try to understand possible changes in the composition of faunas in the Middle Paleolithic layers. Environmental Conditions during the Occupation of Klissoura Cave 1 The end of MIS 5 is represented consistently in northern and central Greece by expanded forest conditions (Tzedakis 1993, 1994, 1999; Tzedakis, et al. 2002; Wijmstra 1969). The earliest part of the archaeological record at Klissoura Cave 1 (Middle Paleolithic XVIII-XXb) was probably deposited at this time, based on the lithic and geoarchaeological data (Karkanas, personal communication 4/23/10). However, no dates are currently available for the lowest layers of Klissoura Cave 1, so this assumption may be revised in the future. An alternative possibility is that the Middle Paleolithic XX series dates to MIS 5a, and Middle Paleolithic layers XVIII-XIX were deposited during MIS 4 (Karkanas, personal communication 4/23/10). This means that the earliest layers either formed during the milder forested or mixed forested conditions of MIS 5a, or during the open steppe conditions of MIS 4 (Tzedakis 1993, 1994, 1999; Tzedakis, et al. 2002). For the purposes of this discussion, in keeping with the lithics data, we will assume that Middle Paleolithic layers XVIII-XXb were either deposited in MIS 5a or 4, and overlying undated Middle Paleolithic layers (X-XVII) formed during MIS 4 or 3 (Table 3.1). 133 Table 3.1 Taxa expected in the vicinity of Klissoura Cave 1 by MIS stage. Note that the MIS stage of the lower layers are unknown, which is indicated by "/" and both options are included. MIS 3 is highly variable and local reconstructions are available for the UP layers but not the MP. P = present, A = absent, ? = unsure. 134 In general, environmental data from southern Greece correspond fairly well with those from northern and central Greece for the latter half of MIS 3 and later periods, though the records are quite incomplete. Small disagreements among records likely reflect local conditions, as some areas seem to lag behind others in vegetation and climatic changes (Figure 3.2). The general picture for MIS 3 in Greece seems to be a high degree of climatic variability, but with generally more amicable conditions. Both continental and southern Greece oscillated between mixed forest, steppe communities and warm, forested conditions, with just a few short periods of drying (Albert 2010; Geraga, et al. 2005; Karkanas 2010; Ntinou 2010; Tzedakis 1999; Tzedakis, et al. 2002; Wijmstra 1969) (Figure 3.2). These were the conditions during the later Middle Paleolithic occupations (VIII and possibly X-XVII) of Klissoura Cave 1, as well as the earliest Upper Paleolithic, with the Early Upper Paleolithic (Uluzzian layer V), Aurignacian series (IV, IIIe-g, IIIb-d), Upper Paleolithic (nonAurignacian) industry (III”) and Mediterranean backed-blade industry (III’) deposited during this time. Unfortunately, the limited resolution of the environmental reconstructions coupled with radiocarbon date error ranges from this period makes it impossible to correlate specific environmental fluctuations with these various cultural designations. Toward the end of MIS 3 conditions became drier and steppe communities expanded across Greece, which persisted through the LGM (Geraga, et al. 2005; Hansen 1991; Kotthoff, et al. 2008; Lawson, et al. 2004; Okuda, et al. 1999; Turner and Sánchez Goñi 1997; Tzedakis 1999; Tzedakis, et al. 2002) (Figure 3.2). The LGM is not represented at Klissoura Cave 1 (Albert 2010; Karkanas 2010; Ntinou 2010). After the LGM, particularly in northern Greece, conditions ameliorated and mixed forest-steppe communities became widespread (Lawson, et al. 2005; Lawson, et al. 2004; 135 Turner and Sánchez Goñi 1997; Wijmstra 1969). In southern Greece, it seems that conditions remained dry until shortly after 15,000 BP (Albert 2010; Geraga, et al. 2005; Karkanas 2010; Kotthoff, et al. 2008; Ntinou 2010); this also occurred in some parts of central Greece (Okuda, et al. 1999; Tzedakis 1999; Tzedakis, et al. 2002; Willis 1997) (Figure 3.2). At Klissoura Cave 1, the Epigravettian (IIa-d) was deposited during this period of continued dry, open environments. At the beginning of the Holocene, forests expanded across Greece and the modern Mediterranean-type plant taxa appeared (Geraga, et al. 2005; Hughes, et al. 2006; Kotthoff, et al. 2008; Lawson, et al. 2005; Lawson, et al. 2004; Okuda, et al. 1999; Turner and Sánchez Goñi 1997; Tzedakis 1993, 1994; Tzedakis, et al. 2002; Willis 1997) (Figure 3.2). No botanical or phytolith samples are available for Klissoura Cave 1 for the Mesolithic (3-5a), though based on widespread climatic amelioration in the rest of Greece, it is likely that conditions also became more mild in the Argolid. BIOTIC COMMUNITIES AND RESOURCE AVAILABILITY Understanding the habitat requirements of prey species at Klissoura Cave 1 is important in interpreting the suite of taxa hunted by the site occupants at different times during the Pleistocene. Ultimately, this will aid in determining if changes in species composition at the site were due to environmental causes or human hunting pressures. Ungulates and small game are particularly significant, as carnivore species are fairly rare in the assemblages (see Chapter 5). Large game species found in the assemblages include European fallow deer (Dama dama), ibex (Capra cf. ibex), European wild ass (Equus hydruntinus), aurochs (Bos primigenius), wild pig (Sus scrofa), red deer (Cervus elaphus), roe deer (Capreolus capreolus), and chamois (Rupicapra rupicapra). These taxa can be 136 broadly categorized as grazers and browsers, though considerable flexibility exists in their diets. Grazers tend to consume more grasses and sedges, while the diets of browsers are composed of more leaves and vegetation of woody plants and forbs (Janis 2008; Searle and Shipley 2008). It might be expected, particularly with taxa that adhere more strictly to these categories, that grazers would be more successful in dry, open habitats, while browsers prefer wetter, forested environments. Either group should be tolerant of mixed vegetation. European fallow deer are intermediate feeders. Grasses compose the bulk of their diets, though they also include herbs, broad-leaf tree browse, shrubs, mast, and forbs on a seasonal basis (Feldhamer, et al. 1988:317; Nowack 1999:1099). They inhabit a range of environments, including woodlands, mixed and broad-leafed forests, grasslands, scrub, savanna, and areas with subalpine vegetation (Feldhamer, et al. 1988:317). Red deer similarly live in a variety of habitats, including plains and mountainous areas, though they prefer to have access to forested areas, stands of shrubs, tall grasses or reed thickets for shelter (Baskin and Danell 2003:56). They subsist on a broader range of plants than most cervids, feeding primarily on browse, in addition to grasses and herbs (Baskin and Danell 2003:59-61). Their diets are largely seasonal, with grasses, rushes and sedges, heaths, forbs, deciduous browse and conifers common in the summer, and heaths consumed in the winter (Latham, et al. 1999:412). There may also be considerable geographical variation in red deer diets (Baskin and Danell 2003; Latham, et al. 1999), although their ecological niche is wide in most regions. Roe deer prefer woody and mosaic landscapes with dry plains or steppe and adjacent small stands of forest, shrubs or grass thickets, though they avoid dense forests (Baskin and Danell 2003:110). Their diets consist primarily of herbaceous plants, such as forbs, and to a lesser extent grasses, sedges, rushes, heaths, deciduous browse, conifers and 137 ferns; the majority of the latter is consumed during the summer (Baskin and Danell 2003:113; Latham, et al. 1999:412). Ibex are flexible grazers that subsist primarily on grasses in the spring and summer, but will also browse on forbs, leaves, shoots and bark during the winter, though they typically avoid dense forests (Nowack 1999:1223; Roberts 1977:194; Schaller 1977). Ibex are usually associated with high-altitude environments (Nowack 1999:1223; Roberts 1977:193; Schaller 1977), but Phoca-Cosmetatou (2004b) has argued that this may be a product of recent marginalization by habitat depletion. Today, in protected areas, ibex occupy a wide range of ecozones, preferring those that are rugged but not necessarily highaltitude. Chamois are mixed feeders that inhabit alpine meadows during the summer and spring, moving to lower elevations in the fall and winter (Nowack 1999:1213), where they occupy rocky mountain forests or cliffs near the edges of forests (Baskin and Danell 2003:312). In the summer they subsist on herbs, grasses and flowers, and in the winter they consume dry grass, twigs of shrubs, young pine shoots, moss, and lichens (Baskin and Danell 2003:314; Nowack 1999:1213). It is more difficult to understand the ecology of the European wild ass, Equus hydruntinus, because the species is extinct. Recent genetic studies have shown that E. hydruntinus is closely related to Equus hemionus, or the Asian wild ass (Orlando, et al. 2006:2087), so it may be appropriate to use Asian wild ass ecology as a proxy for that of the European wild ass. It should be noted that samples of European wild ass specimens tested for genetic material by Orlando et al. (2006:2084) came from areas that were cold steppe and open, dry grassland at the time that E. hydruntinus inhabited the region. Today, the Asian wild ass lives in semi-desert and steppe environments (Baskin and Danell 2003:5). They 138 subsist mostly in cereal grasses, wormgrass and sedges, though they also consume herbaceous plants and shrubs to a lesser extent (Baskin and Danell 2003:6). Aurochs, the wild progenitors to cattle, are also extinct, though several studies were conducted in an attempt to understand their ecology. Based on their distribution at archaeological and paleontological sites, climatic reconstructions at these localities indicate that they tended to live in low-lying, flat floodplain ecosystems (Hall 2008:189; van Vuure 2005:175). Like modern cattle, and based on morphological observations and early documentation of the animals before they were extinct, aurochs were likely grazers that mostly ate grasses, sedges and rushes (van Vuure 2005:213). Wild pigs have different feeding habits than most other ungulates. They are omnivores, and feed on a variety of invertebrates, small mammals, bird eggs and carrion, as well as tree bark, grasses, seeds and acorns (Baskin and Danell 2003:20; Graves 1984:484). Though they are extremely flexible feeders and are able to tolerate a wide range of habitats, they require dense stands of forest, shrubs or grass for shelter, typically near water holes, rivers or swamps (Baskin and Danell 2003:16; Graves 1984:487). Fallow deer are the most adaptable of the ungulate species found at Klissoura Cave 1, though red deer, roe deer, chamois and ibex are all comfortable with mixed forest/steppe ecozones. Ibex are also found in more open environments, which is probably where European wild ass and aurochs would have occurred as well. Only pigs prefer more forested areas. Small game species found in the Klissoura Cave 1 assemblages includes one of three species of similarly-sized tortoises (Testudo graeca, T. hermanni, or T. marginata), European hare (Lepus europaeus), rock partridge (Alectoris graeca) and great bustard (Otis tarda). 139 Tortoises are fairly flexible in terms of habitat requirements, occupying lush meadows, scrub covered rocky hillsides, woodlands, and stabilized dunes, though they prefer moderately dense vegetation (Arnold and Burton 1992). Hares also live in a wide range of environments, including open fields, grasslands, steppes and woodlands, though they prefer flatter areas (Burton 1991; Mitchell-Jones, et al. 1999). Rock partridges occupy areas with oak, pine and juniper, and open stands of beech, hornbeam, oak and oriental plain, preferring very open and rocky areas near water when nesting (Cramp 1980:659; Handrinos and Akriotis 1997; Vavalekas, et al. 1993). Great bustards are large ground-dwelling birds that typically inhabit more open areas, including steppe environments, lowlands, river valleys and treeless plains (Cramp 1980:659; Heinzel, et al. 1992:128; Johnsgard 1991). Today they are also found in stubble fields (Lane, et al. 2001). In general, the important small game species found at Klissoura Cave 1 tend to be rather non-selective in their habitat requirements. In particular, tortoises, hares and partridges probably were not affected by century or millennial-scale shifts in plant communities in southern Greece, because these changes were never that extreme in this part of the Mediterranean. Great bustards are a notable exception that likely were affected by vegetation changes. They exhibit a clear preference to open areas; with wingspans up to 2.6 m (Cramp 1980:659), they require treeless environments to effectively take flight. CONCLUSIONS In general, changes in the proportions of large game species at Klissoura Cave 1 can serve as a rough indicator of shifts in plant communities in southern Greece during the occupation of the site. This of course does not take into account the distances that hominins 140 were willing to travel to find game, but this will be addressed in Chapter 6, and was probably not a major factor. A summary of game animals expected in different archaeological layers, based on habitat preferences and environmental reconstructions is presented in Table 3.1. In terms of large game, fallow deer were likely the dominant ungulate species in the vicinity of Klissoura Cave 1 throughout the Paleolithic and early Holocene sequence. Certain taxa, including red deer, roe deer and chamois were probably available to hominins occupying the site when there were mixed open and forested communities in the area. Ibex were present in periods of mixed steppe/forest environments, or when it was more open. During periods with dry, open steppe conditions, wild ass and aurochs were more likely to have been an option for human hunters. Wild pigs were probably only available when conditions were wetter and there was an expansion of forests. Based on their habitat requirements, it is likely that small game animals such as tortoises, hares and partridges were available to humans occupying Klissoura Cave 1 throughout the late Pleistocene. These taxa are found in almost every level of the site, indicating that they were present in the environment to some degree (see Chapter 5). There is little reason to believe that environmental changes discussed above would have vastly changed the frequency of these animals in and around Klissoura Gorge, so their varying proportions in the archaeological assemblages likely reflects shifts in human hunting strategies. Conversely, great bustards are sensitive to vegetation shifts, so their presence and absence likely tracks expanding and contracting grasslands. Considering only environmental conditions, and depending on the extent of increased moisture and expanded forests in southern Greece, or if the earliest archaeological layers were deposited during MIS 5a or 4, two possible suites of prey species were probably 141 available during the earliest Middle Paleolithic occupation (XXa-XXb, XVIII-XIX). Fallow deer and wild pigs, as well as tortoises, hares, and partridges are expected in the vicinity of Klissoura Cave 1 if these layers were deposited during MIS 5a. Conversely, if the layers formed during MIS 4, fallow deer, ibex, wild ass, aurochs, and the full suite of small game (including great bustard), are expected. The situation is similar for the overlying Middle Paleolithic layers (X-XVII), if they were deposited during MIS 4 or 3. In these instances, changes in the suite of species found at Klissoura Cave 1 may provide insight into the specific environmental shifts occurring in southern Greece during the Middle Paleolithic. This would be apparent in the relative evenness of ungulate species exploited, as well as the presence or absence of key ungulate species such as wild pig, or wild ass and aurochs. Changes in the representation of tortoises, hares and partridges would likely have more to do with human hunting pressures than environmental variability (see Chapter 5), though the presence of great bustard would probably correlate with an open environment. Radiocarbon dates indicate that the latest Middle Paleolithic layer (VIII), Early Upper Paleolithic (Uluzzian layer V), Aurignacian series (IV, IIIe-g, IIIb-d), Upper Paleolithic (non-Aurignacian) layer III”, and the Mediterranean backed bladelet industry (III’) formed during MIS 3. The climate was highly variable during this interval and oscillated between wet, forested conditions, milder phases of mixed forest-steppe vegetation, and dry, open periods. Mixed forested and opened areas probably persisted in the direct vicinity of Klissoura Cave 1 at this time, though there was slightly more moisture when layer IV was deposited. Environmental fluctuations during this period may again be indicated by the presence or absence of certain ungulate species, or great bustard. 142 The Epigravettian (IIa-d) occupation of Klissoura Cave 1 occurred at a time when dry, open vegetation persisted. It is expected that the full range of small game animals, including great bustards, were present in the environment, in addition to a fairly limited range of ungulate species that include fallow deer, ibex, European wild ass, and aurochs. The Mesolithic (3-5a) that followed was marked by increasing moisture and expanded forests of the Holocene. Faunal resources available in and around Klissoura Gorge would have included wild pig and fallow deer, in addition to small game animals such as tortoises, hares and partridges. In the chapters that follow, game use and prey selection will be examined from the Middle Paleolithic through Mesolithic at Klissoura Cave 1. Based on the environmental reconstructions, habitat requirements of prey species, and predictions discussed above, I will make the case that some changes in game use across the Pleistocene and early Holocene at Klissoura Cave 1 were linked to shifts in vegetation communities in southern Greece, while others were a product of hunting pressures on certain species exploited by hunters in the region. 143 CHAPTER 4: TAPHONOMIC PROCESSES BONE ACCUMULATION AND ATTRITION INTRODUCTION A key issue in evaluating zooarchaeological assemblages is determining whether faunas were accumulated and modified by human foragers, and the extent to which nonhuman processes impacted the remains. At an archaeological site such as Klissoura Cave 1, with unquestionable human activity throughout the sequence evidenced by an abundance of stone tools (Kaczanowska, et al. 2010; Koumouzelis, Ginter, et al. 2001; Sitlivy, et al. 2007) and hearths (Karkanas, et al. 2004; Pawlikowski, et al. 2000), it seems likely that humans were mostly responsible for the collection of the faunal materials. However, there is no guarantee that all animals were introduced into the site by humans. Nor is it certain that nonhuman processes, such as weathering, carnivore damage, or in situ attrition, played no role in shaping the assemblages. Several analyses will be presented here to determine the extent to which, if any, non-human taphonomic agents impacted the faunas of Klissoura Cave 1. These include an evaluation of the input of faunal remains to better understand differences in sample size and faunal density, observations of macroscopic damage caused by weathering, carnivores and rodents, an analysis of in situ density-mediated attrition, and the degree of fragmentation of the assemblages to evaluate potential changes in breakage related to time period or prey body class. Taphonomic processes are those that act on an animal’s remains after it dies and include the pre-burial and post-burial histories of the carcass (Lyman 1994:3, see also Gifford 1981). Taphonomic agents unrelated to human activities include transport and 144 gnawing by carnivores (Hockett 1999; Lloveras, et al. 2008a; Stiner 1991, 1994), digestion (Andrews 1990; Lloveras, et al. 2008a, 2008b, 2008c; Rensberger and Krentz 1988), chemical and mechanical weathering (Behrensmeyer 1978; Lyman and Fox 1989; PhocaCosmetatou 2005), soil corrosion and diagenesis (Karkanas, et al. 1999; Nielsen-Marsh and Hedges 2000; Stiner, et al. 2001; Weiner, et al. 1993), rodent gnawing (Brain 1980, 1981; Hockett 1989; Hoffman and Hays 1987; Maguire, et al. 1980), and root etching (Denys 2002; Fernandez-Jalvo, et al. 2002; Fisher 1995). Other taphonomic processes are the result of human hunting and processing behaviors, including burning, carcass transport, trampling, butchery, bone marrow extraction, and heat-in-liquid grease rendering (Bar-Oz and Munro 2005; Bunn, et al. 1988; Costamagno, et al. 2005; Denys 2002; Fisher 1995; O'Connell, et al. 1988; Outram 2001). Though human processes could easily be included in this chapter, they are explored separately in Chapter 6, for organizational purposes. In an archaeological assemblage, there are three overarching processes that affect faunal remains: accumulation, modification and attrition. Accumulative agents may be human hunters carrying prey to a camp site, carnivores that bring a kill back to a shelter, or a fluvial system that washes objects to a particular area. Modifications include processes that may affect the appearance of bone while not always causing bone loss, such as low levels of carnivore and rodent gnawing, weathering, and human butchering or burning. Agents that cause attrition, or bone loss, include chemical or mechanical processes that lead to bone diagenesis, carnivore gnawing and digestion, and human processing. In some cases, those agents responsible for accumulation or modification may also cause attrition. For example, carnivores can bring kills to a rock shelter or cave, but they can also move materials away from an archaeological site during scavenging or destroy bones entirely. Aside from human 145 subsistence behaviors such as the selective accumulation of meaty body parts or high-ranked species, burning, and butchery damage (see Chapters 5 and 6), the main non-human taphonomic processes that may have affected the Klissoura Cave 1 faunas are weathering and diagenesis, trampling, bioturbation, and carnivore accumulation and gnawing. METHODS A macroscopic survey of the Klissoura Cave 1 faunas indicate that overall there is good preservation of the materials. Excavation recovery methods included screening and the recovery of both shaft and spongy bone fragments. Elements representing a range of structural densities (see Lyman 1984; 1994) were recovered, in addition to fragile fetal elements in many layers (Chapter 6). Variation in bone condition exists through the sequence; the oldest Middle Paleolithic assemblages (XI-XIV and lower) are highly mineralized and some display a metallic blue patina on their surfaces (Figure 4.1). Most Upper Paleolithic remains are covered by a layer of solidified ash that could not be removed without damaging the surface of the bones (Figure 4.2). Attempts to remove the ash on bird bones using acetic acid proved to be ineffective and potentially damaging to the materials (Bocheński and Tomek 2010). This coating likely obscures some of the butchery damage, in particular fine tool marks, though impact fractures and green bone breaks were still easily discernable. Because bone surfaces in the Upper Paleolithic assemblages were often obscured, there are some limitations when it comes to discussing cut marks, which will be addressed in Chapter 6. The skeletal remains are highly fragmented, particularly in the Middle Paleolithic layers, and the only intact elements are small, compact tarsal and carpal bones and a few phalanges. Excavation damage is uncommon in the assemblages, and is 146 obvious when it occurs because the broken edges are noticeably lighter than the rest of the bone. Though the materials are highly fragmented, many of the specimens are easily identified to species or body size class and to skeletal element, based on diagnostic features or “landmarks.” Figure 4.1 Metallic blue patina on Middle Paleolithic long bone fragment. From layer XIX. Bioturbation does not seem to be a major factor at the site. A few micromammals were observed in the assemblages, but their frequencies were quite low and rodent damage is uncommon (see below). In general the stratigraphy is fairly intact and was analyzed carefully (Karkanas 2010). Areas of the site thought to be intrusive pits or mixed during excavation or geoarchaeological analysis were excluded from the zooarchaeological analysis. There is a slight amount of mixing between the earliest Upper Paleolithic and latest Middle Paleolithic layers, evidenced by the presence of a few shell ornaments in the Middle Paleolithic layers VI and VII in the area immediately below the Aurignacian shelter feature (Stiner 2010), so layers VI and VII are excluded from the analysis. The ornaments are small in size (mean 1.28 147 cm), and could have moved between layers when humans disturbed the sediments or fallen into rodent burrows. The lithic industries in Layer VIII and below are Middle Paleolithic. Figure 4.2 (a) Light carbonate fraction on great bustard tibiotarsus. From disturbed zone (layer 6a). (b) Heavy carbonate fraction on great bustard lumbar vertebrae. From disturbed zone (layer 6). The analytical units and techniques for the zooarchaeological analyses follow those outlined in Grayson (1984), Klein and Cruz-Uribe (1984), Stiner (1994) and Lyman (1994). The Wiener Laboratory faunal reference collection at the American School of Classical Studies at Athens and unpublished electronic faunal manuals created by Mary Stiner were used to identify the materials. The basic counting unit in the analysis is number of identified specimens or NISP (Grayson 1984). The term as used here refers to every fragment or whole 148 bone that could be identified to element and body class, or more specific designation. Specimens were identified to the greatest precision possible: genus (e.g., Testudo sp.), species (e.g., Dama dama) or body size category (e.g., medium ungulate or small mammal). In the case of articulated elements or mandible portions with teeth, each individual element that could potentially be separated was counted individually to control for variation in postdepositional processes (Stiner 1994) but with articulations noted. For example, the NISP of a complete right mandible of an adult fallow deer with all of the teeth intact would be eleven. The counting unit NISP is fundamental but is not sufficient for all analyses. It does not, for example, take into account fragmentation or identifiability that may differ among species, or differential preservation (Grayson 1984:21-23; Lyman 2008:29-30). Because of this, some derived units are used in addition to NISP. Minimum number of elements, or MNE, is derived from NISP by recording unique features on each element and tabulating the number of times each landmark occurs in an assemblage for a species or ungulate body size group. For example, if there are four medial distal femur fragments from Dama dama in a given layer, the MNE is four. Side is not taken into account when calculating MNE. In the case of long bones, landmarks on both epiphyses (ends) and diaphyses (shafts) are considered. Another commonly used counting unit is minimum number of individuals, or MNI, which is the smallest number of animals necessary to account for the faunal remains at a site (Grayson 1984:27). It is derived from the highest count of the most commonly occurring element (MNE) in an assemblage, divided by the number of times that particular element occurs in the body. MNI is used less in this study than NISP and MNE, for reasons that will be explained in Chapter 5. 149 Other observations were recorded for each specimen (following Stiner 1994, 2005), including fusion state in the case of bones, wear stages for mandibular teeth, body side, burning damage and intensity (Stiner, et al. 1995), and surface damage from weathering, plant roots, gnawing rodents and carnivores, and tool marks. The coding of elements, portion-of-element, age criteria, and taphonomic variables follows Stiner (1994; 2005). The maximum length of each specimen was recorded to control for the relationship between fragmentation and differences in prey body size. It must be noted that sample totals in various tables in this dissertation vary according to whether the analysis is confined to specific-specific data or includes specimens that could be identified only to body size class. Shattered dental specimens or overly-recognizable bone fragments (such as tiny pieces of tortoise shell or the anterior groove on cervid metatarsals) pose similar problems and are excluded from tables that indicate species frequencies. As an example, Table 4.1 illustrates the overrepresentation of cervids that would result from including the anterior groove of metatarsals. These specimens are included in the taphonomic assessments, however, such as overall degree of burning or fragmentation rates. Faunal assemblages in the different layers at Klissoura Cave 1 vary in size, though not to a great degree (Table 4.2). Some parts of the uppermost layers were examined by previous studies (Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001; Tomek and Bocheński 2002) and were unavailable for analysis. In addition, the excavated area of the site was not the same for the entire sequence because the excavation grid was reduced downwardly from 14 m2 in the uppermost layers to 6 m2 in the lowest layers (see Chapter 7). Lateral variation in material density in the sediments is also noted in layers representing the most intensive occupations or those that experienced less erosion. Finally, a 150 Table 4.1 Fallow deer metatarsal NISP counts and NISP:MNE ratios that include anterior groove fragments and corrected counts that exclude anterior groove fragments, compared to metacarpal NISP counts and ratios. Metacarpals were used for comparison because their structural density is similar and they are equally recognizable in an assemblage as are metatarsals. NISP counts for metatarsals exclude anterior grooves for taxonomic frequencies because they over-represent fallow deer frequencies. Only Middle Paleolithic layers are depicted because they are the most extreme, but cervid metatarsal anterior grooves (and other overly-recognizable fragments such as unidentifiable tooth and tortoise shell fragments) are excluded in all layers. small portion of the abundant materials from the Aurignacian series IIIe-g and nonAurignacian Upper Paleolithic (III”) were excluded from the sample due to time constraints, though the majority of the bones from these layers were analyzed. An analysis of the input of faunal materials relative to sediment volume in the different layers of Klissoura Cave 1 was conducted following Stiner and Munro (2011). This was done in order to understand if sample size differences between the assemblages can be explained by differences in excavated volumes, and the relationship between the accumulation of sediments, artifacts, and faunal remains in each layer. Faunal input (or density) was evaluated by dividing vertebrate NISP counts by the sediment volume in liters for each archaeological layer, or input = NISP/L. Some attempt can be made to correct accumulation rates by dividing input by the time interval represented by depositional layers given a reliable set of radiometric dates, which are lacking in the Middle Paleolithic layers of the site. Accumulation rates for the faunal remains can then be compared to our understanding of the site from a geoarchaeological standpoint. According to Karkanas (2010) the sediments of the earlier 151 Upper Paleolithic have a strong anthropogenic signature, while the Middle Paleolithic and post-LGM layers have a greater input of geogenic sediments. Table 4.2 Rates of faunal input (or faunal density) by layer. Two methods were used to determine rate of bone fragmentation for different layers at the site. Maximum length was recorded for all specimens, assuming they were not broken during excavation. Specimens were then grouped by body class for each layer to correct for structural differences between prey of different sizes and locomotion, for example ungulates and birds. The means for each group were compared by layer to determine if there was a change in fragmentation through the sequence that may reflect intentional bone breakage or post-depositional processes such as trampling. The second method used to examine fragmentation rates between levels is the index MNE/NISP, which provides a standardized value between 0 and 1, where smaller numbers reflect greater rates of fragmentation. The MNE variable is correlated to NISP but the index is valuable for isolating differences in the degree of fragmentation within a prey size class. In examining the fragmentation index, only 152 medium ungulate remains were included because they (mostly fallow deer) represent the bulk of the assemblage throughout the sequence (see Chapter 5). Klissoura Cave 1 is a rock shelter, which makes it an attractive refuge for hominins as well as carnivores. Like humans, large carnivores have the potential to accumulate the remains of large-bodied prey at a site. Small and large carnivores can also impact archaeological faunas by denning, which can mix sediments, and gnawing on human food refuse. Many mammalian carnivores are recorded in Greece during the Late Pleistocene, including Ursus arctos (brown bear), Ursus spelaeus (cave bear), Canis lupus (wolf), Vulpes vulpes (red fox), Crocuta sp. (hyena), Panthera sp. (lion, leopard), Lynx lynx (Eurasian lynx), Felis silvestris (wild cat), and various mustelids such as martins (Martes sp.) and badger (Meles meles) (Tsoukala 1992). Mammalian carnivore influences on an archaeological assemblages can be assessed through several lines of evidence: the presence of carnivore bones that lack human modifications, prey body part profiles that match known carnivore transport preferences (Stiner 1991), fetal or neonate carnivore remains that indicate the shelter was used as a den (Stiner 1994), coprolites, and carnivore damage on the bones themselves (Maguire, et al. 1980; Marean and Spencer 1991; Stiner 1994). Stiner (1991; 1994:249-258) argued that assemblages accumulated by Paleolithic hominins tend to be more complete in their representation of ungulate body parts (with the exception of some headdominated Middle Paleolithic faunas), whereas canid and hyena accumulations often contain more head parts, or feet. Ungulate body part profiles will be addressed in Chapter 6. Observable mammalian carnivore damage includes crenellation, salivary rounding of broken edges, punctures, tooth drag marks, evidence of digestion, and an absence of long bone epiphyses in conjunction with other damage (Binford 1981:44-49, 51-77; Brain 1981; 153 Marean and Spencer 1991; Stiner 1994:140-147). Presence or absence and type of carnivore gnawing was recorded during the analysis of the Klissoura Cave 1 faunal assemblages. Rodents can be involved in both the accumulation and destruction of bone. In particular, wood rats (Neotoma sp.) are common accumulators of small mammal, fish and reptile bones, as well as medium-sized ungulates (Hockett 1989; Hoffman and Hays 1987). Porcupines (Hystrix africaeaustralis, the African porcupine, in particular) are also collecting agents, and they often cause major damage to ungulate remains by gnawing them (Brain 1980, 1981; Maguire, et al. 1980). Though these particular species were not necessarily present at Klissoura Cave 1 and therefore did not disrupt the assemblages, other rodents were in the area and their impact on the faunas must be evaluated. To do this, rodent damage was recorded when present on a specimen. Diurnal raptors (order Falconiformes) and owls (order Strigiformes) are potential accumulators of bird, mammal, and reptile remains in an archaeological assemblage. Many raptors and owls use rock faces and shelters for nesting (Hockett 1996; Lloveras, et al. 2008a). A series of taphonomic studies were undertaken to examine the effects of Falco rusticolus (gyrfalcon), Aquila heliacal (imperial eagles), Asio otus (long-eared owl), Strix aluco (tawny owl), and Bubo bubo (eagle owl) feeding patterns and digestion of smaller avian prey (Bocheński, et al. 1997; Bocheński and Tomek 1994; Bocheński, et al. 1993). These studies indicate that fragmentation rates and the survivorship of particular elements differ depending on the prey species. Taphonomic signatures of specific avian predators may be difficult to discern in an archaeological assemblage, which may represent a palimpsest of many raptor meals. However, damage from digestion is a common feature of the assemblages that many predatory birds accumulate: snowy owls 6-60% of NISP, gyrfalcons 154 67-97%, long-eared owls 55%, tawny owls 82% and eagle owls 84% (damage percentages for the last three raptors only includes femurs) (Bocheński, et al. 1997; Bocheński and Tomek 1994; Bocheński, et al. 1993). If Falconiformes or Strigiformes were the accumulators of bird remains, some digestion damage should be apparent. Similarly, leporid remains should also be expected to exhibit corrosion if they were digested and regurgitated by diurnal raptors or owls (Hockett 1996; Lloveras, et al. 2008a, 2008c). Observations of Falconiformes and Strigiformes nesting sites in North America, Europe and Africa indicate leporid assemblages accumulated by eagles tend to contain mostly hind limbs, and occasionally head parts (Cruz-Uribe and Klein 1998; Hockett and Haws 2002; Lloveras, et al. 2008c). Owl pellets are identified as containing more forelimbs and head parts (Hockett and Haws 2002), but owl nesting sites are observed to have the opposite set of remains (Lloveras, et al. 2008a). Therefore, hare body part profiles dominated by either front limbs or hind limbs may be indicative of raptor nesting. On the other hand, a combination of diurnal raptor and owl feeding behaviors may create the illusion of a more even representation of forelimb and hind limbs in an assemblage, so again digestion damage is necessary for implicating birds of prey in contributing to the accumulation of faunas in an archaeological site. Sampson (2000) analyzed tortoise bones from multiple South African rock shelters and raptor kill sites in order to distinguish between the taphonomic signatures of diurnal raptors, small carnivores and humans. Raptor roosting sites contain higher proportions of intact tortoise crania and axial elements, low frequencies of forelimbs, and few carapace or plastron fragments, since shells tend to be abandoned at kill sites. Small carnivores assemblages, on the other hand, tend to contain few cranial and axial elements, and larger numbers of forelimbs and shoulder girdles (Sampson 2000). The major differences between human and 155 non-human assemblages are a high degree of burning and higher percentages of shell fragments in anthropogenic contexts (Sampson 2000). Another set of taphonomic processes that may alter an archaeological assemblage is weathering, defined by Behrensmeyer (1978:153) as the process of separation and destruction of the original organic and inorganic components of bone by physical and chemical agents. Weathering encompasses a range of characteristics, from cracking and exfoliation, to root damage, to chemical diagenesis. One set of weathering processes, that includes bleaching and erosion, occurs more frequently at open-air sites when remains are exposed to the sun and wind before being buried (e.g., Behrensmeyer 1978; Lyman and Fox 1989). In order to identify surface weathering at Klissoura Cave 1, several different weathering stages for each of the specimens are recorded (following Stiner 2005, but also including root damage). Soil chemistry is also a major factor in the preservation of bone. A number of studies focusing on Paleolithic cave sites have identified the presence of specific minerals in sediments, specifically calcite and dahllite, as indicators of good bone and wood ash preservation (Karkanas 2001; Karkanas, et al. 1999; Schiegl, et al. 1996; Stiner, et al. 2001; Weiner, et al. 1993). Conversely, an absence of calcite and dahllite indicates that bone should not have preserved archaeologically. This principle was applied to Theopetra Cave in northern Greece, and Hayonim and Kebara Caves in Israel to explain the lack of faunal materials in certain areas of the sites (Karkanas, et al. 1999; Schiegl, et al. 1996; Stiner, et al. 2001; Weiner, et al. 1993). To the extent that distributions of faunal remains (see Chapter 7) are heterogeneous at Klissoura Cave 1, particularly as compared to the frequency of lithics, in situ dissolution should be investigated. 156 A final taphonomic process that will be addressed here is density-mediated attrition, which is bone loss as a function of the structural density of bone macrostructures. The structural density of an element portion is the ratio of the mass of the bone to its volume (Lyman 1984:264; 1994:237). The basic premise of density-mediated attrition is that portions of elements with low structural density, for example highly porous long bone epiphyses, should be more susceptible to being damaged or destroyed by mechanical and chemical processes than are portions of elements with high structural density, such as long bone shafts (Binford and Bertram 1977; Brain 1981; Lyman 1994). Mechanical (and possibly chemical) attrition should disproportionately affect specimens with low bulk densities because they have high surface area to volume ratios and more exposed surface to be damaged; therefore, specimens with low bulk density should have reduced survivorship in an archaeological assemblage (Lyman 1984:279; 1994:239). Various methods including photon densitometry, duel energy X-ray densitometry, and computed tomography (CT) have been used to derive bone density standards for comparison to archaeological data (Kreutzer 1992; Lam, et al. 1999; Lyman 1984, 1994). Bone density values have been derived for diverse taxa, including white-tailed deer, caribou, pronghorn, domestic sheep, cattle, pigs, leporids, marmots, deer, and bison, among others (Ioannidou 2003; Kreutzer 1992; Lyman 1984, 1992; Pavao and Stahl 1999; Symmons 2005). Though this list is extensive, it does not include all of the important species in Mediterranean archaeological assemblages. Some criticisms have been leveled concerning the use of bone density values of one species as a proxy for other species in an archaeological assemblage because of recorded variation between taxa that stems largely from locomotive adaptations or the way that mass is distributed across the bodies of different taxa (Ioannidou 2003; Kreutzer 1992; Pavao and Stahl 1999). However, bone 157 density values of similar species are routinely applied to archaeological assemblages, because tests of density-mediated attrition are so valuable and in some cases there are no other sets of data to use (Stiner 2005:86). Some degree of ambiguity in the reference data used in this study cannot be avoided. However, Lam et al. (1999), among others, have argued that differences are mainly between mammalian families rather than within them, or are the result of studies using different methods to derive density values. They compare bone density values for bovids, cervids, and equids and find that intertaxonomic variation is low enough that density values can appropriately be applied to different species with similar morphologies (Lam et al. 1999:356). Therefore, reference data derived for other cervids (i.e., Odocoileus sp.) are generally suitable for analyzing fallow deer survivorship at Klissoura Cave 1. Much of the work seeking to understand differences in the structural density of bone and density-mediated attrition was stimulated by Binford’s (1978) utility curve study in his ethnoarchaeological work with the Nunamiut (see Chapter 6). Utility curves have subsequently been widely applied to faunal assemblages in order to understand transport decisions by past foragers. However, skeletal portions that rank high in utility because of their meat or fat content tend to rank low in structural density (Lyman 1985), so in cases where an assemblage suffered from density-mediated attrition, the representation of remains fit with the interpretation of people transporting carcass portions based on reverse utility strategies (e.g., selectively transporting low-utility elements). Bone loss from density-meditated attrition can also arise from chemical and mechanical weathering processes, as well as carnivore gnawing and human butchery (e.g., Binford and Bertram 1977; Davis 1987; Lyman 1984). The most porous, and therefore least 158 dense, portions of elements are filled with deposits of bone grease. This grease is attractive to carnivores, which is why they tend to preferentially ravage long bone epiphyses and other porous elements (Blumenschine 1988; Marean and Spencer 1991; Marean, et al. 1992), and is also exploited by some humans during bone grease rendering (e.g., Binford 1978; Brink 1997; Lupo and Schmitt 1997; Vehik 1977). Therefore, evidence of density-mediated attrition only indicates that bone loss related to structural density occurred in an archaeological assemblage, but it does not identify the cause as being non-human chemical or mechanical weathering, carnivore gnawing, or human butchery or transport patterns. An evaluation of density-mediated attrition nonetheless is crucial to interpretations concerning transport decisions. Two tests evaluate the potential of density-mediated attrition to explain prey body part representation at Klissoura Cave 1. The first compares the highest tooth-based MNE to bone-based MNE for skull parts in ungulate species in each archaeological layer (following Stiner 1994:99-103) and is based on comparable differences between bone and tooth accumulation and attrition. In the transport of ungulate remains to an archaeological site, it is expected that teeth would tend to remain with the skull, which would lead to a roughly 1:1 ratio of tooth and cranial bone-based MNE. Overall, mammal teeth are less susceptible than bone to most destructive processes because the mineral component of tooth enamel (95%) far exceeds that of any bone (70%) (Currey 1984; Hillson 2005:146; Lyman 1984:72, 79). If significant fragmentation occurred during the processing of the heads by human hunters, or if gnawing or other attritional processes happened post-depositionally, even durable bony features on the skull should break down more often than teeth. Therefore, a significant bias in 159 the ratio of tooth and cranial bone-based MNE in favor of higher tooth-based MNE would indicate some attritional processes affected the assemblages. The second test of density-mediated attrition compares observed patterns in bone representation in an assemblage to independent standards of bone tissue density values obtained by the photon-densitometry technique (Lyman 1984; 1994: see also Lam et al. 1999 on the CT technique). Bone survivorship is evaluated for each portion of an element by dividing the observed frequency of each portion in an assemblage by the expected frequency for this part in a complete skeleton. This value is then standardized against the most commonly represented body part in the assemblage for a given species to determine percent survivorship (Binford 1978; Lyman 1994). Percent bone survivorship was determined for fallow deer and hare remains because they are the most common taxa through the sequence. In this test, and in the discussion of body part representation in Chapter 6, hare data are combined with data for indeterminate small mammals since most of these remains probably come from hare at Klissoura Cave 1, as opposed to similarly-sized animals such as fox or wild cat which are extremely rare in the assemblages. Similarly, remains designated as medium ungulates were combined with taxon-specific fallow deer data, since fallow deer was always the most common medium ungulate species in the faunas by a large margin. In the Aurignacian layer IV assemblage, where there was a significant proportion of other medium ungulates (i.e., ibex), not all remains that could only be identified as “medium ungulate” were lumped with fallow deer data as described previously. Rather, the proportion of medium ungulates specifically identified as fallow deer and ibex in this layer was noted (57.4% and 42.6%, respectively, based on NISP) and 57.4% of elements identified as “medium ungulate” were lumped with the fallow deer data. This is hardly ideal, particularly 160 because transport decisions for ibex and fallow deer may have been different if ibex were hunted further from the site, but it was done in order to avoid over-representing fallow deer remains. Bone density standards for American deer (Odocoileus sp.) were applied to fallow deer from Klissoura Cave 1 (from Lyman 1982, 1984), and the standards applied to European hare were those developed for snowshoe hare (Lepus canadensis) (Pavao and Stahl 1999). Comparing observed and expected body part representations at an archaeological site is useful for identifying biases in zooarchaeological assemblages, but it cannot determine whether human or non-human processes led to the biases. The general type of site must be identified (e.g., hunting station, processing site, residential camp) as well as the agent of transport and modification. Ethnographically, human foragers are shown to be selective about which ungulate body parts will be carried over long distances, whereas small game carcasses tend to be carried to a site whole (e.g., Binford 1978; Bunn, et al. 1988; O'Connell, et al. 1988; Schmitt and Lupo 1995; Yellen 1991). Once human foragers accumulate these remains at a site, the prey items would be subject to processing for consumption as well as post-depositional processes. Transport biases aside, the survivorship of hare and fallow deer bones should be about the same at Klissoura Cave 1 if post-depositional destruction was the main cause of anatomical bias (e.g., Manne and Bicho 2009; Manne, et al. 2005; Munro 2004). If hare bones show no indications of density-mediated attrition but the ungulate remains do, then the case can be made that biases in fallow deer body part representation were caused by selective transport or intensive processing on-site (following Munro 2004). Conversely, several explanations are possible if there is evidence that hare bones underwent density-mediated attrition but ungulate bones did not. One is that in general, hare elements are structurally less dense than deer bones (Pavao and Stahl 1999; see values in Lyman 1994) 161 and are more susceptible to trampling damage. Another possible explanation is that hares were not brought to the site whole, or they were processed and consumed in different areas of the site (e.g., Cochard and Brugal 2004). These hypotheses can be tested to some extent by comparing survivorship rates of small mammals and medium ungulates in a given assemblage, and will be explored further in Chapter 6. In discussing survivorship of different bone elements, the issue of shaft versus endbased MNE is also addressed. It is assumed in this analysis that in most cases, when large animals were butchered in the field, carcasses were partitioned in such a way that most bone shafts were accompanied by at least one epiphyses when transported to the home base. Therefore, it is expected that accumulated remains, barring intense butchery or postdepositional attrition, should include equivalent ratios of proximal and/or distal long bone ends, and shafts, for each appendicular element. In certain cases, examining the ratio of long bone ends to shafts may indicate attritional processes of noncultural or cultural types. In general, based on structural density alone, a slight bias (1:2 or 1:3) of cancellous to compact elements may be expected (using values from Lyman 1984; Lyman 1994; see also values from Lam et al. 1999 in Stiner 2004). A more extreme ratio of 1:8 cancellous to compact element survivorship was estimated from bone ravaging experiments with hyenas (Capaldo 1997; Marean and Spencer 1991; Marean, et al. 1992), refitting studies of Paleolithic assemblages (Marean and Kim 1998), and measurements of limb shaft densities using adjusted computed tomography (Lam, et al. 1998; Lam, et al. 1999). Ratios of ends (based on the maximum from the proximal or distal end) to shafts were recorded for the different layers of Klissoura Cave 1. As with the other tests discussed here, only medium ungulates were 162 taken into account because they have the largest sample size through the sequence. The limb elements considered were the humerus, radius, ulna, metacarpal, femur, tibia and metatarsal. The suite of analyses discussed above can, taken together, provide insight into understanding and interpreting the condition of the faunal remains at Klissoura Cave 1. Once non-human taphonomic biases are taken into account, hominin prey selection, transport decisions, and butchery patterns can be more fully understood. RESULTS Input Analysis and Fragmentation An analysis of faunal input (or density) of Klissoura Cave 1 indicates that there is not a relationship between sample size and faunal input, once excavated volume is taken into account (Table 4.2, n = 14, rp = 0.326, p = 0.255). The layers with the smallest sample sizes (Mesolithic 3-5a, Epigravettian IIa-d, Early Upper Paleolithic V, and Middle Paleolithic X) also have some of the lowest densities of animal remains (<0.50 NISP/L), but larger samples from the Middle Paleolithic layers XI-XIV and XV-XVII also have extremely low rates of faunal input. Middle and Upper Paleolithic layers with comparatively moderate input rates (0.51-.90) include Upper Paleolithic layer III’, Middle Paleolithic layer XXa-XXb. Aurignacian layers (IV, IIIe-g, IIIb-d) and the non-Aurignacian Upper Paleolithic (III”) have the highest rates of faunal input, and there is a trend toward increasing density through time. Middle Paleolithic layers VIII and XVIII-XIX also have high faunal densities. In general, this follows the geoarchaeological evidence quite well (Karkanas 2010), with the largest anthropogenic signature from a faunal perspective occurring mainly in the earlier Upper Paleolithic layers, though use of the site or preservation clearly declines in later periods 163 (Table 4.2). Not surprisingly there is no correlation between input rate and time period (n = 14, rs = 0.033, p = 0.911) which suggests that there was not a steady increase in the use of Klissoura Cave 1 over the years; rather the intensity or duration of occupation reached an apex during the Aurignacian. Mean length and standard deviations for different body classes in each archaeological layer are reported in Table 4.3. Unsurprisingly, the average fragment size for each body class seems to increase as body size increases, but differences in sample sizes make comparisons between body class difficult to quantify. There is no temporal trend of fragmentation size for the overall sample (n = 14, rs = -0.103, p = 0.725) or medium ungulates, the most common prey type at the site (n = 14, rs = -0.402, p = 0.154); no change in fragmentation through the assemblages is apparent. Initial fragmentation was probably related to carcass butchery, but subsequent fragmentation can likely be attributed to trampling or shattering caused by high temperatures as hearths were built on bone-rich surfaces (see Chapter 6). The MNE/NISP fragmentation index is presented in Table 4.4. As with mean length, there is no trend between time and fragmentation index value (n = 13, rs = -0.249, p = 0.411). The index value from the Epigravettian layer (IIb-d) is excluded from the correlation because the sample is small. In addition to the lack of temporal trend for either indication of fragmentation, there is also no correlation between the length of medium ungulate specimens and fragmentation index (n = 13, rp = -0.039, p = 0.899). This means that layers with higher rates of fragmentation do not necessarily have the smallest specimens. 164 Table 4.3 Mean fragment length for different body classes, by cultural layer. Carnivore Damage and Weathering Evidence of carnivore activity is extremely rare in the Klissoura Cave 1 assemblages. A range of small and large carnivore species were identified in the faunas (Chapter 5), but it seems they did not use the site for shelter. No unfused or neonate carnivore remains were found, suggesting that the site was not used as a den. Three hyena coprolites, one in the nonAurignacian Upper Paleolithic (III”), the others in Middle Paleolithic layers (VIII and X) indicates that hyenas visited the cave, but this was an extremely rare occurrence. Only thirty- 165 Table 4.4 MNE:NISP index values for medium ungulate cranial elements and all elements in the Klissoura Cave 1 assemblage. *Note small sample sizes. five specimens from the entire assemblage have evidence of carnivore tooth puncture or drag marks, and these frequencies are far too low to elucidate any kind of trend (Table 4.5). Finally, ungulate body part profiles (Chapter 6) are similar to transport patterns of humans, not large carnivores (Stiner 1991, 1994). Incidences of rodent damage are even rarer than carnivore damage in the Klissoura Cave 1 assemblages, with just three specimens from all layers displaying rodent gnawing (Table 4.5). There is little evidence indicating that diurnal raptors or owls were major accumulative agents in the assemblages. Though several raptor species, specifically Aquila chrysaetos (Golden eagle), Tyto alba (barn owl), Bubo bubo (Eurasian eagle owl) and Asio sp. (eared owls) were identified in the Klissoura Cave 1 assemblages (Bocheński and Tomek 2010), their frequencies are exceptionally low (NISP<8 for the entire Upper Paleolithic and later periods). No cranial puncture marks or digestion damage was apparent on any of the 166 Table 4.5 Frequencies of gnawing and weathering damage on faunal specimens by cultural layer. 167 small game remains. It is possible that some of the micromammals were the remains of owl or raptor meals, but these small creatures were excluded from this study to be analyzed by specialists for environmental reconstructions. Additionally, hare body part analysis does not include a high frequency of hind or front limb bones (Chapter 6), which may be expected in an assemblage accumulated by avian predators. Tortoise remains have high incidences of burning and are mostly carapace and plastron fragments, consistent with human exploitation. Weathering was not a major taphonomic process at Klissoura Cave 1. Cracks and exfoliation that often accompany exposure to the sun are almost completely absent from the assemblages (Table 4.5). Chemical weathering was the most frequently observed process, which often made it appear that the specimens had been burned or etched by acid (Figure 4.3). This type of weathering was the most common in the Middle Paleolithic layers and is almost absent from the Upper Paleolithic and later layers, though overall it is still fairly rare (< 2.5% in any layer) (Table 4.5). Figure 4.3 Chemical weathering on Middle Paleolithic long bone fragment. From layer XIV. 168 Density-mediated attrition Ratios of ungulate tooth to bone-based cranial MNE are presented in Table 4.6 (Figure 4.4) by layer. All ungulate taxa were combined to create a more robust sample (Appendix C contains expanded data). Of the fourteen archaeological layers, four of the samples are too small (MNE<6) to draw conclusions related to density-mediated bias. Of the remaining layers, most have near-even ratios of tooth to cranial bone-based MNE, with a few exceptions. Aurignacian layer IV and Middle Paleolithic layers VIII have tooth-based MNE counts that are double the cranial bone-based MNE (Table 4.6). Middle Paleolithic layer VIII seems to have suffered from some density-mediated attrition based on bulk bone density tests (see below), but there is no explanation for Aurignacian layer IV. One possibility is that crania were more smashed up in this layer. This should be testable, however, by comparing the fragmentation index for cranial parts (Table 4.4) with the proportion of tooth to cranial bone-based MNE (Table 4.6). A higher tooth to bone-based MNE ratio reflects poor survivorship of cranial bones, and a low index value for cranial fragmentation (MNE/NISP) indicates a higher degree of breakage. Therefore, it might be expected that tooth to cranial bone-based MNE ratios should correlate negatively with the fragmentation index if increased processing or post-depositional fragmentation were to blame for an under-representation of cranial bones in the assemblages. This is not, however, the case at Klissoura Cave 1; a Pearson’s correlation indicates no relationship between the cranial fragmentation index and the proportion of tooth to cranial bone-based MNE (n = 9, rp = -0.283, p = 0.461). It seems unlikely, therefore, that high the proportion of tooth to bone-based MNE in Aurignacian layer IV is due to fragmentation of cranial bones, and this anomaly cannot be readily explained. The non-Aurignacian Upper Paleolithic layer III” and Middle Paleolithic layers 169 XVIII-XIX have much higher bone than tooth-based MNE counts. This may relate to greater tooth fragmentation rates in these layers based on excavation damage or post-depositional processes, but this is unclear. Neither of these layers have abnormally high fragmentation rates overall (Table 4.4). Table 4.6 Tooth and bone-based MNE for all ungulates by layer. Figure 4.4 Proportion of tooth to bone-based MNE for all layers at Klissoura Cave 1. Data from Table 4.6. 170 Turning to bulk density, the results of a Spearman’s rank-order correlation between bone density values and percent survivorship for hares and fallow deer/medium ungulates from each layer at Klissoura Cave 1 are presented in Table 4.7. Note that the Epigravettian (IIb-d), Early Upper Paleolithic (Uluzzian layer V) and Middle Paleolithic (X) are excluded from the comparison because of small samples for both hare and fallow deer. Likewise, fallow deer values from the Mesolithic (3-5a), and hare values from the non-Aurignacian Upper Paleolithic industry (III”) and the entire Middle Paleolithic are unavailable due to the low frequency of hares in these assemblages. The hare remains from the Mesolithic (3-5a) and Mediterranean backed-bladelet industry (III’) display a significant, positive relationship between bone density and percent survivorship. The rs2 values, however, indicate that density-mediated attrition has the potential to explain only eight to sixteen percent of the variation in skeletal survivorship from these levels (Table 4.7). For fallow deer in the Upper Paleolithic and later layers, only the middle Aurignacian (IIIe-g) indicates a significant positive relationship between percent survivorship and bone density, and here densitymediated attrition can explain no more than five percent of the variation in deer body part representation (Table 4.7). In addition, hare survivorship does not display a correlation with bone density for this layer, even though hare remains are on a whole more fragile than ungulate bones. In the Middle Paleolithic there is a positive correlation between percent survivorship and bone density values for fallow deer in all of the layers, and rs2 values indicate that density-mediated attrition has the potential to explain 6-26% of the variation in skeletal survivorship in these layers (Table 4.7). Maximum MNE based on long bone ends and shafts for selected long bones from medium ungulates in the Klissoura Cave 1 assemblages are presented in Appendix D. 171 Table 4.7 Spearman’s rank-order correlation between percent survivorship and bone density values for European hare and fallow deer/medium ungulate remains. Data from Appendices A and B. *Significant values Values for different elements are collapsed into all long bone ends and shafts in Table 4.8 and are plotted in Figure 4.5. A 1:1 ratio is expected when shafts and ends are equally wellpreserved because values are used for the end with the highest MNE. In almost all layers, except for the Middle Paleolithic (XVIII-XIX), MNE values for long bone shafts and ends are either equal, or end-based MNE is much higher than shaft-based MNE. A closer look at the data reveals that most of the layers with higher end-based MNE values are from the 172 Upper Paleolithic layers, while the Middle Paleolithic layers have nearly a 1:1 ratio between shaft and end based MNE (Table 4.8). Potential explanations for this are discussed below. Table 4.8 End and shaft-based MNE values for medium ungulate long bones by cultural layer. Elements include humerus, radius, ulna, metacarpal, femur, tibia and metatarsal. Figure 4.5 Proportion of shaft-based and end-based MNE for all layers at Klissoura Cave 1. Data from Table 4.8. 173 DISCUSSION AND CONCLUSIONS Several interesting points concerning site use and the condition of faunal remains at Klissoura Cave 1 are apparent in the results presented above. Perhaps the most striking feature is the extent to which the faunas were primarily accumulated and modified by Pleistocene hominins. This bodes well for interpretations of small game use and human transport and butchery patterns presented in later chapters. Fragmentation rates of faunal materials indicate no temporal trend, so they cannot be taken to indicate an increase or decrease in the intensity of processing activities or other sources of mechanical damage over time. The high degree of fragmentation at Klissoura Cave 1 is probably largely due to postdepositional factors, such as trampling or compacting by sediments, though relationships between fragment size and butchery intensity will be explored in Chapter 6. In terms of site occupation intensity, rates of faunal input support geoarchaeological conclusions set forth by Karkanas (2010), who interprets the Middle Paleolithic, Epipaleolithic and Mesolithic sediments as having a higher natural input and the Upper Paleolithic layers as having a more intense anthropogenic signature. Occupation intensity can refer to the duration of time a site was used, how often it was used, or how many people were utilizing the site at a given time. Distinguishing between these patterns of use in the archaeological record is difficult, particularly because an increase in occupation intensity could represent something as minor as ten people as opposed to five using a site for a week, or a site being occupied for weeks rather than days. In this context, it simply means a general increase in site use, which could encompass any of these behaviors. According to faunal density data, the site was occupied most intensively during the Aurignacian and the overlying non-Aurignacian Upper Paleolithic (III”). The Aurignacian layers (IV, IIIe-g) of Klissoura 174 Cave 1 also contain dozens of clay-lined hearth features, and a possible rock-delineated structure (Karkanas, et al. 2004; Koumouzelis, Ginter, et al. 2001; Pawlikowski, et al. 2000; Stiner 2010). There was a subsequent decline in occupation intensity, particularly following the depositional hiatus of the Last Glacial Maximum (Karkanas 2010). This decline may indicate the use of different kinds of sites or different uses of the landscape in later periods, perhaps even other caves in Klissoura Gorge. The scarcity of carnivore remains, evidence for carnivore denning, and modification of the faunas is striking, as it is clear that carnivores were present and active on the late Pleistocene landscape in Southern Greece (Gamble 1986; Payne 1975; Reisch 1976; Stiner and Munro 2011; Tsoukala 1992). It is possible that the Klissoura Gorge with its many caves provided ample options for sheltering carnivores, so that the occasional human presence at Klissoura Cave 1 was enough to discourage carnivores from using the site. Another option is that carnivores used the deeper, narrower parts of the cave for shelter, which is an area that was not included in the excavation. If this were true, we might still expect to see appreciable frequencies of carnivore damage on the archaeological assemblages, which is certainly not the case. Evidence for modifications made by birds of prey are also rare in the faunas, as is rodent gnawing. Surface weathering of faunal remains was not a factor in the Klissoura Cave 1 assemblages. Also, there is no indication of differential bone loss in specific parts of the site (see Chapter 7) so chemical diagenesis was probably not a major problem either. Results of tests evaluating density-mediated attrition are somewhat mixed, though evidence of density-mediated bias does not indicate the underlying cause (i.e. human or nonhuman) of attrition. Overall, deer bone survivorship is uncorrelated or only weakly correlated with bone density in the Upper Paleolithic layers. The cranial bone-based MNE nonetheless 175 is much lower than expected as compared to tooth-based MNE for Aurignacian layer IV; this discrepancy cannot be explained by in situ attrition or heavy fragmentation of cranial parts. Hare bone survivorship does correlate with bone density in the Mediterranean backedbladelet industry (III’) and above, so there may be some density-mediated processes at play in these later layers. A lack of correlation between deer/medium ungulate bone survivorship and bone density in layer III’, however, suggests that any density-mediated processes in the upper layers may have only affected the structurally least dense or fragile small elements, in this case hare remains. A discussion of body part representation in Chapter 6 will allow us to evaluate some of the potential variation in bone survivorship as a product of human behavior. Correlations between percent survivorship and bone density for ungulate remains in the Middle Paleolithic elicit several questions concerning game use in these earlier periods. 1) Are fallow deer elements missing in the Middle Paleolithic layers because of densitymediated attrition or transport decisions by human foragers? 2) Are hare remains lacking in this lower layers of Klissoura Cave 1 because they were not exploited by Middle Paleolithic hunters, or are they absent due to attritional processes? First, it must be pointed out that there is only a bias in tooth to cranial bone-based MNE in the latest Middle Paleolithic layer (VIII) (Table 4.6). At the same time, there is actually a greater representation of cranial bone-based MNE compared to tooth-based MNE in Middle Paleolithic layer XVIII-XIX. It seems likely, then, that though there is some correlation between body part representation and skeletal density in the Middle Paleolithic, some of the differential representation of anatomical parts was influenced by human transport decisions or butchery patterns, which can also relate to density-mediated processes (see Chapter 6). As for hare remains in the Middle Paleolithic layers, if density-mediated attrition was a factor, we would expect to find only the most 176 dense elements. Despite the fact that sample size is small, hare remains that are present in the Middle Paleolithic layers span an entire range of structural densities, including many that are far less dense that any fallow deer elements (Appendix E). An interesting and rather counterintuitive bias in the MNE of under-represented long bone shafts versus over-represented ends was observed in several of the assemblages, particularly in the Upper Paleolithic layers. The most likely explanation for this is the macroscopic condition of the Upper Paleolithic remains. Most of the Upper Paleolithic faunas were covered with an ashy layer that obscured many macroscopic markings on the specimens. It is likely that this same ash also hid diagnostic shaft features, such as nutrient foramina and muscle attachments. This provides yet another argument in favor of relying on both long bone shaft and end features for analyzing zooarchaeological assemblages, which is already done by most faunal analysts (Bar-Oz 2005; Binford 1978; Brain 1981; Bunn and Kroll 1986; Delpech 1998; Morlan 1994; Rogers 2000; Stiner 1991, 1994, 2005; Todd and Rapson 1988), because it is never clear at the outset what kind of analytical biases may be encountered in an assemblage. Though some non-human taphonomic processes affected the faunas of Klissoura Cave 1, they were very minor overall. Based on the discussion outlined in this chapter it seems unlikely that broad patterns of human behavior related to changes in species use, transport decisions or butchery patterns were obscured by non-human taphonomic processes acting on the assemblages. 177 CHAPTER 5: RELATIVE SPECIES ABUNDANCE AND DIVERSITY OF PREY TYPES INTRODUCTION Changes in human subsistence strategies as reflected by prey choice over a long archaeological sequence are informative of changing environmental conditions, shifts in human demography, or as we see at Klissoura Cave 1, both. Using multiple methods, including analyzing prey groups based on body size and flight behavior (following Stiner, et al. 2000), tests of species diversity, ratios of prey types, biomass comparisons, and input analysis (see Chapter 4), this chapter evaluates diachronic change in prey choice from the Middle Paleolithic through Mesolithic at Klissoura Cave 1. Following predictions laid out in Chapter 3, it is expected that some of the changes in the species representation of large game animals, as well as great bustard, track environmental changes. Conversely, shifts in the overall composition of the small game component, and changes in the proportion of large and small game through time, relates more closely to increased occupation intensity at the site or in the region as a whole. METHODS Several different approaches are employed to evaluate prey choice over time at Klissoura Cave 1. Number of identified specimen (NISP) counts for all species identified in the eight Upper Paleolithic and later layers and six Middle Paleolithic units are presented (Appendix F). The taxa are divided into eight broad prey categories based on body size and behavioral characteristics; there are four different size classes for ungulates, small and large 178 carnivores, small slow-moving game, and small fast-moving game (Table 5.1). It is predicted that human foragers would rank animals according to differences in their handling costs and return rates, with large-bodied ungulates being preferable (see Chapter 1, Table 5.1). Carnivores are considered separately because they are behaviorally different than other small and large game, and are very uncommon at Klissoura Cave 1 though they represent human prey (as opposed to dying naturally at the site). The small, slow-moving category includes tortoises (Testudo sp.), which are easy to collect by foragers of all ages and conditions. The small, fast-moving game category includes hares (Lepus europaeus) and birds of various sizes (most notably great bustard, Otis tarda, and rock partridge, Alectoris graeca), species that tend to require a significant technological investment (e.g. snares or nets) in order to be exploited efficiently (Cannon 2000; Jones 2006; Madsen and Schmitt 1998, but see also Ugan 2005 for differences between terrestrial game, fish and invertebrates). Class designations including small birds (<0.1 kg) and indeterminate snakes were excluded from the groupings because it is not certain that they were exploited by humans; however, they exist in such low frequencies (NISP<15 in the entire assemblage) that their inclusion would have little effect on these interpretations. The four major small game species found at Klissoura Cave 1 have very different life history characteristics, habitat requirements, and flight responses; the bustard is further distinguished by its large size (see Table 5.1). Tortoises have low capture costs and high return rates because they are slow-moving; therefore, they are expected to be preferred prey when available (Stiner, et al. 2000). They are a slow-growing species, taking a long time to reach reproductive maturity and continuing to grow throughout their lifetimes. They are sexually dimorphic in terms of size, with females tending to be larger (Blasco, et al. 1986-87; 179 Hailey, et al. 1988; Lambert 1982; Swingland and Stubbs 1985). Therefore, the most obvious individuals on the landscape, and the most desirable for human collection, are the large, older females that are the reproductive base of the population (Stiner, et al. 2000). Climatic variation would never have been great enough to eliminate Mediterranean tortoises from southern Greece; they are found in human diets throughout much of the Middle and Late Table 5.1 Taxa included in each body class category and their respective weight ranges and predicted ranking. Mass values from Nowack (1999) and Silva and Downing (1995). 180 Pleistocene in the eastern Mediterranean and Middle East (Reynaud Savioz and Morel 2005; Speth and Tchernov 2002; Stiner 2005, 2009; Stiner, et al. 2000). Therefore, a decrease in tortoises in an archaeological sequence may be the result of overexploitation by human foragers. Though they are categorized as small, fast-moving animals, great bustards potentially have higher return rates than either hares or partridges because they are appreciably larger (Table 5.1). Like tortoises, however, bustards have low rates of reproduction and development and rely disproportionately on older females to perpetuate the species (Morales, et al. 2002). Females over six years of age are twice as successful at reproducing as younger birds, even then they average only 0.40 chicks per year and are unlikely to breed two years in a row (Morales, et al. 2002). Successful individuals tend to be long-lived and population turnover rates are comparatively low; in this sense bustard populations are similar to tortoises in that they are easily stressed with the removal of a few older females. Heavy exploitation of either of these taxa could not be sustained for long. Rock partridge and European hare likely have similar return rates to one another, which are lower than the other two small game taxa, because they have rapid flight responses when confronted by a predator. Both species are prolific and are able to reproduce in the first year of life. Partridges have yearly clutch sizes of between twelve and eighteen eggs (Vavalekas, et al. 1993), and hares give birth to one to four litters per year of up to five leverets (Burton 1991). Their productivity as modeled against Mediterranean tortoises can be as much as seven times greater (Stiner, et al. 2000). 181 Examining richness and evenness (i.e. diversity) in a faunal assemblage is a common way to evaluate relative species abundance and to control for sample size differences in zooarchaeological assemblages (Grayson and Cannon 1999; Grayson and Delpech 1998; Nagaoka 2001). There are many tools for analyzing species diversity, and various terms have been used in different ways in ecological and anthropological literature (see Lyman 2008:175 for a review). In this presentation, taxonomic richness refers to the number of taxa identified to the genus level, such that an assemblage with ten different taxa is richer than an assemblage with eight provided that sample sizes are similar. Genus is used rather than species because the species is not altogether certain for some taxa such as Capra (though it is most probably C. ibex). Additionally, there may have been up to three species of similarsized tortoises (Testudo graeca, T. hermanni, and T. marginata) in southern Greece and the remains could only be identified to Testudo sp. In the case of diversity indexes, the level at which taxonomic identifications are made is less important than maintaining consistency with the taxonomic level at which they are tallied (Lyman 2008:174). A related method to evaluate species abundance is taxonomic evenness, which measures how individuals distribute across different categories, or taxa (Smith and Wilson 1996). Taxonomically even assemblages have a similar number of individuals for each of the different genera. One technique for evaluating the evenness of archaeofaunas is using the reciprocal of Simpson’s Index (Grayson and Delpech 2002; James 1990; Jones 2004; Schmitt and Lupo 1995; Stiner 2001). It is calculated as 1/D, or 1/Σ(ρi)2, where ρ is the proportion of taxon i in the total assemblage (Levins 1968; Simpson 1949). A lower value for 1/D indicates the disproportionate representation of one taxon. This simple index evaluates species richness and evenness simultaneously and corrects for sample size differences among layers in an 182 archaeological assemblage. Prey diversity, as it relates to the input analysis introduced in Chapter 4, is also examined. Presenting the minimum number of individuals (MNI) is a common technique in faunal analyses (Grayson 1984). Like many useful zooarchaeological counting units, it is derived from NISP, though it suffers more than any other commonly used unit from aggregation error; MNI values fluctuate depending on how stratigraphic or cultural units are grouped, which can have a major effect on how an assemblage is interpreted (Grayson 1984; Lyman 2008:46). Another problem with MNI is that it tends to overestimate the importance of rare species in an assemblage. Perhaps the most disconcerting problem with MNI from the perspective of someone seeking to understand food transport patterns is the assumption that carcasses, particularly those from large game animals, were brought to a site whole. An entire body of ethnographic literature (see Chapter 1 for a review) tells us that this is not necessarily the case. In this study, minimum number of elements (MNE) is preferred for large mammal biomass calculations, as discussed in Chapter 4. MNI is not used in most of the analyses below (particularly for large game), though MNI values for each of the species found at Klissoura Cave 1 are presented in Appendix F. Minimum number of individuals is used in this chapter for some biomass calculations (discussed below), because it is a fair assumption that small game animals were brought to the site in their entirety, provided that the study samples are suitably large. A biomass comparison allows for a rough estimation of the amount of meat utilized by foragers occupying Klissoura Cave 1. Often, biomass is calculated based on the mass (weight) of archaeological bone for each species or group of species in an assemblage (Reitz and Cordier 1983; Reitz, et al. 1987). This method was not employed at Klissoura Cave 1 because the 183 lowest Middle Paleolithic layers are heavily mineralized, and bones in certain Upper Paleolithic layers are covered with thick concretions. As mentioned above, calculations of small game biomass are based on the assumption that the animals were transported to the site whole, and are derived by multiplying the average live mass of a given species by the MNI for that taxon in each stratigraphic layer (Appendix G). Estimated biomass values for large game species do not rely on the assumption that entire carcasses were brought to the site, and are calculated by multiplying MNE for each taxon in a stratigraphic layer by a standardized value of live mass of a given species divided by the number of elements diagnostic to species level in a given carcass (Appendix G). For example, fallow deer have 55 elements that can be identified reliably to species (this excludes axial elements and carpal/tarsal complex bones that may not be recovered archaeologically). The average live weight of an adult fallow deer is approximately 70 kg, so the standardized value is 70/55 elements, or 1.27 kg per element. It must be stressed that this biomass calculation is only a rough indicator because it includes bone and other inedible tissues, ignores variation in individual body size, and assumes that edible meat is spread evenly over an ungulate body, which it clearly is not. These factors will be examined in Chapter 6 with a focus on the most commonly occurring species in the assemblage. For the purposes of this chapter, however, MNE-based biomass estimations can be used on a broad basis because they standardize NISP for body size and give a more reliable indication of which species provided the bulk of the meat diet. Species represented by an NISP<5 for a given stratigraphic layer are excluded from this analysis. PREY REPRESENTATION IN THE MIDDLE PALEOLITHIC THROUGH MESOLITHIC LAYERS AT KLISSOURA CAVE 1 184 The Klissoura Cave 1 faunal sequence spans much of the end of the Pleistocene, from about MIS stage 5a until the final deglaciation. This time interval represents variable climatic conditions that supported a wide range of species near Klissoura Cave 1. Ungulate species available in Klissoura Gorge, as well as the adjacent upper Berbati Valley and Argive Plain, included European fallow deer, ibex, European wild ass, aurochs, red deer, and wild pig. Roe deer and chamois occur in low numbers in Aurignacian layer IV, and roe deer is present again in the Mesolithic. Small game animals occur throughout the sequence, with tortoises being common in the Middle Paleolithic and Early Upper Paleolithic. European hare occurs in low frequencies in the Middle Paleolithic, particularly in layers XI-XIV, but is ubiquitous in the Upper Paleolithic and later assemblages, when birds such as rock partridge and great bustard were increasingly important. Large and small carnivore remains are found in low frequencies throughout the sequence. They include red fox (Vulpes vulpes), wild cat (Felis silvestris), Eurasian lynx (Lynx lynx), and martens (Martes martes or M. foina). Large carnivore remains are mainly leopard (Panthera pardus), along with scant remains of wolf (Canis lupus) in some layers, and a single fragment of brown bear (Ursus arctos) in the Middle Paleolithic (layer XVIII). Three hyena coprolites were recovered, one in the Upper Paleolithic, the other two in Middle Paleolithic layers, indicating that hyenas visited the cave, but their presence was ephemeral. There are absolutely no indications of carnivore dens at the site. A possible shaft fragment from a hominin (Homo sp.?) ulna was found in the lowest Middle Paleolithic layer that contained fauna (XXa, XXa1, XXb). Faunal assemblages from different stratigraphic layers at Klissoura Cave 1 vary in size. Most notably, the Epigravettian (IIa-d) contains a small sample. The Mesolithic (3-5a), Early Upper Paleolithic in layer V and Middle Paleolithic layer X are also fairly small. Input 185 analysis (Chapter 4) indicates that sample size does not necessarily reflect occupation intensity, for example the lower Aurignacian (layer IV) and the Middle Paleolithic layers XVIII-XIX have a similar input rate (measured in NISP count per liter), though the former has a total sample size of over 4,000 and the latter is around 2,500 (Table 4.1). Tests to correct for variable sample sizes are undertaken below, and it is still possible to reconstruct aspects of subsistence from the layers yielding small samples. Relative species abundance (NISP) through the sequence at Klissoura Cave 1 is presented in Figure 5.1 (from Appendix F). Several trends are apparent in the data. During the Middle Paleolithic, fallow deer was the most commonly exploited species at the site. There is a period of increased prey diversity (evenness) in the Early Upper Paleolithic (V) and lower Aurignacian (layer IV). Also during the Early Upper Paleolithic (V), small game becomes a more important part of the diet and increases in frequency until the Epigravettian (IIa-d) when it actually overtake fallow deer by NISP count. The trends become even clearer when the Linnaean taxonomic groups are collapsed into the prey categories outlined above. Figure 5.2 shows the abundance by NISP of the eight prey categories. The medium ungulate category dominates for the majority of the sequence, though large ungulates attain greater importance during Aurignacian layer IV. Small, fast-moving game becomes the most common prey category by count during the Epigravettian (IIa-d) and Mesolithic (3-5a). Taking this analysis a step further, it is possible to compare changes in the proportion of large game to small game (combined slow and fast prey categories), as well as changes in the proportion of small, slow-moving game to small, fast-moving game through the sequence. 186 Figure 5.1 NISP counts for the major taxa at Klissoura Cave 1 by layer. Taxa are in descending order of average mass. 187 Figure 5.2 NISP counts for the major taxa at Klissoura Cave 1, grouped by body class. 188 Table 5.2 shows the proportion of small game to ungulate NISP in the Klissoura Cave 1 sequence. The proportion of small game is graphed as a scatter plot against time in Figure 5.3. Warm, mild, and cool phases following the marine oxygen isotope curve (MIS) (see Chapter 3) are indicated at the top of the graph. As no dates are currently available for the Middle Paleolithic, points are spaced evenly on the X-axis until MIS 5a, which is thought to be the stage during which the earliest cultural layers were deposited at the site (Karkanas, personal communication). Using a Spearman’s rank-order correlation, the relationship between time period and the proportion of the total assemblage that is composed of small game is significant and positive (r = 0.754, p = 0.002, n = 14); small game comprise a greater part of the diet at Klissoura Cave 1 by NISP in the later time periods. Table 5.2 Proportion of NISP counts per layer for small game: large game (ungulates), and small slow: small fast-moving prey types. *Small samples excluded from Figure 5.3. The relative frequency of small game is plotted by period in Figure 5.4, with the two earliest Middle Paleolithic layers, as well as Middle Paleolithic layer X excluded due to the small sample sizes for small game (NISP<25). Small, fast-moving animals including the 189 Figure 5.3 Proportion of small game to ungulate NISP, plotted against time. The relationship is significant and positive. Data from Table 5.2. European hare, rock partridge, and great bustard are depicted using dark shades of grey and black to indicate their similar handling costs, while tortoises are differentiated by white bars. Hares were exploited in all periods but their importance increased with time after Early Upper Paleolithic (V), though they are also common in Middle Paleolithic layer XI-XIV. Conversely, tortoises were a significant food item only in the Middle Paleolithic and Early Upper Paleolithic, and are completely absent from the assemblages by the Epigravettian (IIad) (Figure 5.4). Rock partridges and great bustard gained importance in Aurignacian layers (IIIb-d) and in the period defined by the Mediterranean backed-bladelet industry (III’). The most productive small prey, hares and partridges, were used heavily by the Upper Paleolithic and Mesolithic hunters, except during the Early Upper Paleolithic (layer V). 190 Figure 5.4 Relative NISP counts for small game animals by layer. White fill indicates small, slowmoving species, grey or black fill indicates small, fast-moving game. In general, Figure 5.4 indicates an overall shift from small, slow-moving prey species in the Middle and Early Upper Paleolithic to small, fast-moving animals in the Upper Paleolithic and later layers. The proportion of small, slow-moving game is graphed as a scatter plot against time period in Figure 5.5 (data from Table 5.2), again removing the layers with the small samples mentioned above. A Spearman’s rank-order correlation between time period and the proportion of the small game fraction that is composed of slow-moving prey is significant and negative (r = -0.952, p = 0.000, n = 11). Small, slow-moving game animals were a greater part of the diet at Klissoura Cave 1 by NISP in the earlier time periods, and were replaced by fast-moving small species in the later part of the sequence. Also, earlier time periods in general display more variability in the small game animals that were exploited; by the Aurignacian and later layers tortoises drop out of the assemblages and small, fast-moving game animals were the exclusive focus. 191 Figure 5.5 Proportion of small slow to small fast game NISP, plotted against time. The relationship is significant and negative. Data from Table 5.2. Carnivores make up a small proportion of the Klissoura Cave 1 faunal series. Taxa with an NISP>1 for the entire assemblage are depicted in Figure 5.6. One specimen identified as brown bear was found in Middle Paleolithic layer XVIII and a single stone or pine marten specimen was recovered from Aurignacian layer IV (Appendix F). The most common carnivore species in the series are wild cat, red fox and leopard, though their frequencies are always low (NISP<25). These taxa may have been utilized for meat and fur, though their occurrence is low. It is notable that none of the carnivore remains are from young animals, so it is unlikely that any of the remains are the result of denning behavior (see Chapter 4). In addition to vertebrate taxa, land snails were exploited during the Upper Paleolithic at Klissoura Cave 1 (Starkovich and Stiner 2010). No quantitative account of land snails are 192 Figure 5.6 NISP counts for carnivore species at Klissoura Cave 1 by cultural layer. available, but a sample from squares AA4-BB4 was analyzed by M. C. Stiner. The most common taxon is Helix figulina, a large edible snail found in southeastern Europe. Land snail shells are not found in the Middle Paleolithic layers, and they are rare in the early Upper Paleolithic. Shells from the Early Upper Paleolithic and Aurignacian series IV-IIIg show no evidence of human modification. Shell size in these layers varies, and some shells have small perforations from predators. The number of land snail shells in the deposits increases with time. They are moderately abundant in Aurignacian layers IIIb-d and increase greatly in the Mediterranean backed bladelet industry (III”). Snail shell frequencies are low in the Epigravettian (IIa-d), like other faunal remains, and peak in the Mesolithic (3-5a). Land snail shells from the younger layers are biased toward large individuals (2.3-2.4 cm modal diameter), and display high frequencies of broken aperture rims (75-95%). No snail shells are burned. It is unclear exactly when land snails start to be commonly exploited at Klissoura 193 Cave 1, though their use is confined to the Upper Paleolithic and later periods. Land snails are considered a low-return resource because, though they are easy to collect, they are extremely small and labor intensive to process. The overall increase in small game exploitation and the specific targeting of small, fast-moving prey species and small invertebrates such as land snails indicates a greater reliance on lower-ranked resources in the later layers at Klissoura Cave 1. This may correspond with human population growth in the region or a more intensive use of the landscape immediately around Klissoura Cave 1 in the later Upper Paleolithic and Mesolithic periods. On the other hand, variation in the frequency of great bustards and certain large ungulates may reflect climate-driven changes in environmental conditions (see below). PREY DIVERSITY: RICHNESS AND EVENNESS Taxonomic richness often correlates with sample size (e.g. Grayson 1984; Leonard 1989; Sharp 1990), which simply means that larger faunal assemblages tend to contain more unique or rare species. In evaluating whether or not this is the case at Klissoura Cave 1, logNISP is plotted versus logNtaxa (Figure 5.7, values from Table 5.3). Indeed, a Pearson’s correlation indicates that there is a very significant, positive relationship between the two variables (r = 0.912, p = 0.000, n = 14). Because sample size explains > 80% of variation, richness (N-taxa) values at the species level are not useful at Klissoura Cave 1 for understanding taxonomic diversity in the assemblages because the largest samples contain a greater number of species. 194 Figure 5.7 Ratio of logNISP to logNtaxa in the Klissoura Cave 1 assemblage. The relationship is significant and positive. Data from Table 5.3. Taxonomic evenness also sometimes correlates with sample size (Grayson and Delpech 1998; Grayson, et al. 2001; Nagaoka 2001, 2002b). A Pearson’s correlation between evenness (1/D of all taxa) and logNISP (all taxa) in Klissoura Cave 1 is not significant (r = 0.165, p = 0.573, n = 14) (data from Table 5.3 and 5.4), which indicates that sample size is not driving differences in the evenness index. Potential relationships between sample size and evenness for the ungulate portion of the assemblage as well as the small game component are also tested; no relationship is found in either of these cases (r = -0.395, p = 0.162, n = 14 for ungulates; r = 0.227, p = 0.503, n = 11 for small game). 195 Table 5.3 LogNISP and logN-taxa values for all species, ungulates, and small game species. 196 Table 5.4 Inverse Simpson’s index per layer for all taxa, ungulates, and small game species. *Small sample sizes excluded from Figure 5.10. In order to evaluate if there are any trends in prey evenness at Klissoura Cave 1, the reciprocal of Simpson’s Index for all taxa in each layer is plotted against time in Figure 5.8 (Table 5.4). There is a general tendency for an increase in diversity of species exploited from the Middle to Upper Paleolithic and later periods; this reflects the dominance of fallow deer in the Middle Paleolithic giving way to greater proportions of small game species at the end of the Upper Paleolithic. However, the relationship between evenness and time is not significant (r = 0.477, p = 0.085, n = 14). The outlying point in the Upper Paleolithic represents Aurignacian layer IV, where there was a more even representation of many largebodied ungulates shown in Figure 5.1. Much of the variability in evenness occurs in MIS 3, which experienced many climatic fluctuations and shifts in vegetation (see Chapter 3). It is possible that these rapid shifts are reflected to some degree in the fluctuating diversity of the Klissoura Cave 1 assemblage. 197 Figure 5.8 Inverse Simpson’s index for all taxa plotted against time. No statistically significant relationship exists, but there is a general increase in variability between assemblages in the Upper Paleolithic. Data from Table 5.4. No significant trends in taxonomic evenness are apparent in either the ungulate prey (r = 0.385, p = 0.175, n = 14) or small game species (r = 0.342, p = 0.304, n = 11) at Klissoura Cave 1 (Figures 5.9 and 5.10). In general, evenness values are low in the Middle Paleolithic, while in the Upper Paleolithic and later layers values are variable, which may be climatically driven, or indicate more flexible foraging strategies. The lower Aurignacian (IV), Epigravettian (IIa-d) and Mesolithic (3-5a) occupations exhibit greater degrees of ungulate evenness, which may reflect environmental conditions that will be explored further below. Small game evenness sorts into two groups, one set with higher diversity, the other with lower. The three points in Figure 5.10 that are higher diversity likely reflect two different processes. One of the points represents the Early Upper Paleolithic (Uluzzian layer 198 Figure 5.9 Inverse Simpson’s index for all ungulate taxa plotted against time. No statistical relationship is apparent, though certain Upper Paleolithic and later assemblages are significantly more even than the other layers. Data from Table 5.4. V), which has a high proportion of both tortoise and hare. The other points in Figure 5.10 represent Upper Paleolithic layer III” and Aurignacian layer IIIb-d, both of which have higher frequencies of great bustard and rock partridge (Figure 5.4). The latter set of points is likely a reflection of drier climates and open vegetation in the vicinity of Klissoura Cave 1. The declining importance of tortoises, on the other hand, cannot be explained by climatedriven changes in the presence of these species in southern Greece. The decrease in tortoise numbers in relation to hares corresponds with a shift from high ranked small, slow-moving species to low ranked small, fast-moving species at the site. 199 It is possible that changes in taxonomic evenness are simply a result of changes in site occupation intensity. For example, more prey species that are rare in the environment may appear in an assemblage as human hunters spend more time at a site, resulting in a negative correlation between input and evenness. This is not the case at Klissoura Cave 1, as the relationship between evenness and input is not significant for all taxa (r = -0.007, p = 0.980, n = 14). Likewise, there are no relationships between ungulate or small game evenness and faunal input (r = -0.335, p = 0.241, n = 14 for ungulates, r = 0.184, p = 0.587, n = 11 for small game). Layers with low evenness values are not necessarily those that have higher inputs of fauna; rather, evenness may relate to environmental factors or prey choice. Figure 5.10 Inverse Simpson’s index for small game species plotted against time. No temporal trend exists, but note the cluster of points that represent higher small game diversity. The earlier point indicates a more even representation of hares and tortoises, while the later points indicate layers with an even representation of hares and birds (great bustard and rock partridge). Data from Table 5.4. 200 BIOMASS COMPARISON Analyzing prey biomass provides a method for evaluating the importance of prey species with different body sizes to forager diets. It goes beyond NISP counts, which are valuable for understanding changes in prey representation, but does not necessarily account for difference in the body size of prey being exploited by past hunters, or for differential fragmentation rates between animals of different sizes (see Chapter 4). For example, the NISP of hare and fallow deer may each be 50 for a given assemblage. This might result in an MNI of three hares and one fallow deer (for this example, assume that the animals were brought to the site whole). Even though the number of hares in the assemblage is greater, the contribution of fallow deer meat to the diet is greater because the animal is much larger than even three hares (Table 5.1). The biomass contribution of each major prey taxon at Klissoura Cave 1 is presented by layer in Figure 5.11, and is arranged similarly to Figure 5.1. Once again, fallow deer is the most important ungulate taxon by mass for the majority of the sequence, with the exception of Aurignacian layer IV, and small game seems to increase in importance by the late Upper Paleolithic (Figure 5.11). A major difference between the figures, however, is that the importance of other ungulate species in most of the Middle Paleolithic layers, and some Upper Paleolithic layers, is much more apparent when biomass is considered (Figure 5.11). The portion of the diet comprised of small game is much more modest using this method, even in the Epigravettian (IIa-d) and Mesolithic (3-5a) assemblages that are dominated by small game NISP counts (Figure 5.11). Temporal changes in the relative proportion of different groups of species at Klissoura Cave 1 remain significant when considered in terms of biomass. The proportions of 201 Figure 5.11 Estimated biomass for each important taxa at Klissoura Cave 1 by layer. Note that small game biomass is calculated based on MNI, while large game biomass is calculated using MNE. 202 ungulate to small game (all prey types) biomass are listed in Table 5.5. The proportion of small game biomass is plotted against time in Figure 5.12. A Spearman’s rank-order correlation indicates that the relationship between small game biomass and time is significant and positive, excluding the Epigravettian (IIa-d) outlier (r = 0.669, p = 0.012, n = 13). Changes in the proportion of small game biomass mirror changes in the comparisons based on NISP, with small game becoming increasingly important through time. Despite the rising importance of small game in the diet, it must be noted that small game never exceeded so much as a quarter of the total meat diet by these calculations, with the exception of the Epigravettian period (IIa-d), where small game comprises more than half of the meat biomass (Figure 5.12). Figure 5.12 Proportion of small game to ungulate biomass, plotted against time. The relationship is significant and positive. Data from Table 5.5. 203 Table 5.5 Proportion of biomass values per layer for small game: large game (ungulates), and small slow: small fast-moving prey types. *Small samples excluded from Figure 5.13. Considering only the small game component and dividing prey by evasion tactics, it becomes clear that changes in relative biomass mirror those of NISP counts (Table 5.5, Figure 5.13). Figure 5.13 again excludes the three Middle Paleolithic layers with small sample sizes. In general, there was a shift from small, slow-moving species to small, fastmoving species by meat weight through the sequence at Klissoura Cave 1 (Figure 5.13). The relationship is significant and negative (r = -0.915, p = 0.000, n = 11). There are subtle differences between the results based on NISP and those based on biomass, but the larger conclusions are the same. Specifically, changes based on biomass show the importance of a range of ungulate species in some layers, even if fallow deer are the most common species as measured by NISP. However, both approaches indicate that across the Paleolithic sequence at Klissoura Cave 1, small game animals became increasingly important, and later foragers relied heavily on small, fast-moving game in particular. 204 Figure 5.13 Proportion of small slow to small fast game biomass, plotted against time. The relationship is significant and negative. Data from Table 5.5. CONCLUSIONS Shifts in the representation of different prey groups at Klissoura Cave 1 likely signify changes in species availability as a function of both climate change and hunting pressures exerted by Pleistocene humans. No trends in species diversity are apparent in the assemblage, but there is a shift in the proportions of certain types of game through the sequence. One of the central goals of this presentation is to tease out differences between environmental impacts on biotic communities and human demographic pressures that may have influenced the availability of desirable species. It is relevant at this point to return to the predictions presented in Chapter 3 in an attempt to evaluate whether some shifts in game use at Klissoura Cave 1 track changing environmental conditions, and if others may be attributed to humans. 205 Table 5.6 includes the predictions set forth in Table 3.1, with a simplified summary of species actually found in different cultural layers at the site. As discussed further below, most species key to understanding changes in human demography are flexible in terms of their habitat requirements, and are therefore not useful for environmental reconstructions. This reduces the risk of tautologies when trying to distinguish between human or environmental factors as causing changes in the composition of the faunas at Klissoura Cave 1. The earliest phases of human occupation are thought to have been deposited at Klissoura Cave 1 as early as MIS 5a, though some layers may also have formed during MIS 4 (XXa-XXb and XVIII-XIX). During milder MIS 5a, wild pigs, fallow deer, tortoises, hares and partridges should have been available in the vicinity of the site, while during cool, dry MIS 4, fallow deer, ibex, wild ass, aurochs, tortoise, hare, partridge, and great bustard are expected. What in fact is found is a dominance of fallow deer, though small amounts of steppe species including aurochs and European ass, as well as ibex in the oldest Middle Paleolithic layers, while wild pigs are more abundant in Middle Paleolithic layer XVIII-XIX. Few small game species were recorded. All available climatic records from this period come from northern or central Greece, so it is possible that forest expansions recorded elsewhere during MIS 5a were less widespread in the south, leading to a more mixed environment in Peloponnese that would have supported the range of ungulate species mentioned above. Clearly, forests persisted to some degree as evidenced by wild pig, but it seems that these forests were probably broken up by open areas. Based on the faunal evidence alone, Middle Paleolithic layers XXa-XXb and XVIII-XIX cannot confidently be assigned specifically to MIS 5a or 4. 206 Table 5.6 Taxa found in Klissoura Cave 1 by cultural layer, with predictions from Table 3.1 in parentheses. P = present, A = absent, U = uncommon (NISP < 5). Grey shading indicates deviations from predictions. Roe deer and chamois are excluded because of small samples overall. *Epigravettian NISP = 67. 207 Middle Paleolithic layers XV-XVII, XI-XIV and X either formed toward the end of MIS 4 or beginning of MIS 3, and layer VIII was deposited during MIS 3. The sequence lacks dates from the Middle Paleolithic, and it is difficult to distinguish between MIS 4 and cool periods of MIS 3. Oxygen isotope stage 3 was highly variable and biotic communities shifted between dry, open periods that would have supported steppe species such as aurochs, European ass, ibex and great bustard, wet, forested periods in which wild pigs thrived, and periods of mixed forest-steppe that supported the widest range of prey species: red deer, roe deer, chamois and ibex. Fallow deer, tortoises, European hare and rock partridges are wellsuited to any of these climatic regimes. Unfortunately much of the climatic variability is difficult to correlate with the resolution available from the archaeological evidence, and it is not clear which layers were formed during MIS 4 or 3. In some instances, taxa present at Klissoura Cave 1 may be indicative of climatic conditions during specific phases of occupation. Environmental conditions at Klissoura Cave 1 seem to have been similar for the duration of the Middle Paleolithic MIS 4/3 (layers XV-XVII, XI-XIV and X) and 3 (VIII). Fallow deer continue to dominate the assemblages, though steppe species including aurochs and European ass, and forest-preferring wild pigs are also present in all of the layers. Red deer and ibex are also found, as well as trace amounts of roe deer. It is unclear if the lack of great bustard reflects human hunting decisions, or the environment not being sufficiently dry and open. However, most Middle Paleolithic assemblages in the Mediterranean region lack human-associated bird remains (but see Blasco and Fernández Peris 2009). From the composition of species, it seems that most of the Middle Paleolithic layers were deposited at a time when the Klissoura Gorge and surrounding area was a mosaic of 208 forest and steppe communities. The climate may have been on the drier end of the spectrum to still support mixed plant communities, as it has been established elsewhere that decreased moisture generally correlates with decreases in primary productivity and a corresponding decrease in mammalian taxonomic richness (Brown 1973, 1975; Grayson 1998, 2000; Meserve and Glanz 1978). Richness may not be the most appropriate measure to use at Klissoura Cave 1 because it correlates so closely with sample size, but evenness values are quite low during the Middle Paleolithic, which may indicate that diverse ungulate resources were not usually widely available, especially when conditions were slightly drier. An increase in moisture is typically followed by an increase in plant productivity, which is reflected in an increase in the taxonomic richness of animal communities, but only to a point, since environmental heterogeneity is also important (Abramsky and Rosenzweig 1984; Brown 1973, 1975; Meserve and Glanz 1978). This seems to be the case by the beginning of the Upper Paleolithic. The Early Upper Paleolithic (V) and lower Aurignacian (IV) layers display higher levels of species diversity (evenness). Fallow deer and small game taxa, excluding bustard, are represented in these occupations, alongside species that prefer steppe communities (aurochs and European ass), forests (wild pigs), and mixed forest-steppe habitats (red deer, ibex, chamois and roe deer). This range of species, as well as local climatic reconstructions (Albert 2010; Ntinou 2010), indicates that the plant communities continued to be mixed forest-steppe, but the taxonomic evenness suggests that conditions were slightly wetter than they were during the final Middle Paleolithic occupations. The middle Aurignacian (IIIe-g) and Upper Paleolithic layer III” contain a similar set of fauna even though they are not consecutive stratigraphically. In many ways, the large game composition is reminiscent of the MIS 4/3 Middle Paleolithic series, with a dominance 209 of fallow deer and low incidence of aurochs, European ass, red deer and ibex. These species, and the lack of wild pig, indicate that there were probably mixed forest-steppe communities in the region, and the environment was slightly drier than it was when the first two Upper Paleolithic layers were deposited. Ungulate species evenness is also low in these periods, like the later Middle Paleolithic layers. These Upper Paleolithic layers differ from the Middle Paleolithic, however, because there is a higher proportion of small game. The upper Aurignacian (IIIb-d) is stratigraphically above the middle Aurignacian (IIIe-g) and below the non-Aurignacian Upper Paleolithic industry (III”), but its fauna has more in common with layer III’, the Mediterranean backed-bladelet industry that overlays the non-Aurignacian Upper Paleolithic (III”). The large game signature is similar to the Upper Paleolithic layers discussed previously, with an abundance of fallow deer and modest remains of other ungulate species that prefer steppe or mixed habitats. These two layers are unique, however, because they contain the highest proportion of great bustard remains. There is little reason to believe that Pleistocene humans had ignored great bustard earlier in the Upper Paleolithic; other small, fast-moving species such as hares are found throughout the sequence. It seems likely that the difference was in prey availability in the gorge area; the upper Aurignacian (IIIb-d) and the Mediterranean backed-bladelet industry (III’) may have been deposited during a period of climatic deterioration, when there was a further expansion of open, steppe communities. These are also the layers in which rock partridge is the most abundant, though partridges were exploited in other Upper Paleolithic periods. Perhaps partridges were caught with the same technologies used to hunt bustards and they were selected as part of an optimal foraging strategy. It is likely that rock partridges flourished during these periods near the site because, though they are tolerant of a wide range of 210 habitats, they prefer open areas for breeding. A major difference between the two layers is the increased abundance of hares, a versatile species, in the Mediterranean backed-bladelet industry (III’) layer, which cannot readily be explained by environmental factors. Local climatic data that coincide with the Epigravettian (IIa-d) occupation of Klissoura Cave 1 indicate dry conditions at this time (Albert 2010; Geraga, et al. 2005; Karkanas 2010; Ntinou 2010). Significantly, there is an absence of great bustard from this layer. Ungulate evenness is comparatively high, probably because fallow deer frequencies decline relative to other prey types. Steppe taxa and those that require some forests are apparent, and their presence indicates that even if the climate was dry, some mixed communities persisted. The Mesolithic (3-5a) was deposited during the early Holocene, which was a time of increased moisture across much of Greece. The species representation in Klissoura Cave 1 does not necessarily reflect this; the same suite of species found in earlier times of mixed environments are still apparent, and in fact wild pigs are absent from the assemblage. Conditions may have been similar in MIS 5a, when expanded forests were expected, but are not actually reflected in the local ecosystems of southern Greece. The Mesolithic (3-5a) faunal assemblage at Klissoura Cave 1 is similar to that found in the Epigravettian (IIa-d), with a more even representation of ungulates driven by the decrease in fallow deer, and an increased importance of European hare. From an environmental perspective, the major changes that stand out in the Klissoura Cave 1 faunal series are an increasingly wet environment during the early phases of the Upper Paleolithic, and a few severely dry periods within the Aurignacian and later industries. Other trends in game use are more readily explained by human hunting pressures. A major 211 shift in the large game component at Klissoura Cave 1 that cannot be readily explained by climatic change is the relative decrease in the proportion of high-return ungulates to lowerreturn small animal taxa through the sequence, which bottoms-out in the Epigravettian (IIa-d) and Mesolithic (3-5a). This change is exemplified mainly by a decrease in fallow deer, a species that dominates the Middle and Upper Paleolithic assemblages by NISP and biomass until the Epigravettian (IIa-d), when it is surpassed by hares and other ungulate species. Chapter 6 explores this shift in more detail, considering, for example, whether people were travelling further for ungulate resources, or if younger animals were exploited more in later periods. Both of these possibilities may be indications of resource stress. Another important area of inquiry concerning Paleolithic human diets is changes in small game use through time (Munro 1999, 2004; Stiner 2001, 2005, 2009; Stiner, et al. 2000). At Klissoura Cave 1, small game use changed in fundamental ways over the course of the Paleolithic. In addition to a trend towards greater exploitation of small game animals at Klissoura Cave 1, it has been shown elsewhere that targeting specific kinds of small game taxa is a sensitive indicator of intensification efforts (Munro 1999, 2004; Stiner 2001; Stiner, et al. 2000). Specifically, a shift from easy-to-catch small, slow-moving (i.e., higher return) animals to small fast-moving (i.e., low return) prey is often indicative of increased human population pressures. During the Middle Paleolithic at Klissoura Cave 1, people relied more heavily on tortoises to supplement their diet, even though it was always incidental to large game hunting in terms of absolute meat mass. The proportion of tortoises in the small game assemblages slowly declines until they disappear from the record after the Mediterranean backed-bladelet industry (III’). Interestingly, great bustards also disappear from the sequence at this time, even though the Epigravettian climate was dry and open, conditions preferred by 212 this species. A commonality between tortoise and bustards is their low reproductive rates, so it is possible that human population densities reached a point in these later periods that such vulnerable taxa were essentially hunted out in the vicinity of Klissoura Gorge. Conversely, the frequency of hares, which are prolific but low-return, increase through the sequence. It should be noted that hares are common in one of the Middle Paleolithic layers (XI-XIV) by NISP count, though they only represent two individuals. It is possible that hares were occasionally hunted by Middle Paleolithic hominins, or they may be intrusive into these levels (see Chapter 6 for a discussion of low frequencies of hare burning as compared to other prey groups in the layer). Along with an increase in small, fast-moving game there is an increased use of low-ranking land snails in the later Paleolithic and Mesolithic layers. Snail exploitation began in the later Aurignacian or during one of the Upper Paleolithic blade/bladelet industries and intensified in later periods. The Klissoura Cave 1 faunal assemblages provide a long sequence over which climate change and human hunting decisions can be evaluated. Ideally, many local highresolution environmental studies would be available, in addition to a tight series of dates spanning the entire archaeological sequence. The available data, particularly from the Klissoura Cave 1 Middle Paleolithic, currently fall short of this. Despite this shortcoming, the faunal series at Klissoura Cave 1 can reasonably correlated to larger climatic shifts that are well-documented across Greece and the Mediterranean Basin, specifically climatic amelioration during Aurignacian layer IV and a dry, open environment during the Epigravettian (IIa-d). From the faunal sequence, a convincing case can be made for a more intensive use of animal resources by human hunters utilizing Klissoura Cave 1 over the course of the Paleolithic. Multiple lines of evidence, including changes in the frequency of 213 ungulate taxa, in particular fallow deer, shifts from high-ranked to low-ranked small game species, and the disappearance of taxa sensitive to disruptions in their reproduction all point to the same conclusion: human population densities in southern Greece were on the rise in the Paleolithic to the extent that new foraging strategies and intensification efforts were adopted to keep up with this growth. At the same time, biotic communities were forever altered by human hunters that would turn to agricultural lifeways in a few short millennia. 214 CHAPTER 6: PREY SELECTION, TRANSPORT DECISIONS AND BUTCHERY DAMAGE INTRODUCTION Many decisions accompany the subsistence pursuits of hunter-gatherers, from the location selected for foraging to the animal targeted in a herd. Once a prey animal is captured, the meat is divided and often transported back to a base camp or habitation site. At camp, the carcass is processed, which may involve some combination of defleshing, bone marrow extraction, cooking, and bone grease rendering. All of these behaviors play a part in determining the composition of faunal assemblages. The series of decisions that goes into killing and processing an animal may be influenced by many factors, including the size of the animal, the distance from the base camp where the kill occurs, how many members are in the group, and how badly food is needed (see Binford 1978; Bunn, et al. 1988; O'Connell, et al. 1988; 1990 for ethnographic examples). I return in this presentation to patch choice and central place foraging models that were first discussed in Chapter 1; such models are useful for evaluating transport decisions and butchery patterns across a carcass. Diachronic change in food collecting behaviors may indicate intensification efforts, which are often related to resource stress. Intensification can take many forms: the selection of lower-ranked prey species (Chapter 5), hunting lower-return individuals within a species (e.g., juveniles) (Broughton 2002; Butler 2001; Davis 1987; Munro 2004; Speth and Clark 2006; Stiner, et al. 2000), hunting forays further from the base camp (Bird and Bliege Bird 1997; Kelly 1995; Metcalfe and Barlow 1992; O'Connell, et al. 1988, 1990), and intensified carcass processing such as grease rendering (Binford 1978, 1984b; Broughton 1999; Burger, 215 et al. 2005; Lupo and Schmitt 1997; Manne and Bicho 2009; Munro and Bar-Oz 2005; Potter 1995; Saint-Germain 1997; Stiner 2003b). Findings reported in Chapter 4 make it clear that the Paleolithic and Mesolithic faunas at Klissoura Cave 1 were accumulated and subsequently modified primarily, if not exclusively, by humans. This chapter explores human activities in greater detail, by examining the degree of burning on faunal remains, butchery damage, transport decisions and prey mortality patterns. Large and small game data are evaluated separately because different considerations must be taken into account in the analyses of transport and butchery as a function of prey body size. For example, small animals are easily transported whole to a site by one person (Bartram, et al. 1991; Yellen 1991), whereas larger animals require a greater effort and are often butchered in the field to maximize the edible meat that can be carried to the base camp by a small group of hunters (e.g., Binford 1978; Bunn, et al. 1988; O'Connell, et al. 1988, 1990). This chapter examines two major stages of human decisionmaking in meat procurement and processing: decisions that hunters make in the field in regards to prey choice and field processing at or near kill sites, and butchery strategies that are applied once a carcass has been brought to the residential site. Kill and residential sites may contain a complementary set of activities, and only the residential site is represented at Klissoura Cave 1. A single large skeleton may be divided between a kill site and home base, with bulky or low-utility elements abandoned at the kill, and meaty or easily carried elements transported to a residential site. Evidence of defleshing or disarticulation may begin at the kill site but continue at a base camp. Though some snacking and cooking can occur at kill sites, burning, intensive marrow processing, and grease rendering are typically activities that are confined mainly to residential sites. Based on the large number of hearths, the 216 Aurignacian layer IV structure, the location of the site in a rock shelter, and many other indications, it is reasonable to assume that Klissoura Cave 1 was used primarily or exclusively as a base camp or residential site. Therefore, it is expected that the zooarchaeological assemblage should reveal processing activities that occurred at the residential site, and the suite of skeletal elements that were selected for transport back to the shelter. Both lines of evidence are useful indicators for understanding whether hominins at Klissoura Cave 1 were making decisions based on optimizing their returns or intensively harvesting food resources. Many observations associated with human activities were recorded during the analysis of the Klissoura Cave 1 faunas, including burning damage, tool marks, impact fractures, and fresh bone breaks. Damage on faunal specimens of both large and small prey species confirms that the faunas were collected and modified by humans (Chapter 4), a point that is expanded upon below. In the analyses presented in this chapter, NISP counts often vary between tables and with values presented elsewhere in this dissertation, following points made in Chapter 4; shattered dental elements or certain bones that are highly recognizable because of their texture (e.g. tortoise shell fragments) are used in taphonomic analyses, such as burning frequency, but not in considerations of species representation. Some of the tests presented below, such as those that evaluate butchery damage, exclude dental specimens because their fragmentation often is not directly related to food processing. Other tests target specific anatomical elements, as noted in the sections that follow. BURNING DAMAGE 217 Burning stages on the faunal remains were recorded following Stiner and colleagues (2005:45; 1995:226) (Table 6.1). Seven stages were recorded, including unburned, three different degrees of carbonization (semi- to fully blackened), and three different degrees of calcination (semi- to fully whitened). Browning of bones was not recorded as burning, since many other factors, such as staining, can produce coloration in archaeological bone. Published bone burning experiments indicate that fires using hardwoods common in the Mediterranean easily reach a temperature of 900 °C (Stiner, et al. 1995). Bone becomes calcined only when it is in direct contact with fires at these temperatures (Stiner, et al. 2005:50; Stiner, et al. 1995:234). Burning damage does not necessarily indicate that meatcovered bones were cooked to the point of charring, or even that bones were deliberately thrown into a fire after a meal. Experimental studies indicate that hearths built on sediments containing bone fragments burn bones up to 5 cm below the surface, depending on the sedimentary matrix and duration of burning, though these specimens rarely reach the point of calcination (Bennet 1999; Stiner, et al. 2005:50; Stiner, et al. 1995:230). This means that archaeological faunas can be carbonized long after they are deposited, though the highest levels of burning must come from direct contact with fires and may concentrate in the vicinity of hearth areas in sites that are visited repeatedly. There is certainly the potential for post-depositional burning at Klissoura Cave 1, with the abundance of hearth structures through the layers (Karkanas, et al. 2004; Koumouzelis, Ginter, et al. 2001; Pawlikowski, et al. 2000). 218 Table 6.1 Burn frequency and degree for different classes of animals at Klissoura Cave 1 by layer. 219 Table 6.1, cont. Burn frequency and degree for different classes of animals at Klissoura Cave 1 by layer. 220 Table 6.1, cont. Burn frequency and degree for different classes of animals at Klissoura Cave 1 by layer. 221 Bone is also useful as a fuel source in some circumstances; experimental studies indicate that it extends the life of a fire depending on how much bone is added (Théry-Parisot 2002), though the addition of small fragments tends to smother a fire (Mentzer 2009). Experimental studies using bone as fuel yields assemblages with high frequencies of calcined bone (76.1%) and high rates of fragmentation (56.9% smaller than 2 cm), particularly of cancellous bone (Cain 2005; Costamagno, et al. 1998; Costamagno, et al. 2005; ThéryParisot 2002). Establishing a connection between experimental bone burning studies and archaeological evidence of bone as a fuel source is difficult, however, because calcined bone tends to preserve poorly in archaeological contexts (Costamagno, et al. 2005; Stiner, et al. 1995), and cancellous bone is more susceptible to density-mediated processes (see Chapter 4). Mineralogical studies of archaeological sediments may help bridge the gap between experimental work and the actual use of bone as fuel in past with the identification of microscopic ash or calcined bone (Mentzer 2009). Burning frequencies are presented for different categories of species (i.e., tortoise, medium ungulate, etc.), as well as for each layer as a whole to determine if certain species categories are burned more often than others. This also helps to identify potentially intrusive animals, such as carnivores. Tooth and bone elements are included in this comparison. An association between degree of fragmentation and burning is well-established at many archaeological sites (Cain 2005; Costamagno, et al. 1998; Costamagno, et al. 2005; Stiner 2005:96-97; Stiner, et al. 1995:224; Théry-Parisot 2002). Mean fragment length for burned and unburned specimens at Klissoura Cave 1 is presented to determine if there is a similar difference at the site, particularly in relation to hearth areas (see also Chapter 7), and to evaluate the possibility of burning bone for fuel. Burn damage within species was also 222 examined to see if certain elements were treated differently. Tortoise remains in the Hayonim Cave and Meged Rockshelter assemblages of Israel, for example, show differences in burning frequencies for plastrons (flat, belly plate) versus carapaces (dorsal shell) (Stiner 2005:99), and significant differences in burning rates are found from marrow rich bones at Kebara Cave (Speth and Clark 2006) and at Qesem Cave (Stiner, et al. 2009). Differences between the degree of burning for medium ungulate crania and postcrania is also examined. Teeth are excluded from this particular burning analysis because they tend to shatter when burned, which would bias the results. Burning Damage on Small Game Small game burning damage is presented in Table 6.1. Trends in burning are difficult to discern for some species because the sample sizes are small. This is the case with tortoise remains for most of the Upper Paleolithic and Mesolithic layers. In the early Upper Paleolithic and Middle Paleolithic layers, tortoise remains are heavily burned, often at nearequal or higher rates than the assemblage as a whole (with the exception of Middle Paleolithic layers VIII, XI-XIV and XXa-XXb, the latter of which is a fairly small sample) (Table 6.1). The extent of burning of tortoise carapace and plastron fragments is presented in Table 6.2. In slightly more than half of the layers with samples of shell fragments larger than NISP = 20, plastron fragments exhibit equal or higher rates of burning than carapace specimens. Differences in the frequency of carapace and plastron burning is not as striking as that observed at some sites in Israel (Stiner 2005), though smaller tortoise shell samples in 223 Table 6.2 Burn frequencies for tortoise shells in layers with larger sample sizes. most layers may explain the discrepancy. High rates of burning on tortoise specimens is consistent with results obtained from Kebara and Hayonim Caves in Israel (Speth and Tchernov 2002; Stiner 2005). Upper Paleolithic and Mesolithic hare specimens exhibit about the same frequency of burning as the rest of the animal remains in each layer, while Middle Paleolithic hare remains are burned much less often than their respective whole assemblages (Table 6.1). Hare remains are fairly uncommon in the Middle Paleolithic layers at Klissoura Cave 1, with the exception of layer XI-XIV. The fact that hare specimens in all Middle Paleolithic layers are less burned than the bones of other animals suggests that they may have been intrusive into the layer, though a lack of avian or mammalian carnivore damage on hare remains (Chapter 4) indicates that they were not introduced to the site by predators. In layers that contain the largest bird assemblages (Mediterranean backed bladelet industry III’ and Aurignacian layers IIIb-d and IV), birds have considerably lower rates of burning than the average rate for these layers (Table 6.1). Carnivores also have extremely low incidences of burning, with the exception of Aurignacian layer IV, which may be explained 224 by sample sizes (Table 6.1). This may indicate that at least some of the remains were intrusive, or that they were cooked or disposed of differently than ungulates. Burning Damage on Ungulates For the Klissoura I faunal assemblage as a whole there is a statistically significant trend toward a decrease in burned bone over time (Table 6.1; n = 14, r s = 0.556, p = 0.039). This is reflected in the degree of burning for ungulates, which comprise the bulk of the assemblage through most of the sequence. Burning rates are similar for medium and large ungulates, except when samples for large ungulates are small (Table 6.1). Calcined bones were recorded in nearly all layers, indicating that some specimens were in direct contact with fires. The average lengths of burned and unburned specimens are presented in Table 6.3; burned specimens are consistently smaller than their unburned counterparts in every layer except the Epigravettian (IIa-d), which has an extremely small sample. Table 6.3 Mean length for burned and unburned identifiable specimens by layer. 225 Variation exists in the frequency of burning damage through the stratigraphic sequence, with the Middle Paleolithic and Aurignacian layers containing more burned bone than the later layers. This variation probably relates to occupation intensity as opposed to specific cooking behaviors; high frequencies of burning are common in layers that contain large numbers of hearth features (see Chapter 7), though burn frequency does not correlate with faunal input rates (n = 14, rp = 0.047, p = 0.874). Table 6.4 indicates the frequency of burning for medium ungulate crania versus postcranial specimens by layer; the differences are plotted in Figure 6.1 in layers with large enough sample sizes. Antlers are excluded from this analysis in case they were collected after they were shed, a strong possibility given the importance of antler for osseous tools. In most of the Upper Paleolithic layers, cranial elements are burned more often than postcranial elements (with the exception of Aurignacian layer IIIe-g). No such difference is apparent in the Middle Paleolithic; in general postcranial specimens have higher rates of burning, though burn frequencies are fairly close between the two sets of body parts. Differences in burning in Upper Paleolithic layers may indicate the systematic roasting of medium ungulate crania. There is little evidence to indicate that bone was used as a fuel source at Klissoura Cave 1. The average fragment size of burned bone is smaller than unburned bone, which is the case at other Paleolithic sites (e.g., Stiner 2005:92). The mean fragment size at Klissoura Cave 1 is larger than 2 cm in all layers except for one, however, and generally greater than expected if bone was burned for fuel (Cain 2005; Costamagno, et al. 1998; Costamagno, et al. 2005; Théry-Parisot 2002). There is no overabundance of burned cancellous bone in Klissoura Cave 1, especially in the Middle Paleolithic, though this may be the result of 226 Table 6.4 Burn frequency and degree for cranial and postcranial elements by layer for medium ungulates. *Small samples not included in Figure 6.1. 227 Table 6.4, cont. Burn frequency and degree for cranial and postcranial elements by layer for medium ungulates. *Small samples not included in Figure 6.1. 228 Figure 6.1 Frequencies of burned cranial and postcranial specimens at Klissoura Cave 1. In the Upper Paleolithic layers crania are more commonly burned. density-mediated attrition. Calcined bone is expected if bone is used as a fuel source (Costamagno, et al. 1998), but it is present in low frequencies in all layers. Costamagno et al. (2005) stress that the unidentifiable component of an assemblage is as important for evaluating bone burning as the identifiable portion, but calcined bone is also highly susceptible to degradation (Stiner, et al. 1995). More significant in the case of Klissoura Cave 1 is that fact that burned bones do not correlate spatially with the presence of hearth features (see Chapter 7), or in discrete areas that may indicate fireplace cleaning events. At this stage, there is no evidence that bone was burned as fuel at Klissoura Cave 1. 229 BUTCHERY DAMAGE AND OSSEOUS TOOLS: CUTS, CONES, IMPACTS, WORKING, AND OTHER FRACTURES Evidence of butchery is perhaps a more direct indicator of human modification of a faunal assemblage than burning, since it is not always clear if burning damage is associated with cooking activities, site cleaning, disposal of bone, or if burning occurred postdepositionally. Several kinds of butchery damage were recorded in the Klissoura Cave 1 faunas. Cone, or hertzian fractures and impacts are typically associated with the extraction of marrow from a mammalian long bone cavity (Binford 1978; Blumenschine and Selvaggio 1991:153-156; Brain 1981:141-142; Stiner 1994) or breaking open a tortoise shell for meat (e.g. Stiner 1994). Such fractures are occasionally caused by carnivore chewing or crushing, but are more often associated with human butchery activities, specifically the striking of cortical bone shafts with a hammerstone. Large carnivores occasionally create cone fractures, but Paleolithic humans produce them at much higher frequencies (Stiner 2005:93). Tool marks were recorded in the assemblage, though observed frequencies in the Upper Paleolithic and later assemblages likely underestimate the true number of cuts because many specimens were partly or wholly covered with concretions of ash (see Chapter 4). In this case, the presence of cone fractures or impacts and cut marks are more instructive than raw frequencies of such damage in establishing that humans modified faunal remains. Green bone fractures or "fresh breaks" occur postmortem, but while bone still maintains its natural fibrous composite structure (Currey 1984; Johnson 1985:160). Such breaks are indicative of damage shortly after the death of an animal, before the bone dries (Davis 1985:66-67). Many different fracture types are observed on archaeological assemblages (see Davis 1985; Gifford-Gonzalez 1989:188; Lyman 1994:318-324; Marshall 230 1989; Shipman, et al. 1981). Here I focus on fresh spiral fractures on tortoise bones and transverse fractures on mammalian and bird bones. In this case, spiral fractures are classic green bone breaks. Transverse green fractures are similar, but occur when a bone is broken transversely to its main axis using a shearing action (Gifford-Gonzalez 1989:188; Stiner 2005:93). This type of fracturing may relate to the partitioning of carcasses by hominins during butchery for transport or sharing (Stiner 1994). Carnivores produce green bone fractures in much the same way that they cause cone fractures. Some investigators have also linked transverse fractures on small mammals to human processes, such as on the long bones of hares and rabbits during marrow removal (Hockett 1994); transverse fractures on small animal bones may also relate to dismemberment (Stiner 2005:95). In the results presented below, butchery activities on ungulate remains are from all boney skeletal elements. Frequencies of cut marks on small animals are likewise counted on all bone elements, while transverse fractures on small game are only evaluated on selected long bones, including femora, humeri, tibiae (or bird tibiotarsi) and radii. Spiral fractures are presented for all tortoise shell fragments in the assemblages. Butchery Damage on Small Mammals and Carnivores Frequencies of cut marks on small animal and carnivore remains are presented in Table 6.5. Cut marks are present but rare on all species throughout the assemblages. The low frequencies of tool marks may be a combination of small sample sizes for small animals and carnivores in most layers combined with surface encrustations on specimens, especially in 231 Table 6.5 Butchery damage on small game at Klissoura Cave 1 by layer. the Upper Paleolithic layers (see Chapter 4). For example, the largest samples of hare and bird bones are from the Upper Paleolithic and later layers, where ashy concretions were common on the remains and frequencies of cut marks are extremely low (Table 6.5). One cut marked specimen was recorded for hares in the Middle Paleolithic, in layer XXa-XXb. A few tortoise elements exhibit cut or impact marks (Figure 6.2), but again frequencies are quite low. Evidence of human modifications on small animal remains comes mostly from high frequencies of transverse fractures on mammal and bird long bones and spiral fractures on tortoise shells (Table 6.6). Transverse fractures occur on well over 50% of hare long bones (with the exception of Middle Paleolithic layer XI-XIV) and 60% of bird long bones in assemblages with sample sizes greater than ten (Table 6.6; see also Bocheński and Tomek 2010 for a further discussion of damage on bird remains). Transverse fractures are also 232 observed on carnivore remains, though samples are small for any given layer (NISP < 5). Humans traces are the only significant modifications of the small animal remains, indicating that humans were the primary collectors of small game remains at Klissoura Cave 1. Low frequencies of fractures on Middle Paleolithic hare remains supports the hypothesis that hare remains were intrusive during much of this time period. Figure 6.2 Tortoise carapace fragment with crushing and an impact fracture. From Middle Paleolithic layer XVI. Butchery damage is found on all groups of small animals at Klissoura Cave 1, including the small mammals, reptiles, birds and carnivores in the assemblage. Cut marks are rare on small game remains, which is partially explained by poor bone surface visibility on Upper Paleolithic and later layers (Chapter 4). Another possibility for hares in particular is that elements expected to have cut marks associated with skinning (e.g., feet) are generally lacking in the assemblage (see below). 233 Table 6.6 Transverse fractures on small animal limb bones (humerus, radius, femur and tibia/tibiotarsus) by layer. Butchery Damage on Ungulates Butchery damage summaries for medium and large ungulates are presented in Table 6.7. Conservative criteria are used to evaluate damage, so cone and impact fractures are less common in the Klissoura Cave 1 assemblages than in other Mediterranean Paleolithic sites (see Stiner 2005:107), occurring on 1-3% of medium ungulate remains in most layers (Table 6.7). Impact damage is observed on a slightly higher proportion of large ungulate specimens in layers with sample sizes larger than n = 20, at rates of 1-4.5% (Table 6.7). There is no trend in the frequency of impact damage on either the medium or large ungulate remains (n = 14, rs = -0.525, p = 0.054 and n = 11, rs = -0.301, p = 0.368, respectively), though the p-value for medium ungulates is nearly significant, with a decrease in impact marks in later periods. This correlation, to the extent that it indicates a significant relationship between medium ungulate remains and time period, has the potential to explain 28% of the variation in the 234 Table 6.7 Butchery damage on medium and large ungulate remains. *Note small sample sizes. assemblages, which may reflect small samples in the youngest layers, or be related to decreased visibility of bone surfaces in Upper Paleolithic layers. In any case, butchery damage does not increase through time, which indicates that bone marrow was not necessarily more intensively accessed in later time periods. Frequencies of tool marks on the Upper Paleolithic assemblages are difficult to assess because of ashy concretions on the surfaces of bone in these layers. Cut marks are found on 235 0.3-1.1% of Upper Paleolithic medium ungulate remains and are almost entirely absent on large ungulate specimens (Table 6.7). Because tool mark frequencies for the Upper Paleolithic layers are unreliable due to low visibility of bone surfaces, no further analyses were attempted on them. Tool marks are more common in the Middle Paleolithic layers on both medium and large ungulate remains; cut marks are present on 0.8-3.4% of medium ungulate specimens and 1.1-6.4% of large ungulate remains in samples larger than NISP = 30 (Table 6.7) (see Figure 6.3 for an example of cuts on a medium ungulate specimen). Tool mark counts for medium ungulates are listed by element in Table 6.8 by layer. Most cut marks were recorded on unidentified long bone fragments. Aggregated cut frequencies for the entire Middle Paleolithic are illustrated in Figure 6.4 for elements with a combined NISP > 10 cuts. Most of the tool marks occur on the hind limbs, mandible, humerus or ribs. Transverse fractures are common in the ungulate assemblages at Klissoura Cave 1, with 12-46% of medium and large ungulate remains exhibiting such damage (Table 6.7). There is a statistically significant trend in the frequency of transverse fractures on ungulate specimens, with fewer fractures in the older layers (n = 14, rs = -0.895, p = 0.000 for medium ungulates and n = 11, rs = -0.664, p = 0.026 for large ungulates). This may indicate the development of finer carcass partitioning for transport, sharing, cooking, or marrow processing strategies in the later periods. Worked Bone and Antler The use of bone and antler as raw materials for tool manufacture is rare in Europe and western Asia during the Middle Paleolithic (Villa and d'Errico 2001), though these 236 Figure 6.3 Example of cut marks on medium ungulate remains. Specimen from Middle Paleolithic layer XII. Table 6.8 Cut marks on medium ungulate elements in Middle Paleolithic layers. 237 Figure 6.4 Cut frequencies on fallow deer elements in the Middle Paleolithic layers at Klissoura Cave 1. Includes elements with n > 10 cuts. technologies appear by 80,000 to 100,000 years ago in Africa (McBrearty and Brooks 2000). The use of osseous technologies in Europe is associated first with basal Upper Paleolithic and transitional technologies, such as the Châtelperronian, Uluzzian and Streletskaian, among others. Osseous technology becomes ubiquitous during the Aurignacian (Bar-Yosef, et al. 2006; Conard and Bolus 2003; d'Errico, et al. 2003; d'Errico and Laroulandie 2000; d'Errico, et al. 1998). Worked bone is present in the Klissoura Cave 1 assemblages, but only in the Upper Paleolithic and later deposits (Koumouzelis, Ginter, et al. 2001). Nearly all of the osseous tools and manufacturing debris were separated from the faunal remains as the time of 238 this study. Simple counts of worked bone and antler by layer are included in this chapter, though a more in-depth study of osseous tools from a technological standpoint is ongoing by Rozalia Christidou. Overall, worked bone or antler is rare in the assemblages (Table 6.7). One possible worked bone was recorded in Middle Paleolithic (XI-XIV) (Figure 6.5); the rest of the worked bone come from Upper Paleolithic and later layers. Most of the osseous technology at Klissoura Cave 1 is made from antler, so proportions of antler relative to total ungulate NISP counts are compared in Table 6.9 along with percentages of worked antler by layer. Antler is present in appreciable frequencies, given that only adult male cervids develop antlers, and they are present on the animal only on a seasonal basis. Tools made of antler were found primarily in the middle and lower Aurignacian layers (IIIe-g and IV), though a few isolated tools were also found in later layers (Table 6.9). Aurignacian layers IIIe-g and IV also contain high frequencies of antler fragments, suggesting some on-site production of osseous tools. The non-Aurignacian Upper Paleolithic layer (III”) has a high proportion of Figure 6.5 Worked bone from Middle Paleolithic layer XIV. 239 Table 6.9 Counts of all antler and worked antler by layer, and corresponding ungulate NISP counts for comparison. antler specimens, though none of it is worked (Table 6.9). In general, antler is rare in the Middle Paleolithic layers, indicating that male fallow deer were infrequently targeted for hunting or that male deer were hunted in seasons when antler was unformed or in the early stages of growth. BONE PROCESSING One measure of intensification of meat resources is the extent to which a carcass is processed for its edible tissues (e.g., Binford 1978, 1984b; Broughton 1999; Burger, et al. 2005; Lupo and Schmitt 1997; Manne and Bicho 2009; Munro and Bar-Oz 2005; Potter 1995; Saint-Germain 1997; Stiner 2003b). As discussed in Chapter 1, the patch choice model predicts that the highest-return tissues are utilized first. Lower-utility animal products are typically only exploited if their processing costs are lower than travel and handling costs for 240 additional prey items. Heavy exploitation of low-return tissues may reflect resource stress or intensification. Organs and muscle are the easiest tissues to extract and require a fairly minimal degree of butchery. Cold marrow processing is slightly more labor intensive but still quite simple; using a hammerstone, marrow-rich long bones are crushed, granting access to nutrient-filled bone marrow cavities, and creating cone and impact fractures discussed above (e.g., Binford 1978; Blumenschine and Selvaggio 1991; Brain 1981:141-142; Stiner 1994, 2005). This level of processing began in the Lower Paleolithic, possibly as early hominins scavenged ungulate carcasses. Cold marrow processing from medullary cavities is nearly ubiquitous in Paleolithic sites from all subsequent periods and regions (e.g., Bar-Oz and Munro 2007; Binford 1984b; Blumenschine and Magrigal 1993; Boyle 2000; Brain 1981; Costamagno, et al. 2006; Egeland and Byerly 2005; Gaudzinski-Windheuser and Niven 2009, and discussion therein; Hockett and Bicho 2000; Lupo 1998; Miracle 2005; Munro and Bar-Oz 2005; Niven 2007; Stiner 1994, 2005; 2009, among many others). In order to assess the degree of cold marrow processing at Klissoura Cave 1, ungulate long bones are ranked by the volume of their marrow cavities. Large ungulates are excluded from this analysis because of small sample sizes. MNE values are listed for each element, and the percent of unopened elements are indicated. It is expected that elements with the greatest amount of bone marrow tend to be opened more frequently. It is also predicted that when resources are intensified, perhaps during periods of stress, elements containing smaller amounts of marrow will be broken at high frequencies. An even more intensive use of large game carcasses is heat-in-liquid bone grease rendering, a process that seeks to unlock small pockets of animal fat in the pores of spongy bone elements. During grease rendering, cancellous parts of elements are pulverized and 241 boiled, typically by placing heated stones into water and liberating the oil, which can be skimmed from the top of the mixture, purified and stored (e.g., Binford 1978; Brink 1997; Lupo and Schmitt 1997; Saint-Germain 1997; Vehik 1977). Bone grease rendering is extremely labor intensive (Church and Lyman 2003), and is typically associated with increased seasonality (Enloe 2003a; West 1996, 1997) or with resource intensification (BarOz and Munro 2005; Manne and Bicho 2009; Munro 2004; Nakazawa, et al. 2009; Stiner 2003b). Notably, bone grease rendering has not been found in the Mediterranean region earlier than the Gravettian (Manne and Bicho 2009; Manne, et al. 2005). At Klissoura Cave 1, the most likely layers to contain bone grease rendering (e.g., the Epigravettian and Mesolithic) have such small samples that it is difficult to test for grease rendering. Further, supporting artifacts typically found associated with grease rendering such as fire cracked rock or hammerstones and anvils (e.g., Manne and Bicho 2009; Manne, et al. 2005) are not common at Klissoura Cave 1 (Kaczanowska, et al. 2010). Ungulate Marrow Processing Evidence of bone marrow processing is based in part on the percentages of opened medium ungulate long bone medullary cavities (Table 6.10). No major long bone elements are complete in any of the layers, indicating that medium ungulates were processed consistently for medullary marrow in all periods. Some phalanges were recovered intact, however. These elements contain the least amount of marrow (see marrow utility indices in Binford 1978). Table 6.11 presents the percentage of unopened phalanges by layer provided 242 Table 6.10 Percent of medium ungulate limb bones not opened prior to discard, by layer. 243 Table 6.11 Percent of unopened medium ungulate phalanges for layers with MNE values larger than 10. that the assemblage was relatively large. In general, first phalanges are processed more heavily than second phalanges, which were in turn opened more frequently than terminal phalanges (Table 6.11, Figure 6.6). This reflects the decrease in nutritional return that comes from opening third phalanges, as they contain less marrow than the proximal phalanges. No temporal trend exists for processing any of the phalanges, though there is an increase in opened third phalanges during Middle Paleolithic layer XI-XIV and especially during Aurignacian layer IV. This may indicate a more intensive use of animal carcasses, particularly during Aurignacian layer IV, which was a period in which occupation intensity or site use changed . A few differences in marrow processing strategies were observed in at Klissoura Cave 1, though marrow was intensively utilized throughout the sequence. There is no evidence for bone grease rendering, though by the time this strategy is expected to appear in Eurasia (during the Gravettian or later), there is an overall shift to lower-ranked small game resources at Klissoura Cave 1 and much less frequent use of the shelter (see Chapter 5). One 244 Figure 6.6 Percent of unopened first, second and third phalanges through the layers at Klissoura Cave 1. First phalanges, which contain the highest marrow content, are consistently processed through the layers. Second phalanges were opened more often in the late Middle Paleolithic and Aurignacian layers. Terminal phalanges were only intensively processed during the late Middle Paleolithic and Aurignacian layer IV (data from Table 6.11). notable trend in butchery strategies is a statistically significant increase in transverse fractures in later time periods. This could relate to more standardized carcass processing strategies in the Upper Paleolithic, or at least more efficient segmenting for transport, cooking, or while extracting bone marrow. Another difference is the increase in processing of terminal phalanges, an element with an extremely low marrow yield, during Aurignacian layer IV. 245 PREY TRANSPORT Evaluating the representation of prey body parts at an archaeological site is instructive in understanding the decisions surrounding the transport of prey to a base camp or habitation site, assuming post-depositional taphonomic factors did not produce the damage (see Chapter 4). Central place foraging models are often used to understand transport decisions involving the movement of specific portions of large-bodied prey (Broughton 1999; Cannon 2003; Nagaoka 2005; O'Connell, et al. 1988, 1990; O'Connell and Marshall 1989; Speth 1991; Speth and Scott 1989). As local resources are depleted, hunters travel farther from residential sites to procure large game. The further they travel, the more selective foragers tend to be about which portions of a carcass are carried to the habitation site. If the meat must be carried over long distances, foragers typically focus on transporting fewer high-utility skeletal elements, or elements that provide the greatest nutritional return (Binford 1978; Broughton 1999; Cannon 2003; Nagaoka 2005; Speth 1991; Speth and Scott 1989; Stiner 1991, 1994, 2005). A bias toward high-utility skeletal portions at Klissoura Cave 1 may indicate increased travel times between the site and hunting location. Small-bodied prey can easily be carried to a site in whole form (Bartram, et al. 1991:102; Yellen 1991:5). Examining body part profiles for small animals, such as hares or birds, is valuable because it contributes to our understanding of taphonomic processes at a site and the possible presence of discrete activities areas within a site. Of course, the hypothesis that small game animals were transported to the site whole must also be evaluated. Utility Indices 246 There are many methods for examining prey transport decisions. One is based around utility indices introduced by Binford (1978). Based on his work with the Nunamiut in Alaska, Binford (1978) derived a series of indices for caribou (Rangifer tarandus) and dall sheep (Ovis dalli) in order to rank the utility of different classes of edible tissue (meat, marrow, and bone grease) by body part for these two species (see Appendix H for values, from Binford 1978). Two combined indices were derived from the three tissue-specific indices mentioned previously: the general utility index (GUI) considers the combination of meat, marrow and bone grease associated with different elements or portions of elements, and the modified general utility index (MGUI) takes into account food-poor articulated elements that tend to be transported as riders with larger, food-rich elements. GUI and MGUI values both correlate significantly with meat index values; MGUI also correlates with bone grease index values (see Table 6.12). Raw index values are standardized to a scale between one and a hundred, and the normed MGUI values become %MGUI. In order to compare utility indices with archaeological data, %MAU is constructed by calculating MAU values for each element, and dividing this by the maximum MAU value in the assemblage: (observed MNE/expected MNE)/(maximum MAU). Binford (1978:81) then constructed a series of models of transport strategies based on scatter plots of %MGUI versus %MAU. This analytical model was quickly applied to archaeological assemblages by other zooarchaeologists (Landals 1990; Speth 1983; Thomas and Mayer 1983), though many were critical of aspects of the technique, particularly its application to taxa for which no index values had been derived (Chase 1985; Jones and Metcalfe 1988; Metcalfe and Jones 1988). It was later realized that certain relationships between %MGUI and %MAU (specifically the “reverse utility curve”) 247 Table 6.12 Pearson’s correlation values between Binford’s (1978) utility indices. * Indicates significant relationships. could be caused by post-deposition attritional processes in archaeological assemblages rather than human transport decisions (Grayson 1988, 1989; Lyman 1985). Additionally, some utility curves (specifically the gourmet strategy and reverse gourmet strategy) particularly as conceptualized by Lyman (1994:228) could never occur because so few elements in an ungulate assemblage have %MGUI values above 80, which is required for these two models to be realized. Despite these potential pitfalls, it is useful to examine relationships between utility indices and %MAU in zooarchaeological assemblages, if density-mediated effects have already been largely excluded. Density-mediated attrition was not a significant problem for the ungulate remains in the Upper Paleolithic and later layers at Klissoura Cave 1. On the other hand, possible density-mediated biases in Middle Paleolithic layers must be kept in mind during this analysis. Correlations between %MAU and high utility values, particularly meat, GUI or MGUI may be indicative of transport strategies that focus on high utility elements. Although these indices include bone grease, this does not seem to have been a motivating factor for the transport of certain ungulate elements to the site, as there is no evidence of bone grease rendering. Thus, the main goals in the transport of ungulates would include meat and medullary marrow. 248 Percent MAU values for fallow deer element portions are listed in Appendix I by layer; corresponding meat, marrow, grease, GUI and MGUI values for reindeer are in Appendix H (from Binford 1978). Spearman’s rank-order correlation values for %MAU and the respective utility indices listed above are presented in Table 6.13. Few correlations are significant, with the exception of marrow index values in Aurignacian layer IV and Middle Paleolithic layer XXa-XXb (Table 6.13). It is possible that marrow-rich elements were preferentially transported to the site during these periods for the added nutritional value of bone marrow, particularly during Aurignacian layer IV when there are other markers of intensification (e.g., marrow processing ungulate terminal phalanges). On the other hand, there is clearly no evidence for preferential transport of elements with overall high utility. This is especially the case in Middle Paleolithic layer VIII, where there is a significant, negative relationship between meat and GUI index values and %MAU, indicating a lack of high-utility elements at the site during this period. It must be noted that several lines of evidence indicate that density-mediated attrition was a factor in layer VIII (Chapter 4). Scatter plots of %MAU and %MGUI are presented by layer in Figure 6.7. For the most part, the plots do not resemble any of Binford’s (1978) models of transport strategies. However, some of the plots, particularly non-Aurignacian layer III”, Aurignacian layer IIIb-d and Middle Paleolithic layers VIII and XI-XIV, slightly resemble the reverse utility curve. The reverse utility curve reflects an abundance of low-utility elements and underrepresentation of high-utility elements, suggesting the preferential transport of lowutility elements, possibly as a result of opportunistic scavenging by hominins. It must be noted, however, that Middle Paleolithic layers VIII and XI-XIV have evidence of densitymediated attrition so it is possible that post-depositional bone loss accounts for the bias in the 249 Table 6.13 Spearman’s rank-order correlation values for fallow deer %MAU and caribou utility indices (from Binford 1978). *Statistically significant relationships. 250 Figure 6.7 Scatter plots of %MGUI and %MAU values for different layers at Klissoura Cave 1 (following Binford 1978). 251 representation of remains. The Upper Paleolithic layers (III” and IIIb-d) may simply resemble the reverse utility curve because of low incidences of two high-utility elements (the two points in the bottom right of each plot), as no other lines of evidence indicate that lowutility elements were preferentially transported during these two periods (see below). Broughton (1999) introduced the mean food utility index (FUI) as a method for assessing the transport of high-utility body parts to an archaeological site. He uses FUI values derived by Metcalfe and Jones (1988) for caribou, which are based on Binford’s (1978) data. FUI is equivalent to MGUI and for this reason is closely correlated with the meat index (see above). Following Broughton (1999), mean FUI values are derived for fallow deer in each layer at Klissoura Cave 1 (Appendix J). This is done by multiplying fallow deer MNE for each element by the element’s respective FUI value (from Broughton 1999). Values for all elements in a layer are then summed, and divided by the total MNE for the layer to determine mean FUI. Note that Broughton (1999) multiplies FUI by relative skeletal abundance (RSA), which is an NISP-based value, and MNE is used in this study. Mean values are plotted in Figure 6.8; the horizontal line indicates mean FUI for a complete skeleton (Appendix J). Only one layer has a higher mean FUI value than that for a complete skeleton, which may indicate increased transport distances and the preferential movement of high-utility elements. Based on mean FUI, during most periods, high-utility elements were not preferentially brought to the site, suggesting that fallow deer may have been hunted near Klissoura Cave 1, though density-mediated attrition may be a factor in some layers (see Chapter 4). Another issue is that transport goals may not have been based strictly around meat utility, which is explored further below. 252 Figure 6.8 Mean FUI for each layer, following Broughton (1999). Horizontal line indicates the average FUI for a complete cervid. Data from Appendix J. Body Part Profiles A final method for interpreting prey transport decisions is the analysis of body part representation at an archaeological site. Following Stiner (1991), the prey body plan is collapsed into anatomical regions that are logical packages for the purpose of meat transport and discard, or dismemberment. In the case of cervids or bovids, nine regions are considered: antler or horn, head, neck, axial skeleton, upper front limb, lower front limb, upper hind limb, lower hind limb and feet (see Appendix K for elements included in each region). The hare body plan includes eight regions, the ones mentioned previously, excluding antler/horn (Appendix L; Munro 2001). Partridge bodies are divided into eight regions: head, neck, axial, pectoral girdle, wing, upper hind limb, lower hind limb and feet (Appendix M; Munro 2001). 253 Minimal animal units (MAU) (Binford 1978) or standardized MNE (Stiner 1991), which are equivalent units, are determined by dividing the observed MNE for an element or anatomical region (regardless of body size) by the expected MNE value for that same element (or region) in a complete skeleton. This is done for different ungulate species, hare and partridges in each of the stratigraphic layers with sufficiently large sample sizes. When the results are plotted as bar charts, a complete anatomical representation is represented by bars of equal height for all body regions. Dividing the body in this way also bypasses many of the problems associated with variable structural density because each region contains corresponding high and low-density portions (Stiner 2002:982; 2004:129; 2005:182). Similarly to what was done in the taphonomic analysis in Chapter 4, hare remains are combined with all small mammal axial element data, and all medium ungulate axial elements are combined with fallow deer data. This done because hare is by far the most common small mammal and fallow deer dominates most of the assemblages; it is likely that less diagnostic axial elements came from these species. In the Aurignacian layer IV, where ibex are also well-represented, medium ungulate axial elements are split between fallow deer and ibex proportional to the representation of each species in the layer by MNE. In this layer red deer data are combined with large ungulate axial elements (Appendix N). Small Game Body Part Profiles Body part profiles for partridge from Mediterranean backed bladelet industry layer III’ and the upper Aurignacian layer IIIb-d are presented in Figure 6.9 (data from Appendix M). Partridge remains in both layers are dominated by elements from the pectoral girdle and lower hind limb elements; feet, heads and necks are absent in the assemblage. This is similar 254 to what is reported by Bocheński and Tomek (2010); they find high frequencies of partridge tibiotarsi, humeri, coracoids, pelves, ulnas and tarsometatarsus by NISP. Their study included a larger sample of avian remains from Klissoura Cave 1 than the assemblage presented here, though the sample included in this study is representative of all layers from the site. Figure 6.9 Anatomical profiles for partridges in the Mediterranean backed-bladelet industry (III’) and upper Aurignacian (IIIb-d) (data from Appendix M). Hare body part profiles are presented in Figure 6.10 (data from Appendix L) for most Upper Paleolithic and later layers (excluding the non-Aurignacian Upper Paleolithic layer III” and the Early Upper Paleolithic layer V) and Middle Paleolithic layer XI-XIV. Recall from Chapter 4 that density-mediated attrition on hare remains was only a potential factor in the Mesolithic (3-5a) and Epigravettian (IIa-d) layers (Table 4.1). Indeed, a Kolmogorov- 255 Smirnov (K-S) test on MAU in Mesolithic layer 3-5a indicates that hare anatomical representation is not uniform (Table 6.14). This is also the case for Aurignacian layer IV. Even in layers where the data do not differ from a uniform representation, hare Figure 6.10 Anatomical profiles for hares in the Upper Paleolithic and later layers and Middle Paleolithic (XI-XIV). Note the lack of foot elements in the UP and Mesolithic and neck and axial elements in all layers (data from Appendix L). Table 6.14 K-S tests for hare body part profiles by layer. *Asymp. Sig. > 0.05, data do not differ from a uniform distribution. 256 neck, axial and foot elements are poorly represented, particularly in Upper Paleolithic and later layers. The representation is slightly less biased in the Middle Paleolithic (XI-XIV), though this layer exhibits an overrepresentation of upper hind limbs and a lack of axial elements. This observation may contradict the assumption that hare carcasses were brought to the site complete. Alternatively this pattern may be due to limited spatial sampling in the excavations (see Chapter 7). Figures 6.11 and 6.12 provide a more detailed element-by-element comparison of foot and neck/axial bone representation for hare from Klissoura Cave 1. In both figures, elements are plotted in descending order of the most dense portion of each element; these figures illustrate the range of structural densities found in these portions of the hare skeleton (see values in Pavao and Stahl 1999; Appendix A). There is no bias in the representation of foot elements with respect to structural density (Figure 6.11) in the assemblages, with the possible exception of the Mesolithic (3-5a). In general, the entire foot region of hares is almost entirely absent throughout the later part of the sequence. Hare foot elements are not as severely underrepresented in the Middle Paleolithic (XI-XIV) (Figure 6.11). Likewise, the lack of neck and axial elements in the Upper Paleolithic and later assemblages cannot be explained by variation in bone density (Figure 6.12). Examining the post-cranial axial skeleton indicates that the innominate is consistently more common than other elements in the Upper Paleolithic and later layers, though it is not the densest element. Axial elements from the Middle Paleolithic (XI-XIV) are extremely scarce. Crania and mandibles are included in Figure 6.12 because they are not as severely underrepresented as the postcranial 257 axial skeleton in most layers (Figure 6.10), even though they could be removed during skinning or transport. Anatomical profiles for hares are surprisingly uneven, with feet and postcranial axial elements generally missing from the assemblages. It was established in Chapter 4 that density-mediated processes cannot explain the lack of hare elements in all Upper Paleolithic and later layers at Klissoura Cave 1, and discussions above indicate that foot and vertebral elements with the highest structural density values are no more common than those with low structural densities. Either hare carcasses were not transported to the site in a whole state or the excavation did not include all processing and refuse areas. Because hare foot bones are small, it also is possible that they were not retrieved during excavation or screening. This may be the case for phalanges, but probably does not explain the lack of metapodials, which are comparatively quite large. Cochard and Brugal (2004) argue that lagomorph feet may be absent from archaeological assemblages because they were removed at the kill site, or because they were deposited in an area of the site not used for cooking. The presence of the innominate and head parts, along with an absence of vertebral and rib elements in the Klissoura Cave 1 assemblages may be the result of postcranial axial elements breaking down into unidentifiable fragments (Cochard and Brugal 2004), whereas innominate fragments remained diagnostic. In fact, the general composition of hare body parts in the Upper Paleolithic through Mesolithic is similar to the representation found in the “zone de préparation culinaire” described by Cochard and Brugal (2004), with the exception of abundant long bone epiphyses at Klissoura Cave 1. 258 Figure 6.11 Plot of the foot region of hares from Klissoura cave 1 by layer. MAU (or standardized MNE) values for elements included in the foot are presented in order of descending structural density from the most dense scan site of each element. The expected MAU is presented based on the highest MNE value for all skeletal elements for each layer. Bone density values from Pavao and Stahl (1999). Figure 6.12 Plot of the axial region of hares from Klissoura cave 1 by layer. MAU (or standardized MNE) values for axial elements are presented in order of descending structural density from the most dense scan site of each element. The expected MAU is presented based on the highest MNE value for all skeletal elements for each layer. The head region is also plotted to indicate that in some layers it is present though other axial elements are not. Bone density values from Pavao and Stahl (1999). 259 Partridge body part profiles are even more biased than hare remains, but the sample sizes are large enough to analyze only two layers. It is difficult to determine if this is related to field butchery, spatial variation in site use, or taphonomic processes. Ungulate Body Part Profiles Preferential transport of ungulate elements from anatomical regions based on nutritional return, or food utility, is of central importance to this discussion. In order to quantify food utility, I turn to GUI values provided by Binford (1978). Values used here were originally calculated for caribou, which, as a cervid is the closest species to fallow deer for which utility values are derived. GUI is preferred over MGUI in this study because preferential transport is the main question. GUI correlates significantly with the meat index (see above), though it also takes into account marrow and bone grease, so inherent in using GUI is the assumption that meat was the main goal of foraging excursions, which may not always have been the case. For each of the nine ungulate anatomical regions, GUI values are summed for elements in a given anatomical region (Appendix H). Ungulate anatomical regions are arranged in order of increasing GUI values. Note that previous presentations of portions of these data (Starkovich 2009; Starkovich and Stiner 2010) did not arrange anatomical units in this order. Fallow deer body part profiles, which includes data for indeterminate medium ungulate axial elements, are presented by layer in Figure 6.13 in ascending order of GUI values (from Binford 1978; data from Appendix H). K-S tests indicate that fallow deer anatomical representation is uniform in layers Mediterranean backed bladelet layer III’, Aurignacian layer IIIb-d, and Middle Paleolithic layers VIII and XXa-XXb (Table 6.15). Not 260 surprisingly, antlers, which have the lowest cumulative GUI values of any anatomical region are poorly represented in all layers. This may also be due to hunters targeting fewer adult males, or hunting during the summer, after antlers are shed. The upper hind limb region, which has the highest cumulative GUI value of any anatomical region, is well-represented only in the Mediterranean backed bladelet industry layer (III’), the middle Aurignacian (IIIeg) and Middle Paleolithic layer XVIII-XIX (Figure 6.13). These layers also contain high numbers of lower hind limbs, which may have been attached to the upper limbs during transport. Three assemblages, from the lower Aurignacian (IV), Middle Paleolithic layer VIII and Middle Paleolithic layer XXa-XXb, are unusual because they contain high numbers of lower front limbs, without corresponding frequencies of upper front limbs (Figure 6.13). Density-mediated attrition may be a biasing factor in the Middle Paleolithic layers, but a similar relationship does not exist for Aurignacian layer IV (Chapter 4). The lower Aurignacian is unusual in other ways, however: tooth to bone-based MNE counts indicate an extreme bias toward teeth in this layer (Chapter 4), and it has the widest range of ungulate species of all sizes (Chapter 5). Figure 6.14 depicts the anatomical representation of ibex and red deer in the Aurignacian layer IV (data from Appendix N). Both ibex and red deer have a more uniform representation of anatomical regions as compared to fallow deer in layer IV (Table 6.15). Horn and antler elements are lacking from both taxa, as are axial elements for red deer and upper front limbs for ibex. Lower utility regions tend to be more heavily represented for both taxa (Figure 6.14). GUI values for the nine ungulate anatomical regions fall into four categories: low, medium, high, and highest utility values (Table 6.16). Because the differences among these 261 Figure 6.13 Body part profiles for fallow deer in the layers with large sample sizes, ranked by increasing GUI values (data from Appendix K). 262 Table 6.15 K-S tests for ungulate body part profiles by layer. *Asymp. Sig. > 0.05, data do not differ from a uniform distribution. Figure 6.14 Body part profiles for red deer and ibex in lower Aurignacian layer IV, ranked by increasing GUI values (data from Appendix N). 263 four categories are quite large, a second set of body part profiles were constructed based on aggregating the nine anatomical regions into four groups of increasing utility. This provides a clearer method for determining if higher or lower utility regions were transported to Klissoura Cave 1. If foragers selectively transported body parts based on their general utility, a greater representation of high-utility anatomical regions is expected in the archeofaunas. Figure 6.15 presents medium ungulate body part profiles aggregated by cumulative GUI values (from Binford 1978). It is striking that the highest utility regions based on GUI, specifically the upper hind limbs and axial skeleton, are poorly represented in all layers. This is not surprising, given the lack of axial elements in the assemblages (Figure 6.13), and may be indicative of post-depositional attrition. Alternatively this could reflect processing strategies at the kill; perhaps meat was removed from the axial skeleton and transported to the site, while bulky axial elements were abandoned in the field. The aggregated body part profiles for most layers are dominated by high-utility regions, followed by medium- and lowutility regions. Medium utility anatomical units are dominant in Aurignacian layers IV and Table 6.16 Cervid anatomical regions (following Stiner 1991), ranked in order of GUI values derived for caribou (from Binford 1978). 264 Figure 6.15 Body part profiles for fallow deer in the layers with large sample sizes, collapsed into regions with low, medium, high and highest GUI values (from Table 6.16). Middle Paleolithic XV-XVII. This may reflect the importance of an additional resource: bone marrow. The aggregated body part profiles discussed here are based on the GUI, which 265 correlates closely with the meat index (Table 6.12). Marrow index values, on the other hand, are highest for the lower front and lower hind regions (Appendix H), which were aggregated into the GUI-based medium- and high-utility regions, respectively. It seems, therefore, that body part transport decisions may have been based on more than just quantity of meat, at least during some time periods. Patterns of ungulate carcass transport and butchery differ markedly from those for smaller game animals. Whereas small game animals can easily be transported to a site in whole form, the transport of ungulate body parts often has several motivating factors including the value of meat, bone marrow and bone grease. In zooarchaeology, the specific skeletal elements are by necessity used as a proxies for the presence of meat. However, field butchery strategies may involve removing portions of meat from bulky skeletal elements that are never introduced to a habitation site. Further, the desire for certain nutritional requirements (e.g., fatty bone marrow during lean times) or availability of specific technologies (e.g., vessels for bone grease rendering) may expand the goals of foraging excursions beyond simple meat procurement. There is no clear preference at Klissoura Cave 1 for the transport of the highest-utility elements, though high- and medium-utility elements (based on GUI) are consistently present in the site during all layers. MORTALITY PROFILES Establishing ungulate mortality profiles in an archaeological assemblage provides information related to herd structure and human hunting strategies, as well as the seasonality of the occupation of a site. Several lines of evidence yield complementary data, including 266 tooth eruption and wear patterns, long bone epiphyseal fusion, and the presence of fetal or neonate remains in an assemblage. Age profiles are constructed using eruption and wear schedules for teeth, which are well-established for many mammals (Deniz and Payne 1982; Grant 1982; Hillson 2005; Lowe 1967; Magnell 2006; Severinghaus 1949). The dental ages of ungulates are estimated from the mandibular deciduous fourth premolar (dP4) or the mandibular fourth premolar (P4). It is possible to use the mandibular third molar (M3) in place of P4, though the use of the dP4 and P4 is preferred in this study because the deciduous tooth must be shed before the adult tooth comes into wear, which is not always the case with the M3. Additionally, the dP4 and P4 are highly diagnostic elements even when broken. Tooth side is taken into account, and only permanent teeth that have some occlusal wear are considered among the permanent tooth specimens to avoid double counting of individuals (Stiner 1990). Wear stages follow the three-cohort system presented in Stiner (1990), with the age cohort data collapsed into the following categories: juvenile, prime-aged adult and old adult. The juvenile stage includes all animals that died between the time of birth and the shedding of their dP4. The division between prime-aged adults and old adults is set at approximately 65% of the potential life span (see Stiner 1990). No attempt is made in this system to assign numeric ages to tooth specimens. Tooth eruption schedules are not known for all ungulate species at Klissoura Cave 1, so similar species are used as proxies. Standards for white-tailed deer, published by Severinghaus (1949), are applied to fallow deer and data on red deer are from Lowe (1967). Developmental data from Hillson’s (2005) discussion of domestic cattle are applied to auorchs, data from Angora goats provided by Deniz and Payne (1982) are used for ibex and 267 chamois, and data from Magnell (2006) are applied to wild pigs. Tooth eruption and wear data are available in Appendix O. Figure 6.16 Tripolar graph indicating living structure and mortally models in an ungulate death assemblage. Once ungulate tooth wear stages are established and individuals are organized into the three-cohort system, the relative proportions of ungulate of different ages are plotted as percentages of 100 in a tripolar graph. This approach allows comparison of the archaeological data to known living structures and mortality patterns (Figure 6.16, following Stiner 1994; 2005). Generally, hominin hunters tend to target prime-aged adult ungulate prey or to choose age groups randomly upon encounter. This pattern was well-established early in the Middle Paleolithic and spans into the Holocene in many regions (Adler and Bar-Oz 2009; Adler, et al. 2006; Enloe 1997; Gaudzinski and Roebroeks 2000; Hoffecker, et al. 1991; Klein and Cruz-Uribe 1996; Pike-Tay, et al. 1999; Speth and Tchernov 1998; Stiner 1990, 268 1994, 2005). In certain situations, particularly during the Natufian in the Levant, human hunters relied more heavily on juvenile animals, which in some instances is linked to human hunting pressures and resource stress (Davis 1983; Munro 2001, 2004; Speth and Clark 2006), although seasonality and other motives (such as skin-hunting, see Stiner 1990) must also be taken into account. Tooth-based age data for ungulate species are presented in Appendix O; Table 6.17 contains an abridged version of these data. Sample sizes by layer are too small to reach any firm conclusions regarding age structure at Klissoura Cave 1 by cultural layer, but collapsing data from the Middle Paleolithic layers and the Upper Paleolithic (and one Epigravettian specimen) into two groups indicates somewhat different hunting patterns between the two periods. Relative proportions of juvenile, prime-age adult and old adults exploited in the two periods at Klissoura Cave 1 are plotted on tripolar graphs in Figures 6.17 and 6.18. Figure 6.17 includes all ungulate species at the site and Figure 6.18 only includes fallow deer. Ninety-five percent confidence intervals that take into account sample size are included in the plots (following Weaver, et al. 2011). Two plots are presented because the living structure of fallow deer differs from many other ungulate species (e.g., Stiner 2005:Fig. 11.11). In both cases, prime-aged adult animals were targeted ubiquitously in the Middle Paleolithic layers, while the Upper Paleolithic and later layers have a non-selective mortality pattern. There is little overlap in the confidence intervals of the samples, in either plot, indicating different hunting strategies in the two periods. Another method for determining age-at-death is examining fusion states of long bone epiphyses. Long bone elements fuse in a known, predictable sequence for many species. Fusion data are useful in understanding juvenile age structures at Klissoura Cave 1 and 269 Table 6.17 Age distribution for ungulates in the Klissoura Cave 1 assemblages, based on tooth eruption data (see Appendix O for more detail). 270 Figure 6.17 Proportions of ungulate age groups exploited at Klissoura Cave 1 during the Middle and Upper Paleolithic based on tooth eruption and wear. During the Middle Paleolithic, prime-aged adult animals were targeted. More juveniles were incorporated in Upper Paleolithic subsistence pursuits. Circles indicate 95% confidence intervals (following Weaver et al. 2011) (data from Table 6.17). Figure 6.18 Proportions of fallow deer age groups exploited at Klissoura Cave 1 during the Middle and Upper Paleolithic based on tooth eruption and wear. During the Middle Paleolithic, prime-aged adult animals were targeted. Upper Paleolithic patterns are closer to the living structure of fallow deer. Circles indicate 95% confidence intervals (following Weaver et al. 2011) (data from Table 6.17). 271 helping to establish the season during which a site was occupied. Though all available fusion data for ungulates are presented in Appendix P, five elements are the focus of the analysis: proximal first phalanx, distal tibia, distal calcaneum, proximal femur, and distal metapodials (following Stiner 2005:211, though radii are excluded due to small samples in Upper Paleolithic layers). Percentages of unfused specimens are calculated by element and the values plotted in Figure 6.19. Middle Paleolithic layers and the Upper Paleolithic and later layers were combined into two groups, respectively, to create more robust samples, though this may mask variation within each time period. As with tooth eruption data, fusion schedules are not known for all of the ungulate species in the Klissoura Cave 1 assemblages, so in some cases related taxa are used as proxies. Fallow deer fusion data are from Kersten (1987) and the reference data applied to red deer come from Schmid (1972). Bone fusion data applied to ibex are taken from Noddle’s (1974) study of domestic goats, fusion data from domesticated equids (from Silver 1969) are used for wild ass, and data from domestic pigs (Schmid 1972; Silver 1969) are applied to wild pigs. Table 6.18 lists fusion stages for selected fallow deer long bones (Stiner 2005:210). As with aging data based on tooth wear and eruption, layer-specific fusion sample sizes are small so the Middle Paleolithic and the Upper Paleolithic and later layers are again combined into two groups. These data are then graphed in Figure 6.19, which is comparable to Figure 11.12 in Stiner (2005:211). No major differences are apparent between the Middle and Upper Paleolithic layers based on the state of fusion of fallow deer specimens. 272 Figure 6.19 Percentages of juveniles based on five selected elements, based on epiphyseal fusion, following Stiner (2005), figure 11.12. Selected elements fuse between 12 and 24 months (Kersten 1987). No major differences exist between the ages of juvenile animals exploited in the Middle and Upper Paleolithic (data from Table 6.17). Infant and neonate individuals are also represented in the death assemblage. The small size of fetal/neonate individuals probably belong to unborn animals, based on comparisons to specimens in known-age collections (Figure 6.20). Fetal or neonate remains by layer are combined in Table 6.19; most of the remains are from fallow deer. Proportions of adult to fetal or neonate fallow deer remains are plotted in Figure 6.21. There seems to be 273 Table 6.18 Fusion schedules for five elements for fallow deer. Note radii are excluded from Figure 6.19 because of small samples in the Upper Paleolithic layers. a higher proportion of adult to fetal or neonate fallow deer in the Middle Paleolithic layers, though the trend is not quite significant (n = 9, rs = -0.633, p = 0.067). Several interesting points are apparent concerning mortality profiles at Klissoura Cave 1. Ungulate mortality profiles are instructive in establishing seasonality of occupation, as well as shifts in the ages of individuals targeted by human hunters. Antler is present mainly in the Upper Paleolithic and later layers. Male fallow deer possess antlers from roughly July to April (Chapman and Chapman 1975), so some individuals in the assemblage were hunted during these months. However, antler may also have been collected and curated over long periods for tool manufacture, though the presence of some unshed antler bases still attached to crania indicates that this was not always the case. Fetal remains recovered in all layers of the site are extremely small compared to modern stillborn specimens (Figure 6.20). Fallow deer and ibex typically give birth in May or June (Braza, et al. 1988; Schaller 1977; Spiess 1979), so the size of the fetal remains indicates that the site was used during the winter or early spring. Similarly-sized fetal specimens are found throughout the Middle and Upper Paleolithic layers, so the shelter was 274 Figure 6.20 Comparison of fallow deer fetal remains from Klissoura Cave 1 and a modern stillborn goat. (top) Fetal femur from Upper Paleolithic disturbed zone (6-7a), (bottom) Fetal humeri from Middle Paleolithic layers (VII) (bottom) and (XVa-XVI) (top). used during the cold months throughout its occupation history, though it could have been occupied at other times as well. Ungulate mortality data based on tooth eruption and wear indicate generally different exploitation patterns between the Middle and Upper Paleolithic layers, with a dominance of prime-aged adult ungulates in the Middle Paleolithic, and a non-selective pattern in the later periods. Both cases fit comfortably within the range of observed ungulate mortality patterns for the Middle Paleolithic through Epipaleolithic across the Mediterranean, including 275 patterns specific to fallow deer (Stiner 2005:206). Examining mortality profiles from sites in Italy, Israel, Lebanon and Turkey does not indicate a clear shift in ungulate hunting patterns in later time periods, until the Neolithic (Stiner 2005). However, Klissoura Cave 1 provides an interesting contrast, as the same ungulate species (fallow deer) dominates the assemblages throughout most of the sequence. The shift from a prime-dominated to non-selective Table 6.19 Fetal elements identifiable to species by layer. 276 Figure 6.21 Proportion of fetal to adult fallow deer by MNI. Data from Table 6.19. mortality pattern could represent mild pressure on deer populations in the later periods or it might relate to how hunters exploited herd structures (e.g., focusing on bachelor males versus maternal herds). CONCLUSIONS Human modifications on the faunal remains from Klissoura Cave 1 are apparent throughout the sequence. The assemblages were principally collected and altered by human hunters as they targeted prey, transported carcasses to the site, and butchered the remains. The zooarchaeological evidence clearly indicates that Klissoura Cave 1 was used largely as a 277 base camp or residential site for the duration of its occupation. However, some of the occupations were more intensive than others (see Chapter 7). Small and large game animals were treated differently with respect to transport and butchery. Changes in butchery patterns and transport decisions offer only subtle, if any, indicators of resource intensification at Klissoura Cave 1 from the Middle Paleolithic through Mesolithic periods. Much of the variation, such as changes in the frequency of burning on faunal remains, may be explained by changes in site use, which are addressed in Chapter 7, or limited horizontal sampling during the excavations. Some shifts in butchery patterns, such as an increase in transverse fractures on ungulate long bones, point to more systematic butchery efforts or standardization in later time periods. One pattern that is clearly different between the earlier and later periods is the selection of individuals by age within a herd. Based on central-place foraging models, it is expected that foragers utilizing Klissoura Cave 1 would transport higher-utility elements to the site, the further away they traveled to procure game. According to several analyses presented above, foragers did not preferentially transport the highest-utility elements during any time period. This suggests that either foragers at the site did most of their hunting nearby, or that the GUI or FUI are not the best proxies for ranking elements based on their utility. For example, axial elements have some of the highest utility values in an ungulate skeleton (Appendix H; Binford 1978). However, the axial skeleton is rather bulky and may be abandoned in-field. Vertebrae are useful for bone grease rendering, but might not be transported in the absence of this technology. There is no evidence for bone grease rendering at Klissoura Cave 1. Further, nutrient-rich bone marrow may be an important goal of a foraging excursion, and bone marrow was routinely processed at Klissoura Cave 1 in all time periods. This is particularly 278 the case in Aurignacian layer IV, when ungulate third phalanges were routinely opened for marrow, despite their low nutritional yield. Interestingly, this is one of the layers at the site with a positive correlation between %MAU and the bone marrow index. Following the patch-choice model, when Aurignacian layer IV was deposited, the marginal benefit from intensively processing bone marrow provided a higher return than the combined travel and handling costs of pursuing the next kill. This may indicate increased travel distances at this time, though this is not evidenced from body part transport strategies (though marrow-rich elements were preferentially transported during this period). The patchchoice model may not be relevant to the absence of bone grease rendering at Klissoura Cave 1, though the exploitation of low-utility bone grease often indicates resource stress or intensification. Resource stress is indicated in certain periods based on intensive marrow processing (e.g., Aurignacian layer IV) or the exploitation of low-return species (e.g., the Mesolithic and Epigravettian). However, no evidence has been found for bone grease rendering as early as the Aurignacian, and Mesolithic and Epigravettian occupations were too ephemeral at Klissoura Cave 1 to test for grease rendering. Middle Paleolithic hunters at Klissoura Cave 1 targeted the same ungulate species as were exploited during later periods, but they chose a different subset of the population. More fetal or neonate remains were identified in the Middle Paleolithic layers, so overall pregnant females were hunted more frequently, before that year’s offspring could be born (but there is only a weak temporal trend). It is possible that this was the result of occupying the site during the winter, when females were in better physical condition than males. Season of occupation, in addition to a focus on maternal herds, may explain the prime-aged bias in the Middle Paleolithic sample. It is likely, based on the size of fetal remains in all layers of the sequence, 279 that the site was utilized during the winter or early spring throughout the sequence. In the Upper Paleolithic, human hunters exploited ungulates less selectively with respect to age and possibly with a slight bias to juveniles. Coupled with the inclusion of lower-ranked prey species in later periods discussed in Chapter 5, this latter observation might provide an argument for resource intensification linked to the growth of human populations in southern Greece by the end of the Paleolithic, with important variation within time periods in the sequence that relates to site use. 280 CHAPTER 7: SPATIAL VARIATION WITHIN CULTURAL LAYERS INTRODUCTION Spatial variation is often apparent across an archaeological site, reflecting different activity areas or preservation biases in different parts of the site. Though specific activity areas are more easily discerned at later archaeological sites with permanent architecture or more formal structures, lateral variation is documented at some Paleolithic sites (Enloe 1991; Enloe and David 1989, 1992). At the Magdalenian site of Pincevent in the Paris Basin, for example, Enloe and colleagues (1991; Enloe and David 1989, 1992) found through refitting faunal elements that reindeer carcasses were shared across the site, with certain high or lowranking carcass parts tending to cluster around different hearths. This is interpreted as evidence of provisioning some members of the group who were perhaps unable to hunt on their own, such as widows or grandparents (Enloe 2003b). Spatial distributions indicative of sharing are also found at other forager sites, in ethnographic and archaeological contexts (e.g. Binford 1984a; Waguespack 2001; Yellen 1977). Admittedly, these cases are single-occupation or short-term sites with extraordinary preservation. It is unrealistic to assume that a similar degree of resolution will be observable at a deeply stratified site that represents a palimpsest of behavioral events. However, examining variation within the archaeological layers potentially serves other functions, beyond the unique circumstances mentioned above. First, analyzing the spatial distribution of archaeofaunas at a site is a means by which taphonomic processes (Chapter 4) can be better understood. An example of this is the identification of a bone-poor zone at Hayonim Cave, where the faunal remains underwent chemical diagenesis in one area of the site and through 281 much of the depth of the sequence (Stiner, et al. 2001). This area likely would not have been obvious if faunal remains were considered in bulk units by layer and could even have been misidentified as a cleared activity area. A second function that an intra-site spatial analysis serves is to evaluate whether different areas of the site tended to be used for specific activities based on the physical characteristics of the location, such as the preferential placement of hearths to protect them from wind. In this case we expect to see hearths or other features constructed in the same area through the years. It is possible that faunal remains are also unevenly distributed, reflecting distinct activities within a site. In this chapter, the lateral distribution of faunal remains at Klissoura Cave 1 is examined in order to determine if there are any areas lacking bone that may be explained by chemical diagenesis or exposure to weathering (such as outside the drip line of the shelter, or in other unanticipated bone-poor zones). Variation in site use is also discussed in terms of hearth building, and whether animal processing was focused in certain parts of the site. Some archaeological layers, specifically the lower Aurignacian (IV) contain many well defined features such as clay-lined hearths and a possible stone-lined structure (Karkanas, et al. 2004; Koumouzelis, Ginter, et al. 2001; Pawlikowski, et al. 2000; Stiner 2010). The use of space in this layer is examined in terms of these features. Differences in sample size based on discontinuous archaeological layers or the studied sample are taken into account. METHODS Prior chapters have focused on behavioral changes that occurred diachronically through the Klissoura Cave 1 sequence. This chapter focuses on spatial variation of faunal materials and features within each layer. The goals of this chapter are to consider horizontal 282 variation of taphonomic processes within each cultural unit, and to distinguish behavioral differences related to site use or site topography. This is done by constructing a series of plan maps for each archaeological layer that depict a specific set of variables. Each map represents a different cultural layer (i.e., Mesolithic layer 3-5a) and is based on the 1x1 meter grid system used to excavate the site (see Chapter 2). The first map of each layer provides a baseline for understanding the subsequent distribution maps. NISP counts for each 1x1 meter unit are divided by the total NISP count for a given cultural layer in order to calculate the percentage of the assemblage that came from each unit. In some cases, the percent NISP count for an excavated unit is zero, which either indicates that the unit was excluded from this sample because it was included in an earlier study (Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001; Tomek and Bocheński 2002), or that the layer was thin or discontinuous in that part of the site. Values for subsequent maps (e.g., green bone breaks or burning frequencies) are calculated for each 1x1 meter unit. For example, the number of burned bones from square AA1 from Mesolithic layer 3-5a is divided by the total NISP count for that specific provenience in order to calculate the percent of burned bones in that unit. The second map in each set examines the spatial representation of green bone fractures (spits, transverse, or spiral fractures) across the site to determine if processing activity areas are apparent, based on the percentage of fractured medium ungulate specimens in each 1x1 meter square per layer. Medium ungulates were chosen for this analysis because fallow deer are the dominant large prey animal in Klissoura Cave 1 (Chapter 5). Green bone fractures are examined as opposed to tool marks because most of the Upper Paleolithic faunal remains are covered with concretions that obscure the surfaces of the specimens (Chapter 4). 283 Fresh, or green bone, fractures may occur during marrow processing or when a carcass is being partitioned during butchery. A third plan map for each layer indicates the frequency of burning by 1x1 meter unit in a given layer. Areas with high frequencies of burning may indicate clusters of hearths or a discreet dumping area. Conversely, a moderate frequency of burning that is consistent across the site may indicate a high degree of trampling or movement of artifacts across the site. Plan maps drawn during the excavation by K. Sobczyk were examined to determine hearth locations; these are indicated on the burning frequency maps. An “H” appears in the lower left corner of excavation units that contain hearths in each layer. The number of hearths is not specified in each unit because hearth features (particularly during Aurignacian layer IV) are often superimposed upon one another and are difficult to distinguish. No hearth information was available for Middle Paleolithic layers below VIII. The data from each set of maps described above are discussed by archaeological layer at Klissoura Cave 1. Note that bones from squares in column B were not included in any of the faunal samples from this study, though plan maps indicate that they were excavated in the upper layers (K. Sobczyk, plan maps); these units may represent the original test excavation. In the lower Aurignacian (layer IV) a rock-lined structure was identified in squares AA2, AA3, BB2 and BB3, from 145 to 170 cm below the surface (Koumouzelis, Ginter, et al. 2001; Stiner 2010). Stiner (2010:298) notes that the majority of shell ornaments from this layer cluster within the man-made structure. In this layer, burning frequencies and butchery damage (also based on green bone fractures described above) were compared between the structure and surrounding areas. 284 Some of the most interesting features at Klissoura Cave 1 are the dozens of clay-lined hearths in the Aurignacian and overlying Upper Paleolithic industries (III’ and III”) (Karkanas, et al. 2004; Kot 2009; Pawlikowski, et al. 2000). The clay-line hearths co-occur with simple unlined hearths. Understanding the spatial distribution of artifacts in reference to these features is a logical step in understanding site use. However, initial comparisons between faunal remains found in the hearths and specimens found outside the hearths proved to be inconclusive, simply because very few faunal remains were actually recovered in the clay-line hearth features. Karkanas et al. (2004) note that fallow deer, hare, great bustard and rock partridge remains were found associated with the clay-lined hearths. Bird and hare remains, however, have some of the lowest rates of burning in the entire Klissoura Cave 1 assemblages (see Chapter 6), so it is difficult to argue that these specimens were not deposited on the features after they were used. In addition, many of the hearths are superimposed on or overlap one another (Karkanas, et al. 2004; Kot 2009; Pawlikowski, et al. 2000), so it is difficult to assign faunal materials to one specific hearth. Finally, there is no single cluster of clay-lined hearths at Klissoura Cave 1 (with the exception of the rock-lined structure in layer IV, see below) making it difficult to associate presumed activity areas with parts of the site specifically containing clay-lined features. RESULTS The first map of each set is simply an illustration of the excavation units and faunal sample included in the present study. Squares AA4 and BB4 were excavated in the 2005 and later seasons, after the author had joined the Klissoura Cave 1 team, so all faunal remains from these units are included in the analysis. The same goes for the Middle Paleolithic layers, 285 which were completely analyzed by the author. Discontinuity of some of the layers is also apparent, such as the upper Aurignacian (IIIb-d) which does not reach the outer margin of the site, the Early Upper Paleolithic (Uluzzian layer V) which is has a more obvious contact with overlying layers outside of the drip line of the shelter, and Middle Paleolithic (layer X) which is also more obvious toward the drip line (see also Karkanas 2010). The maps show that the excavation area was reduced downwardly from 14 m2 to an 11 m2 area by Middle Paleolithic layer XI-XIV, and again to a 9 m2 area in Middle Paleolithic layer XV-XVII. The final two layers of the excavation only included 6 m2. Mesolithic layers 3-5a Results of the Mesolithic spatial distribution of faunal remains are presented in Figure 7.1. The majority of the sample is from the squares at the outer edge of the shelter, at the drip line and beyond, in three units (Figure 7.1a). Green bone breaks on medium ungulate remains are considerably more frequent at the eastern end of the excavation, which may indicate that more processing occurred in this area (Figure 7.1b). This may partially be a product of sample size, however, since squares AA3 and AA4 only contain 10% of the total sample from this layer. Green bone fractures may be underrepresented in this part of the site; on the other hand, square BB4 contains a quarter of the sample and still has comparatively low frequencies of broken bone. The amount of burning in this upper layer is fairly light and is highest at the outer margin of the site (Figure 7.1c). Few if any hearths were defined in the Mesolithic layers, and these are not associated with the squares containing the highest rates of burning, so any concentrations of mildly burned bone may indicate deflated or eroded hearths or dumping areas associated with fireplace cleaning. 286 Figure 7.1 Plan maps for Mesolithic (3-5a), NISP = 256. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. Epigravettian layers IIa-d The Epigravettian (IIa-d) layers contain the smallest faunal sample in the Klissoura Cave 1 faunal series, only encompassing three excavation units (Figure 7.2a). All of the specimens from squares AA4 and BB4 display green bone breaks, but only half of the remains from AA3 are fractured (Figure 7.2b). Like the situation in the Mesolithic layers, the 287 Epigravettian has a low incidence of burning damage (Figure 7.2c) and only a few hearths were recorded during excavation (K. Sobczyk, plan maps). Figure 7.2 Plan maps for Epigravettian (IIa-d), NISP = 67. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. Mediterranean backed bladelet industry layer III’ The sample from the Mediterranean backed bladelet industry layer III’ is much more robust than any of the overlying samples, though it is still somewhat limited spatially (Figure 288 7.3a). Butchery damage on medium ungulate remains occurs in very high frequencies in all units (Figure 7.3b). Debris from processing activities was discarded in many parts of the sampled area, or processing took place in all parts of the site. Green bone breakage has little relationship with the frequencies of burning damage. Burning frequencies are slightly higher in this layer and grade from less burned toward the back of the shelter (northwest) to more burned in the southeastern excavation units, beyond the shelter drip line (Figure 7.3c). Most of the units included in the faunal sample contain hearth features (K. Sobczyk, plan maps). Figure 7.3 Plan maps for Mediterranean backed-bladelet industry (III’), NISP = 1,359. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. 289 Upper Paleolithic (non-Aurignacian) industry layer III” The spatial distribution of faunal remains from the Upper Paleolithic (nonAurignacian) industry in layer III” is presented in Figure 7.4a. The sample is large but incomplete, as some of the materials were included in earlier studies (Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001). The faunal material included in this study clusters to the northeastern margin of the excavation. The frequency of green breaks on medium ungulate remains is high in all areas of the excavation, particularly in the southern Figure 7.4 Plan maps for Upper Paleolithic (non-Aurignacian) industry (III”), NISP = 1,617. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. 290 and western units (Figure 7.4b). Marrow processing activities may have concentrated in the southwestern part of the excavation areas, but without full horizontal coverage of the site it is difficult to be sure. Burning damage is higher in this layer than in the later occupations, particularly in the outermost excavation units of the cave (AA4 and BB4), in addition to one of the northeastern units (Figure 7.4c). Burn frequencies are lower in the middle of the sample area, particularly in unit AA2. Hearths are found in several units included in the study sample, but do not correspond to squares with high incidences of burning. The frequency of burning and butchery damage do not relate to one another. Upper Aurignacian layers IIIb-d The bulk of the faunal materials from the upper Aurignacian (IIIb-d) come from the northeastern column of excavation units (CC1 and CC2) (Figure 7.5a). Faunal remains were absent from squares AA4 and BB3, which may indicate that layer IIIb-d was very thin or absent in this part of the site. Green bone breaks on medium ungulate bones are very common in this layer, particularly in the northern part of the excavation (Figure 7.5b). The latter concentration could indicate a processing area, or butchered bones may have been discarded in this part of the site. Burning frequencies are low in layer IIIb-d, more along the lines of layers III’ and above (Figure 7.5c). The row of excavation units toward the back of the shelter has particularly low incidences of burning. Hearths were found in only one square in this layer (K. Sobczyk, plan maps). 291 Figure 7.5 Plan maps for Aurignacian (IIIb-d), NISP = 1,158. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. Middle Aurignacian layers IIIe-g The spatial distribution of faunal remains from layers IIIe-g is presented in Figure 7.6a. Most of the remains came from two parts of the site, the two squares in row 4 and unit CC2. Some of the gaps in the faunal map are due to certain materials being excluded because they were studied previously (Koumouzelis, Ginter, et al. 2001; Koumouzelis, Kozlowski, et al. 2001). Additionally, one box of fauna was excluded from this study due to time 292 constraints during the analysis. Green bone fractures on medium ungulate remains are common, particularly in the center of the excavated area (Figure 7.6b). Areas with high incidences of fresh breaks have variable frequencies of burning. The middle Aurignacian layer exhibits a slightly higher rate of burning than the upper Aurignacian (IIIb-d). Many hearths were recorded in this layer, across the extent of the site (K. Sobczyk, plan maps). Burned faunal material is most common in the eastern units that were included in the sample, as well as in square AA2 (Figure 7.6c). Figure 7.6 Plan maps for Aurignacian (IIIe-g), NISP = 1,699. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. 293 Lower Aurignacian layer IV The lower Aurignacian (IV) contains the largest faunal sample of any of the cultural layers at Klissoura Cave 1, and the material is fairly evenly distributed across the horizontal extent of the site (Figure 7.7a). A slightly higher concentration of faunal remains is evident in unit AA2 and in two of the squares directly adjacent to this one. Frequencies of butchery damage are high in layer IV, particularly in the squares at the margins of the sampled area (Figure 7.7b). The areas with the highest incidences of green bone fractures on medium Figure 7.7 Plan maps for Aurignacian (IV), NISP = 3,110. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. 294 ungulate remains are also in the squares with the smallest sample sizes. Burning damage is also common in layer IV, which is not surprising due to the large number of clay-lined and simple hearths in the layer. Burned faunal remains are concentrated in square CC1, as well as in the southeastern units (Figure 7.7c). These are some of the areas with high numbers of hearths, though hearths are recorded in almost all of the units of the excavation (K. Sobczyk, plan maps). Some of units with high concentrations of burned bone may represent dumping areas after fireplace cleaning events. In addition to the spatial distributions described above, the lower Aurignacian (IV) also contains a possible rock-lined structure (see Figure 7.8 for a map of the structure drafted by M. C. Stiner). A set of maps similar to the ones presented for the different cultural layers were constructed for the arbitrary levels in which the structure was observed (145-170 cm below the surface) (Figure 7.9; squares containing the structure are highlighted). Faunal remains tend to concentrate in units outside of the boundaries of the rock-lined structure (Figure 7.9a). It is evident from the plan maps that the majority of the hearths from these levels are outside of the squares that include the structure (Figure 7.8; K. Sobczyk, plan maps). Processing damage in the form of green bone breaks on medium ungulate remains is somewhat lower in the units that contain the structure (Figure 7.9b), so butchering activities may have focused outside of this area. There is no distinguishable pattern between the structure and surrounding units in terms of rates of burning of the faunal remains (Figure 7.9c). This may be because the heavy anthropogenic input of the subsequent Aurignacian occupations quickly filled the structure area after its use was discontinued. 295 Figure 7.8 Drafted plan maps in Aurignacian (IV) from 145 to 180 cm below datum by 5cm cuts. Hearth features are indicated in black and dark grey, limestone rocks in white. The light grey background represents sedimentary matrix. The shelter feature is apparent in cuts 145 to 170, where rocks are common and hearths are absent. From Stiner (2010). 296 Figure 7.9 Plan maps for Aurignacian layers (IV) containing the structure, NISP = 2,049. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. Units containing the structure are outlined in black. Early Upper Paleolithic or Uluzzian layer V The Early Upper Paleolithic, or Uluzzian, layer V is thin and discontinuous (Figure 7.10a) (Karkanas 2010). This layer is fairly diffuse toward the back of the shelter, but has a clean contact with underlying and overlying layers beyond the drip line of the cave (Karkanas 2010). Not surprisingly, the distribution of faunal remains reflects this (Figure 7.10a). Frequencies of green bone breakage are high across layer V (Figure 7.10b). Butchery 297 activity seems to have been higher toward the outside of the shelter, but the limited spatial extent of the sample could undermine this interpretation. Overall, faunal remains in the layer display high frequencies of burning, particularly in the northern and western portions of the sample (Figure 7.10c). These clusters may indicate concentrations of deflated hearths or hearth cleaning areas, but the sample from excavation unit A3 is small. Hearths were found Figure 7.10 Plan maps for Early Upper Paleolithic (Uluzzian V), NISP = 222. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. 298 in many units of the Uluzzian layer, but because the layer is so thin, it is unclear if they actually belong to the hearth-rich lower Aurignacian (IV). Middle Paleolithic layer VIII The faunal sample from Middle Paleolithic layer VIII clusters in the western part of the excavation area, with very few specimens coming from the northern and southern Figure 7.11 Plan maps for Middle Paleolithic (VIII), NISP = 1,569. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. H = unit with hearths. 299 portions of the site (Figure 7.11a). Incidences of green bone fractures on ungulate remains are high across the excavation area, with the lowest frequencies in square BB2 (Figure 7.11b). This may indicate that marrow processing and other activities took place across the site during this period. Hearths were recorded toward the back of the shelter (K. Sobczyk, plan maps), though none are lined with clay. Rates of burning damage on bone are quite variable (Figure 7.11c); some squares with high rates of burning contain hearths (i.e., CC1), while others do not. Perhaps some areas with high frequencies of burning represent areas where hearth cleanings were dumped. Sample size has little effect on rate of burning. Middle Paleolithic layer X Middle Paleolithic layer X has the smallest sample of the Middle Paleolithic layers. It is a fairly small, discontinuous layer that was only found at the outer edge of the site (Figure 7.12a). Faunal remains from this layer are not very evenly distributed across the site, with a concentration of materials in unit AA3 (Figure 7.12a). Green bone fracture frequencies on medium ungulate remains are similar for all units (Figure 7.12b), as are frequencies of burning (Figure 7.12c). In general, this may be a product of sample size or it may be because the layer was horizontally mixed. Middle Paleolithic layers XI-XIV The faunal remains in Middle Paleolithic layer XI-XIV cluster in excavation unit BB2 and adjacent squares (Figure 7.13a). High rates of medium ungulate bone breakage are apparent in all parts of the excavation except the middle area (Figure 7.13b), which may relate to intensive marrow processing in specific areas of the site. The rather uneven 300 Figure 7.12 Plan maps for Middle Paleolithic (X), NISP = 241. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. distribution of bone breakage across the site (particularly as compared to burning frequencies) indicates that the breakage was probably not from post-depositional processes. There is a modest rate of burning on the remains across the layer (Figure 7.13c), which may indicate some amount of horizontal mixing, though fracture patterns suggest otherwise. 301 Figure 7.13 Plan maps for Middle Paleolithic (XI-XIV), NISP = 1,779. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. Middle Paleolithic layers XV-XVII Faunal remains in Middle Paleolithic layer XV-XVII seem to follow a northeastsouthwest gradient. The faunal sample is large, and the frequency of remains is higher in the northern excavation units (Figure 7.14a). Green bone breaks on medium ungulate remains are nearly the inverse of burning damage, with more intensive processing apparent in excavation column BB (Figure 7.14b). Frequencies of burning damage are higher in this layer than in 302 any other, and the most intensely burned areas are in excavation column A (Figure 7.14c). This spatial bias probably is not related to any kind of slope wash or sweeping, or else we would expect similar concentrations for both burning and processing damage. It is possible that more hearths were located in the southwestern part of the site and butchering activities occurred outside of these areas, but hearth distribution data were not available. Figure 7.14 Plan maps for Middle Paleolithic (XV-XVII), NISP = 2,031. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. 303 Middle Paleolithic layers XVIII-XIX Middle Paleolithic layer XVIII-XIX is anomalous relative to all of the other layers at Klissoura Cave 1 because it displays even distributions for all parameters examined in this chapter (Figure 7.15). The sample size for this layer is large, but there is no discernable concentration of faunal remains (with the exception of perhaps excavation unit AA2, see Figure 7.15a). Green bone breaks on medium ungulate remains are common in the assemblage and do not focus in any one part of the excavation (Figure 7.15b). Rates of Figure 7.15 Plan maps for Middle Paleolithic (XVIII-XIX), NISP = 2,199. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning. Darker boxes indicate higher frequencies of listed parameters. 304 burning damage are low if compared with the other Middle Paleolithic layers, and no concentrations of burned materials are evident (Figure 7.15c). One could argue that the even distribution of faunal remains, burn frequency and evidence of butchery is explained by the limited spatial extent of layer XVIII-XIX. However, a similar spatial distribution is not apparent in Middle Paleolithic layer XXa-XXb (see below), which includes the same excavation units, nor is it apparent in some of the younger Paleolithic layers that were also limited spatially. I would argue that this evenness represents a high degree of horizontal mixing in layers XVIII-XIX, possibly from trampling or lateral sweeping of archaeological materials beyond the units included in this sample. Middle Paleolithic layer XXa-b The lowest layer of the Klissoura Cave 1 sequence, Middle Paleolithic layer XXa-b, is spatially limited but has a reasonable sample size (Figure 7.16a). The faunal remains are largely concentrated in the northeastern part of the excavation, in units BB1 and BB2 (Figure 7.16a). This spatial bias may indicate that that the sediments dip or slope to the northeast, or that preservation was a problem in the other units. Green bone breaks on ungulate remains are high in most of the excavation units, with the exception of square BB1 (Figure 7.16b). Burning damage is much higher in the southwestern units, in the squares with a lower representation of faunal remains (Figure 7.16c). It is unclear whether hearths were focused in this area, as maps from this lowest layer could not be located. Interestingly, the square with the largest sample (BB1) has the lowest frequency of burning and butchery damage. 305 Figure 7.16 Plan maps for Middle Paleolithic (XXa-XXb), NISP = 1,106. (a) NISP distribution, (b) frequencies of fresh bone breaks, (c) burning, (d) medium ungulate specimens by body region. Darker boxes indicate higher frequencies of listed parameters. DISCUSSION AND CONCLUSIONS The horizontal extent of the Klissoura Cave 1 excavation is quite limited (maximum 17 m2) but in fact the site was not large. Though few spatial biases in material distributions are apparent, analysis of the horizontal layout of the site by layer is useful on several fronts, including tests for areas of poor bone preservation and loci of repeated activities due to the orientation of the shelter. In addition, evaluating the distribution of the faunal sample 306 horizontally clarifies the limitations of certain layers. The results allow some modest conclusions to be drawn. In terms of site preservation and taphonomic processes, there is no evidence suggesting that faunal remains in certain parts of the site were less likely to preserve than those in other areas, though a lack of data on the volume of sediment excavated makes this difficult to quantify. There is also no indication that remains outside of the drip line of the shelter, which was probably less protected during certain time periods, were less likely to preserve. In fact, many layers in these southeastern units contain the bulk of the faunal materials, sample availability and time limits for primary analysis notwithstanding. Burning damage on faunal remains across the site indicates no consistent vertical clustering between layers. Activity locations often changed, possibly each time the shelter was occupied. The shelter was attractive for habitation, which is clear from the intensity of use during many periods, as well as the flexible arrangement of features within the site. In general, a spatial analysis of the faunal remains from Klissoura Cave 1 does not provide any robust conclusions about site use within the different cultural layers. The results do confirm, however, the conclusion put forth in Chapter 4 that non-human taphonomic processes had little if any effects on the assemblage, and those observed were not confined to specific parts of the site. The spatial analysis also indicates that certain layers, such as Middle Paleolithic layer XVIII-XIX, have a high degree of horizontal mixing, possibly caused by trampling or sweeping. The results of this analysis also add an interesting dimension to our understanding of hearth features and the Aurignacian (IV) rock-lined structure. To begin with, burning and butchering frequencies seem to be unrelated to the spatial distribution of hearths in Klissoura 307 Cave 1. This is not surprising, since hearths locations are highly variable and often overlap. Klissoura Cave 1, or most Paleolithic sites for that matter, cannot be thought of as having long, uninterrupted occupations, even in instances where site use was intense. Each new occupation would introduce more anthropogenic materials into the site, obscuring previous features or patterns of use. As for the possible rock-lined structure, butchery damage is less frequent within the units that contain the structure but burning damage is not, even though hearths do not overlap with this unit. This indicates that faunal refuse filled the structure after it was no longer in use. 308 CHAPTER 8: KLISSOURA CAVE 1 IN CONTEXT: LATE PLEISTOCENE SUBSISTENCE CHANGE IN GREECE AND THE MEDITERRANEAN BASIN INTRODUCTION The preceding chapters have presented an in-depth analysis of the zooarchaeological remains from the Middle Paleolithic through Mesolithic layers at Klissoura Cave 1. There are several goals for this concluding chapter. One is to determine how the Klissoura Cave 1 faunas fit with our understanding of the artifact assemblages of the site, and changes in site use throughout the history of its occupation. In this discussion, the uniqueness of the site is highlighted and placed in the context of the Greek Paleolithic. The second goal is to determine how subsistence data from Klissoura Cave 1 fit with predictions presented in Chapter 1 concerning, among other things, resource depression as a response to human population growth or environmental change. Predictions from prey choice, patch choice and central place foraging models are considered in particular. Also important is a consideration of the full suite of human activities represented by the major layers. The final goal of this chapter is to compare dietary trends at Klissoura Cave 1 during the Late Pleistocene with faunal assemblages from other sites across the Mediterranean Basin. This last part of the discussion provides a broader picture of subsistence change and hominin adaptations over this time span. KLISSOURA CAVE 1 309 Perhaps the most unique feature of Klissoura Cave 1 compared to other Greek Paleolithic sites is the span of the occupations and intact stratigraphy that preserves robust Middle and Upper Paleolithic components, as well as thinner Epigravettian and Mesolithic layers. This allows for large-scale diachronic comparisons of many aspects of site use and changes in stone tool technologies and subsistence patterns. Klissoura Cave 1 is wellpositioned on the landscape, slightly uphill from a perennial stream at the base of a relatively narrow gorge, which made the shelter an attractive camp or habitation site. Prey species were likely targeted as they traveled through the gorge and stopped to drink or feed, or as they passed through the nearby mouth of the gorge from the Argive Plain. There is evidence that site use varied through time, with particularly intensive occupations during the Aurignacian. These and other aspects of site use are expanded upon below. Only slight changes occurred in the lithic assemblages through the Middle Paleolithic, with the lowest layers (XVIII, XIX and XX) having a higher proportion of blades and bladelets than the overlying layers. A few Upper Paleolithic tool types in the youngest Middle Paleolithic layers that were excluded from this study (VI and VII) (Sitlivy, et al. 2007) likely represent localized mixing with younger layers, which is also apparent from the presence of small numbers of shell ornaments in these layers (Stiner 2010). No major changes in faunal use are apparent during the Middle Paleolithic at the site. Hunters focused primarily on prime-aged adult ungulates, fallow deer in particular, and supplemented their diets with tortoises and occasionally hares. Human modifications on carcasses are similar among the Middle Paleolithic assemblages; frequencies of burning and marrow processing are high throughout. A higher proportion of early-term pregnant female fallow deer were hunted during the Middle Paleolithic than in later periods, which may suggest that the site 310 was occupied primarily during the winter, and that adult females were targeted from maternal herds because they were in better nutritional condition than males. The Early Upper Paleolithic or Uluzzian layer V is important for addressing larger questions concerning the Middle to Upper Paleolithic transition. Layer V represents a rare occurrence of the Uluzzian outside of Italy and is several thousand years older (>39,000 BP) than at some of the Italian sites (but see Kuhn, et al. 2010). Unfortunately, the faunal assemblage from this period is smaller than the other layers (NISP < 300), making many comparisons difficult. The Early Upper Paleolithic faunal assemblage has some features in common with the Middle Paleolithic layers, such as comparatively high proportions of tortoise remains and a dominance of fallow deer, though hares are somewhat more important in the Uluzzian, similar to the overlying Upper Paleolithic and later layers. All evidence indicates that Klissoura Cave 1 was occupied intensively during the Aurignacian period. This is particularly striking since the Aurignacian is absent from most Greek Paleolithic sites, particularly those in the north. In Peloponnese, Aurignacian deposits are found at Franchthi Cave and possibly also in Apidima and Kephalari Caves. The Klissoura Cave 1 Aurignacian lithic assemblage is massive (n > 90,000; Kaczanowska, et al. 2010), and the bulk of the ornaments comes from these layers (Stiner, et al. 2010). The Aurignacian also contains hundreds of hearth features, dozens of which are clay-lined structures (Karkanas, et al. 2004; Pawlikowski, et al. 2000), and a possible rock-lined habitation structure in Aurignacian layer IV (Koumouzelis, Ginter, et al. 2001; Stiner, et al. 2010). With the exception of the faunal remains, artifact frequencies tend to cluster inside of the structure, while hearths occur only outside of it. 311 The clay-lined hearth features are one of the most unique components of Klissoura Cave 1, but it is difficult to determine their function because they seem to have been cleaned out after use. Botanical evidence is rare in the features, and the faunal remains found directly on the hearth surfaces do not seem to be directly associated with them. There is no specific reason to believe that the clay-lined hearths served a single purpose; they may have been used for parching plant materials, braising meat, or as brazier-like satellite hearths around a larger simple hearth. One purpose that they probably were not used for is as shallow vessels for bone grease rendering, as evidence of this practice is lacking from the site. What is interesting about these features is their ubiquity throughout the sediments that may span a few millennia at Klissoura Cave 1. To date they are not found at other contemporary sites in Greece or the Mediterranean. The faunal assemblages of the Aurignacian are much more variable than those from the Middle Paleolithic. Aurignacian layer IV is taxonomically richer than the earlier periods, probably because of the appearance of mosaic parkland-forest communities in the area (Albert 2010; Ntinou 2010). Frequencies of hare are also high. In the stratigraphically later Aurignacian layers, fallow deer are the dominant large game species and low-ranked small, fast-game were commonly exploited. There is an increase in rock partridge and great bustard in Aurignacian layers IIIb-d, probably in response to increasingly open vegetation in the vicinity of the site. Evidence for the exploitation of labor-intensive land snails appears at this time. Burned bone is common in the Aurignacian layers, and ungulate crania exhibit higher frequencies of burning than postcrania. Impact and cut-mark data are difficult to quantify in these layers due to ash concretions on the bone surfaces. Bone marrow processing intensifies during Aurignacian layer IV, possibly as a result of prolonged site use. An increase in the 312 importance of marrow extraction is evidenced by the preferential transport of marrow-rich elements and more common processing of marrow-poor terminal ungulate phalanges. There is also a shift in the Aurignacian to a non-selective mortality profile for ungulates, quite different from the prime-dominated patterns of the Middle Paleolithic. This may be partly explained by the main season(s) of occupation at the site. Fetal remains in the Upper Paleolithic layers are similar in size to those found in the Middle Paleolithic, indicating some winter occupations, but fewer fetal remains in general are found in the Upper Paleolithic. The cultural affiliations of Upper Paleolithic layers III” and III’ are distinct from the Aurignacian, but the faunal exploitation patterns of III” do not differ substantially from variation found in the Aurignacian, with the exception of an increase in land snail exploitation. In fact the species representation is very similar to that of Aurignacian layer IIIe-g, with high frequencies of fallow deer and some evidence of hare exploitation, but few partridges and bustards. A major shift in prey representation occurred in association with the younger of the two industries (III’), when small fast-moving game overtakes fallow deer by NISP count for the first time. Great bustard and partridge frequencies increase and slightly higher rates of European wild ass are found, indicating more open vegetation. Tortoise disappears from the faunal assemblages after this layer, as does great bustard. Faunal and lithic densities also decrease at this time relative to the Aurignacian series and layer III”, possibly indicating a decrease in site use. Site occupation intensity drops further during the Epigravettian and Mesolithic periods at Klissoura Cave 1. The lithic assemblages from these periods are modest, and faunal densities are the lowest of the entire sequence. Sample sizes from the youngest layers are small, so much so that some of these assemblages had to be excluded from many of the 313 analyses. Land snails are uncommon in the Epigravettian layer, but peak in the Mesolithic. The ungulate component is more even with respect to species representation in the Epigravettian and Mesolithic layers, because of the decline of fallow deer relative to other taxa, but the faunas are dominated by hare. Relatively few hearths were excavated and burned bones are uncommon in these layers. There is a clear shift in foraging behaviors between the Middle Paleolithic and Aurignacian layers at Klissoura Cave 1 in terms of prey representation, mortality profiles, processing intensity, and burning frequencies on certain ungulate body parts (i.e., crania). Taken in conjunction with the abundance of lithic materials, the large ornament assemblage, dozens of clay-lined hearths and possible structure, it seems that Aurignacian layer IV marked the beginning of a period of exceptionally intensive site use at Klissoura Cave 1. Occupation intensity peaked in layer IV, but remained high throughout the other Aurignacian units and into layers III” and III’. The second major faunal shift occurred after layer III’, when tortoises and bustards completely disappear from the assemblage, and hare becomes the most common prey species at the site. Klissoura Cave 1 was either abandoned during the Last Glacial Maximum (LGM), or there is a depositional hiatus at this time. Use of the site during the Epigravettian and Mesolithic was considerably lighter based on all lines of artifactual and faunal evidence. However, several other caves in the areas were visited by groups during these periods (Koumouzelis, et al. 2004). Many of the important Greek sites that date to the Mesolithic are coastal (Jacobsen 1981; Jacobsen and Farrand 1987; Sampson 1998), and evidence for open water seafaring appeared during this period. Changes in Site Use and Occupation Intensity 314 Several different trends are apparent at Klissoura Cave 1 that relate to occupation intensity and the use of faunal resources. These trends are often at odds with one another and may reflect a combination of changes in the use of the site itself, as well as larger human population changes on the landscape of southern Greece. Figure 8.1 illustrates broad changes in ungulate diversity, the use of small, fast-moving game species, and fluctuations in occupation intensity through the sequence at Klissoura Cave 1. Diversity data are based on the reciprocal of Simpson’s Index values presented in Chapter 5, and small, fast-moving game are from NISP proportions. Changes in occupation intensity are based on faunal and lithic densities, as well as anthropogenic sedimentation and the relative abundance of hearth features in different layers. As mentioned in earlier chapters, “occupation intensity” or “increases in site use” can reflect 1) human groups using a site more often, 2) an increase in group size at the site, or 3) humans spending more time at an individual site. This study does not seek to distinguish between these behaviors; an increase in site use may represent one or a combination of several of these factors. Ungulate diversity roughly tracks environmental changes, as discussed in Chapter 3; diversity is generally higher in periods in which climate was wetter (i.e., Aurignacian layer IV). Diversity values drop when vegetation contracts in dry phases, implying that there are fewer suitable habitats for ungulate species. It might be expected that the site was used more often by humans during these phases, when diverse ungulate prey was abundant on the landscape, but this does not always seem to have been the case. Rather, the intensity of site use is low through the Middle Paleolithic and peaked during the Aurignacian and later Upper Paleolithic phases, followed by a decline after the LGM. It is likely that hominin populations were generally low in the Middle Paleolithic of Greece in general. Greece certainly would 315 Figure 8.1 Illustration of changes through time in ungulate diversity, occupation intensity and small game use at Klissoura Cave 1. have been an early area of colonization by modern humans as they moved across Europe, though the occupation pulse during the Aurignacian in Klissoura Cave 1 post-dates this phase. It must be noted that no increase in site use is associated with the Early Upper Paleolithic (Uluzzian) as compared to Middle Paleolithic layers. It is also possible that southern Greece was a refugium for many plant and animal populations during the mild but climatically variable MIS 3, making it an attractive region for human habitation, especially after 40,000 BP. However, no contemporary sites in Greece preserve a similarly intense Aurignacian occupation, though this may simply reflect the small number of Upper 316 Paleolithic sites that have been excavated in the region. The LGM in Greece was either a period of depopulation or an interval in which sediments did not accumulate (e.g., Farrand 2000; Karkanas 2010). After the deglaciation, Klissoura Cave 1 was occupied much less intensively during the Epigravettian and Mesolithic periods. The proportion of small, fast-moving animals in the Klissoura Cave 1 assemblages does not track the environmental changes or occupation intensity (Figure 8.1). Rather, there is occasional exploitation of hare in the Middle Paleolithic, although some of these animals may be intrusive. Small, fast-moving animals (low-ranked prey) become much more common in the Upper Paleolithic and eventually surpass large game by NISP counts in the Epigravettian and Mesolithic. It is significant that low-ranked small game was hunted more and more in later periods, even though occupation intensity had actually declined. This result likely reflects a regional trend (see also Stiner and Munro 2011), and could indicate that rising human populations stressed higher-ranked faunal resources (e.g., large game and small, slow-moving animals) on the landscape in general. Season of occupation is also relevant in a discussion of site use at Klissoura Cave 1. Fetal remains throughout the sequence are exceptionally small, indicating that the site was often utilized during the late winter or early spring in all culture periods. However, proportions of fetal remains are generally higher in the Middle Paleolithic, as is the representation of prime-aged adult animals. This pattern may reflect the movement of hominin populations as part of seasonal rounds and/or the targeted hunting of maternal herds. Upper Paleolithic and later groups may have occupied the site during a wider range of seasons, opportunistically hunting maternal and bachelor herds. 317 RESULTS FROM THE PERSPECTIVE OF EVOLUTIONARY ECOLOGY In this section, I return to predictions set forth in Chapter 1 about the faunal record at Klissoura Cave 1 based on models from evolutionary ecology. One of the most interesting and valuable aspects of the application of such models is when the archaeological record deviates from the predictions. Deviations may indicate a unique cultural response or a shortcoming of the model in its initial form. Prey Choice Following the prey choice model, it is expected that lower-ranked resources will enter the diet of foragers when high-ranked resources are less abundant in the environment (Emlen 1966; MacArthur and Pianka 1966; Pianka 2000; Stephens and Krebs 1986). A decrease of high-ranked resources can be caused by human hunting pressures, environmental change, or a combination of both. The inclusion of more lower-ranked resources in the diet may manifest as a generally wider diet breadth, which can result in greater evenness in the proportions of prey types in an archaeological assemblage. Another marker of expanding diet breadth is the greater inclusion of low-ranked resources (e.g., small, fast-moving animals) relative to higher-ranked ones (e.g., small, slow-moving game or ungulates). At Klissoura Cave 1, a trend in resource depression and widened diet breadth is apparent from multiple lines of evidence. There is a general increase in prey evenness based on the reciprocal of Simpson’s Index for all taxa, although the trend is not statistically significant. The proportion of prey types changes through the sequence, favoring small quick types and labor-intensive invertebrates with time, thus suggesting an increase in the costs of capture or processing in relation to food yields. There is a significant increase in the 318 proportions of small to large prey animals from the Middle Paleolithic through Mesolithic periods at Klissoura Cave 1, as well as an increase in the small game fraction of fast-moving types at the expense of slow-moving species. The decline of high-ranked prey types over time cannot be explained by the climatic fluctuations, as discussed in Chapter 3. Taken together, these lines of evidence indicate some degree of resource depletion and widened diet breadths at Klissoura Cave 1 from the Late Pleistocene through the Early Holocene. One possible deviation from the prey choice model is the occasional appearance of low-ranked hares in the Middle Paleolithic, particularly in layers XI-XIV, when there are no indications of resource stress. However, hare specimens are typically rare in the Middle Paleolithic assemblages overall (MNI=1 in most layers). Middle Paleolithic hominins were capable of exploiting hares, but this animal never comprised a significant part of the diet at Klissoura Cave 1. Layer XI-XIV has a higher incidence of hare remains than any other layer (MNI=2), but there is reason to believe that at least some of the remains are intrusive. Hare remains in layer XI-XIV have no evidence of cut marks and frequencies of burning are much lower than those for the assemblage as a whole. Early exploitation of small, fast-moving game has been noted elsewhere in the Mediterranean (e.g., Blasco and Fernández Peris 2009; Sanchis Sera and Fernández Peris 2008), and though hares were occasionally exploited in the Middle Paleolithic at Klissoura Cave 1, the major trend is still toward the increased reliance of lower-ranked resources in later periods. Central Place Foraging Central place foraging models attempt to explain the transport decisions that foragers make when bringing food items to a home base (Orians and Pearson 1979; Schoener 1979). It 319 is expected that, if human hunters exploit large game animals near a home base or habitation site, they will move more than just high-utility portions of carcasses. Rather, both high and low-utility elements are brought to the site because the cost of transporting the entire carcass (or large portions of it) is fairly low. The result is a more even anatomical representation at a site. Conversely, if foragers travel a great distance to hunt, they maximize their returns by only transporting the highest utility elements back to the habitation site. If local resource depletion occurs as the result of overhunting or changing environmental conditions, it is expected that there will be an increasing trend toward high-utility elements in an archaeological assemblage (Broughton 1999; Nagaoka 2002a; 2005, see also Faith 2007 for an example of changing transport decisions based on more favorable environmental conditions). It must be noted, however, that though food utility is based on anatomical characteristics (i.e., meat or marrow content of different elements), it is a relative concept that also depends on technologies available to foragers (e.g., bone grease rendering, which currently does not seem to predate the Gravettian, see Manne and Bicho 2009). Several different analytical strategies were employed to examine whether there were shifts in the relative utility of ungulate body parts transported to Klissoura Cave 1, or if body part transport was more complete in certain time periods. Based on other changes in faunal exploitation patterns indicative of resource stress (e.g., low-ranked prey exploitation in the youngest layers or increased marrow processing in Aurignacian layer IV), we might expect preferential transport of high-utility body parts during certain periods. Evaluating the average food utility index values for body parts transported to the site in each layer did not yield any trend that might indicate human hunters exploited large game further from the site in expected periods. In fact, all layers except for one in the Middle Paleolithic had a lower 320 average FUI value than is found in a complete cervid skeleton. This indicates that there was no preferential transport of high-utility body parts (based on FUI values) in most periods. This is also the case when correlations were tested between anatomical representation and MGUI/GUI values; no relationships indicate that there was an increase in the transport of high utility elements in later periods. Similarly, examining changes in ungulate body part representation by anatomical unit does not indicate directional changes in transport strategies through the sequence, though some layers had more uniform distributions than others based on K-S tests (e.g., Upper Paleolithic layers III’, IIIb-d, red deer and ibex in layer IV, and Middle Paleolithic layers VIII and XXa-XXb). There are several possible explanations for the lack of trend in body part transport decisions at Klissoura Cave 1. One is a taphonomic bias; density-mediated attrition is a potential problem, particularly in the Middle Paleolithic layers, which would cause lowdensity, high-utility elements such as vertebrae to be underrepresented in the assemblages. Another explanation concerns the technologies available to foragers occupying the site. There is no evidence of bone grease rendering, so it is possible that bulky vertebral elements were abandoned in the field where the carcass was butchered. Faunal elements at an archaeological site are used as proxies for meat transport, but in fact edible tissues can be removed from bones that are then left at kill sites. There may also be an issue with the assumed goals of foraging excursions, namely that FUI and GUI rely too heavily on the meat index, when in many cases food transport correlates with marrow-rich elements (see below). A final explanation for the overall lack of temporal trend in large game transport patterns is that human hunters at Klissoura Cave 1 simply did not travel greater or smaller distances to hunt large game across periods. Rather, they may have periodically abandoned the site when 321 large game became less abundant. There must be a point where the distance traveled for large game is so great and transport costs are so high that foragers can instead get a higher return from exploiting locally available small game (Cannon 2003). There is evidence for all of these possibilities at Klissoura Cave 1, and to an extent it may be a combination of factors. However, the position of the site on the landscape, next to a perennial stream that would have attracted large herbivores, probably allowed foragers to hunt prey close to the site during most periods. By the end of the sequence, it seems that human groups employed a different intensification strategy than traveling long distances to procure game. Low-ranked prey items played a greater role in forager diets during later periods of the site, and the shelter was certainly used less intensively. By the Mesolithic, it is also clear that other local sites were in use, including other shelters in the Klissoura Gorge, as well as coastal sites such as Franchthi Cave, where marine resources were included in the widening dietary repertoire. Results on Patch Choice based on Carcass Processing Intensity The patch choice model predicts which high-ranked patches on the landscape a forager will exploit (MacArthur and Pianka 1966), and the marginal value theorem determines how much time foragers will spend in a given patch based on the depletion of the patch as a function of time (Charnov 1976; Charnov, et al. 1976). One anthropological application of the patch choice model and the MVT is to butchery or processing intensity of prey carcasses. In this case resource intensification as reflected in increased bone marrow or grease processing over time can indicate human population pressure or resource stress (Burger, et al. 2005; Nagaoka 2005). Changes in processing intensity can be in response to 322 seasonal shortages, or if they occur over a period of time may reflect large-scale trends in carcass exploitation patterns provoked by food shortages. At Klissoura Cave 1, there is no indication that marrow processing intensified through the sequence, though there are certain periods where it was somewhat more important (i.e., Aurignacian layer IV). Ungulate elements with the least amount of marrow (e.g., phalanges, particularly the third phalanx), show no directional trend in the frequencies with which they were opened, while other elements were always processed for marrow. During the formation of Aurignacian layer IV, there is evidence for more intensive processing of terminal phalanges, as well as a preference for the transport of marrow-rich elements to the site. This layer also corresponds with an increase in site use, evidenced by higher quantities of lithics, ornaments and hearth structures. At the same time, ungulate diversity is high in layer IV, so it seems that human foragers exploited the rich environment very intensively during this period. There is no evidence for heat-in-liquid bone grease rendering at the site, which is considered to be another important measure of intensification (Bar-Oz and Munro 2005; Manne and Bicho 2009; Munro 2004; Munro and Bar-Oz 2005; Nakazawa, et al. 2009; Stiner 2003b). Bone grease rendering technologies were probably not part of the cultural repertoire when the site was occupied most intensively (e.g., the Aurignacian). As of this writing, there is no evidence for bone grease rendering in Eurasia earlier than the Gravettian (Manne and Bicho 2009). By the Epigravettian and Mesolithic periods, large game is so rare at Klissoura Cave 1 that potential changes in processing intensity are nearly irrelevant. Resource Depression: Climate Change, Human Population Growth or Both? 323 There is some evidence of resource depression in the vicinity of Klissoura Cave 1 beginning in the Upper Paleolithic and increasing in intensity with the Epigravettian and Mesolithic periods. Very little change in hunting or butchery practices occurred during the Middle Paleolithic at this site, with hominin groups almost exclusively targeting high-ranked resources. These resources included tortoises, which have low fertility rates and whose populations are easily impacted by increased harvesting pressures (Stiner, et al. 2000). By the Upper Paleolithic and later periods, lower-ranked resources comprised a growing part of the diet, and the frequency of high-ranked tortoises declined substantially in the Aurignacian. The rising proportion of prey species that inhabit open areas likely correlates with short-term fluctuations of vegetation types in the area, whereas other long-term trends do not relate to environmental conditions. It is interesting, for example, that great bustard was exploited at low-levels during the Aurignacian but completely disappear from diets by the Epigravettian, an interval in which open parkland communities were expanding (Albert 2010; Ntinou 2010). Human hunting pressures may have depressed populations of these slow-reproducing birds. Tortoises also completely disappear from the diet by the Epigravettian. Because of these changes, and the fact that ungulate diversity increased again in the Mesolithic while exploitation of large game species declined, it seems likely that human foragers were a major cause of resource depletion in the Late Pleistocene and early Holocene in the eastern Peloponnese. It is noteworthy that intensified foraging did not extend to increased processing intensity or longer travel distances for ungulate hunting. Though human population densities were on the rise in southern Greece, foragers could continue to move along if all local animal resources became depleted. It could also be that the gorge remained a prime vantage point for 324 hunters because of its position at the edge of the Argive Plain. In any case, the relentless increase in the use of quick small game with time cannot be explained by environmental changes alone. KLISSOURA CAVE 1 WITHIN GREECE AND THE MEDITERRANEAN BASIN At Klissoura Cave 1 there is evidence for resource depression during the Upper Paleolithic and later periods, particularly when it comes to prey choice. Transport decisions and processing intensity do not change with time, though there are fluctuations in some periods. How do the subsistence trends at Klissoura Cave 1 compare with those found at sites across the Mediterranean Basin? Here I compare general trends in faunal exploitation at Klissoura Cave 1 to contemporary sites that contain robust data sets from the northern and eastern Mediterranean. Sites are included based on their location in the Mediterranean zone defined by Blondel and Aronson (1999:8) (Figure 8.2). Figure 8.2 Map of the four quadrants of the Mediterranean. Shaded areas indicate Mediterranean ecosystem. From Blondel and Aronson (1999). 325 Faunal data from Late Pleistocene sites in Israel provide a high-resolution picture of dietary trends in the eastern Mediterranean. In their work on the Kebara Cave Middle Paleolithic faunal assemblages, Speth and Tchernov (1998; 2001; 2002, see also Davis 1977 for results from earlier excavations that support many points discussed here) conclude that Neandertal groups utilizing the site were successful hunters of large, prime-aged adult game animals during the late Pleistocene, including gazelle, fallow deer, red deer, wild pigs and aurochs. The most common species at the site are gazelle and fallow deer. The authors find that the hunting of male and female fallow deer and gazelles generally reflects their proportions in living populations; temporal shifts in their proportions relate to changes in the season of site occupation (Speth and Tchernov 2001). Neandertals typically transported highutility body parts to the site after low-utility bulky elements were removed and abandoned in the field. In terms of small game use, Speth and Tchernov (2002) find evidence from burning patterns that, in addition to their value as a food source, tortoise carapaces may also have been used as containers. They observe tortoise size diminution at the end of the Middle Paleolithic, which may indicate mild prey population suppression (e.g., Stiner, et al. 2000), though the authors note that the early Upper Paleolithic occupations of Kebara Cave were ephemeral and that other explanations must be explored (Speth and Tchernov 2002). Expanding on these studies, Speth and Clark (2006) note that frequencies of large ungulates such as red deer and aurochs declined steadily over time until they comprise a fraction of the early Upper Paleolithic assemblage. Also, foragers at the site exploited more young fallow deer and gazelles in later periods (Speth and Clark 2006). Based on an extensive study of the Hayonim Cave faunal series, and in conjunction with other pan-Mediterranean work, Stiner (2005) notes three major trends in human 326 subsistence strategies over the course of the Middle and Late Pleistocene through Early Holocene. First is the exploitation of prime-aged adult ungulates that began before the Middle Paleolithic (Stiner 2005; Stiner, et al. 2009). Next is a general trend in small game exploitation. Middle Paleolithic hominin groups spent little time pursuing small game species, except for those with the highest return rates such as tortoises or marine mollusks. Average tortoise body size at Hayonim and Kebara Caves began to decline after about 44,000 BP, as a result of overharvesting larger, more productive individuals, a trend that continued through the Natufian (Stiner 2001, 2005; Stiner, et al. 1999; Stiner, et al. 2000). Low-return small, quick species such as birds become common in the Upper Paleolithic, and by the Epipaleolithic lagomorphs increased in importance (Stiner 2001, 2005; Stiner, et al. 1999; Stiner, et al. 2000). The final trend is the appearance of more efficient hunting technologies including stone and bone-tipped projectiles in the late Middle and early Upper Paleolithic, possibly including snare, net and trap technology late in the Upper Paleolithic (Stiner 2005). Labor-intensive carcass processing technologies, such as bone grease rendering, also appear in later periods (Bar-Oz and Munro 2005; Munro 2004; Munro and Bar-Oz 2005). By the Epipaleolithic, several lines of evidence indicate that resource intensification was widespread in the Levant. At multiple sites, evidence for depression of medium ungulate (i.e., fallow deer) populations began early in the Epipaleolithic and peaked in the Natufian, at time when small ungulates (i.e., mountain gazelles) overwhelmingly constitute the ungulate component of the archaeofaunas (Munro 2004, 2009a, 2009b; Stutz, et al. 2009). Though ungulates remained the dominant prey type in the Levantine sites discussed, the incidence of small quick game also increased starting in the Early Natufian (Bar-Oz 2004; Munro 2004, 327 2009a, 2009b). In addition to changes in the prey species exploited, young gazelles were commonly hunted during the Early Natufian (Munro 2004, 2009a, 2009b). Bone marrow was intensively processed across the Epipaleolithic and Natufian periods at the Israeli sites (BarOz and Munro 2007), and there is strong evidence for bone grease rendering at this time (Munro 2004, 2009a; Munro and Bar-Oz 2005). In the Mediterranean zone of Turkey, recent work at the Paleolithic sites of Uçağızlı Caves I and II (Kuhn, et al. 2009; Stiner 2009) and Epipaleolithic Öküzini and Karain Caves (Atici 2009) provides large faunal data sets from the Late Pleistocene. Uçağızlı Caves I and II contain complementary Middle Paleolithic (II), Upper Paleolithic (I) and Epipaleolithic (I) sequences. The most common prey types in the assemblages are fallow deer and bezoar goat, whose abundances through the sequence oppose one another, depending on climate regime, followed by roe deer and a range of other ungulate and carnivore species in smaller numbers. There is a decline in the evenness of ungulate prey through the sequence that follows the disappearance of large ungulates such as aurochs and red deer (Kuhn, et al. 2009; Stiner 2009). A striking trend is apparent in the small game component for these sites. Small, slowmoving species such as tortoises and shellfish are present throughout the sequence, though tortoise abundance drops in the Upper Paleolithic. Conversely, small fast-moving lagomorphs, birds and fish only become important in the Upper Paleolithic and Epipaleolithic layers (Kuhn, et al. 2009; Stiner 2009). Little variation is noted in body part transport decisions through the sequence, with axial elements and phalanges consistently underrepresented. A change in butchery practices occurred during the early Upper Paleolithic; in earlier periods most cut marks were diagonal to the main axis of elements, 328 whereas in the Ahmarian most cuts have an axial orientation, possibly reflecting the removal of large, intact portions of meat, possibly for smoking (Stiner 2009). Epipaleolithic faunal data from southwest Turkey at the sites of Öküzini Cave and Karain Cave are presented by Atici (2009). The data reveal a nearly exclusive focus on wild sheep and goat for most of the Epipaleolithic, which Atici (2009) argues to be evidence for hunting specialization. There is a general diversification of the diet toward the end of the Epipaleolithic in the area, with the inclusion of low and high-ranked game such as roe deer, wild boar, tortoises, hares and partridges into the meat diet. The author notes that this dietary expansion occurred during a time of climatic amelioration and that it may be the result of increased availability of a wider range of species (Atici 2009). In Greece, excluding Klissoura Cave 1, the most substantial faunal data sets available are from Franchthi Cave (Payne 1975; Rose 1995; Stiner and Munro 2011), Kastritsa (Kotjabopoulou 2001), Klithi (Gamble 1997; 1999), Boïla (Kotjobopoulou 2001) and Cave of Cyclope (Mylona 2003; Powell 2003; Trantalidou 2003). Of these, Franchthi Cave is the only site that preserves a long sequence of faunal remains from the Aurignacian through Mesolithic (and a substantial Neolithic component that is not the focus here). The Franchthi Cave faunas provide data on long-term subsistence trends. Ungulates such as red deer and equids dominate the Upper Paleolithic layers of Franchthi Cave, and shifts in their frequencies track environmental change (Payne 1975; Stiner and Munro 2011). By the Epigravettian, ungulate diversity declines and the meat diet widens to include small game animals such as land snails, shellfish, coastal fish species, and pond turtles (as opposed to tortoises, which are found earlier) (Stiner and Munro 2011). Across the Mesolithic, more open or deep water fish such as barracuda and tuna were exploited (Rose 1995; Shackleton 329 1990; Stiner and Munro 2011). The subsistence picture from Franchthi Cave is that of expanding diets, incorporating more low-ranked small game resources with time. The Cave of Cyclope also has abundant Early and Late Mesolithic remains (Mylona 2003; Powell 2003; Trantalidou 2003). Ungulates, specifically goat and sheep, are common throughout the sequence. A small number of pig remains are found in the later layers, and land snails occur in the earlier layers (Trantalidou 2003). Abundant fish remains are found throughout the Mesolithic but they increase in frequency in the Late Mesolithic (Mylona 2003; Powell 2003). As an island site marine resources are expected, but the Cave of Cyclope confirms greater use of low-return small game, such as fish, in Mesolithic diets. In Italy Middle Paleolithic hominins were successful hunters of large, prime-aged adult ungulates in addition to opportunistic scavengers (Stiner 1994). The transport of ungulate body parts is variable in the Italian Mousterian and seems to relate to specific food acquisition tactics (e.g., hunting or scavenging) at different sites. Small game animals do appear in the Middle Paleolithic assemblages, but they are generally restricted to easy to collect prey such as tortoises and marine mollusks (Stiner 1994). In northern Italy, Middle Paleolithic faunal assemblages are likewise dominated by prime-aged adult ungulate faunas, though carnivores and small game such as lagomorphs and birds are also observed in low numbers (Fiore, et al. 2004). The authors note that many of the bird and lagomorph remains lack evidence of human modification, so it is unclear if they were introduced to the sites by carnivores or other natural processes. A review by Alhaique et al. (2004) that spans the Lower Paleolithic through Neolithic at sites around the country indicates that ungulates were important throughout the Late Pleistocene and early Holocene, and that the frequencies of small game increased in later periods. 330 Ungulate exploitation continued in the Upper Paleolithic and Epipaleolithic at Italian sites, as did small game collection, though the complexion of the small game component changed (Stiner 2001; Stiner, et al. 1999; Stiner, et al. 2000). By the early Aurignacian, avian faunas appear in the assemblages in considerable numbers, and by the Epipaleolithic lagomorphs were heavily exploited. The inclusion of these small, fast-moving game species was at the expense of slow-moving types, presumably because the latter were increasingly rare in the environment. Declines in tortoise and shellfish in human diets may have been the result of overharvesting; limpet diminution is apparent by 23,000 BP at Italian sites (Stiner 2001; Stiner, et al. 1999; Stiner, et al. 2000). However, shellfish gathering as a major food source persisted at some Italian sites (Colonese, et al. 2011). In an analysis of late Upper Paleolithic sites across Italy, Phoca-Cosmetatou (2003b) does not find an across-the-board widening of diet breadth, but rather increasingly variable foraging patterns, one of which is specialized ibex hunting. In northeast Italy, Phoca-Cosmetatou (2009) stresses the high degree of variability in late Upper Paleolithic subsistence strategies and notes no specific temporal trends in prey exploitation patterns if site use is not taken into account. The French Paleolithic has been intensively studied and many high-quality data sets are available. I limit my discussion here to those sites in the southern part of the country, in the classic Mediterranean environmental regime (Figure 8.2). During the Middle Paleolithic in southern France, hominins were highly mobile hunters that exclusively targeted large ungulate prey (Boyle 2000; Valensi 2000). Boyle (2000) notes that regional variation in the prey species exploited relates to geographic factors (i.e. montane regions typically associated with ibex hunting). Species dominated by a single prey type are common, but may not reflect specialized hunting reported elsewhere for the Upper Paleolithic. Boyle (2000) does not 331 exclude this possibility, but notes that seasonality data are currently lacking from many Middle Paleolithic sites, so it is unclear if such monospecific faunas simply reflect season of use. The author finds little evidence for scavenging in faunal assemblages from the region based on the proportions of body parts found at Mousterian sites (Boyle 2000). At Lazaret Cave, Valensi (2000) notes that Middle Paleolithic hominins hunted prime-aged adult red deer and ibex, transporting the carcasses in whole form to the site for butchery. Ungulate exploitation persisted in the Upper Paleolithic of Mediterranean France, and species targeted largely reflect their environmental availability (Costamagno 2003; Surmely, et al. 2003). The use of small game seems to correspond with the onset of the Upper Paleolithic (Costamagno and Laroulandie 2004). Small animals occur in trace amounts at a few Middle Paleolithic sites, though they may have been accumulated by carnivores. Small game is likewise rare in the Châtelperronian but increases in frequency in later periods. By the Magdalenian, the exploitation of birds, lagomorphs and fish is common (Costamagno and Laroulandie 2004). Forager diets widen even further in the Mesolithic to include shellfish, in addition to other small game resources. Several faunal trends are apparent in Late Pleistocene sites of Mediterranean Spain. Tortosa et al. (2002) provide an extensive review of Paleolithic faunal data from southern Spain. They find that Middle Paleolithic hominins exploited a diverse range of ungulate faunas, typically prime-aged adults or larger subadults (Tortosa, et al. 2002). Faunal richness at Middle Paleolithic sites is often deceptively high because many human occupations, particularly at rock shelters and caves, are interspersed with carnivore accumulations. Small game is rare in the Middle Paleolithic, but recent evidence from Bolomor Cave indicates that hominins at the site exploited tortoise, rabbit and duck as early as MIS 6 (Blasco 2008; 332 Blasco and Fernández Peris 2009; Sanchis Sera and Fernández Peris 2008). Species diversity drops between the Middle and Upper Paleolithic as the result of contracting territories of some herbivore species and the extinction of many large carnivores (Tortosa, et al. 2002). Local fluctuations of ungulate taxa is probably in response to shifting environmental conditions. In the Upper Paleolithic, a range of small game animals appear in forager diets, especially lagomorphs, but also including birds, fish and shellfish (Hockett and Haws 2002; Tortosa, et al. 2002). During the Gravettian and Solutrian periods specialized deer, ibex and horse hunting appears at certain sites. Adult and subadult ibex and red deer were targeted at special-function sites in the Magdalenian and Epipaleolithic. In the later Upper Paleolithic and Epipaleolithic, marine resources including mammals and shellfish appear in greater numbers (Tortosa, et al. 2002). It has been argued that many of the dietary differences between Middle and Upper Paleolithic hominins in Spain stem from different land use strategies and increased territoriality (Tortosa, et al. 2002; Villaverde, et al. 1996). During the Middle Paleolithic, Neandertal groups had access to a greater range of large game resources, but by the Upper Paleolithic, human foragers shifted between deer and ibex (i.e., coastal and mountainous areas) on a seasonal basis. The authors conclude that rabbit exploitation is tied to this change in land use and stress that it is not the result of human demographic pressure or resource intensification (Tortosa, et al. 2002; Villaverde, et al. 1996). Whether or not this interpretation is correct, the overarching trend fits with what is found in other parts of the Mediterranean. 333 Also on the Iberian Peninsula is the Rock of Gibraltar, which has preserved a concentration of large Paleolithic cave sites. Faunal data are available from the late Mousterian layers from Vanguard Cave and the late Middle Paleolithic and post-LGM Upper Paleolithic layers from Gorham’s Cave (Stringer, et al. 2008). The Vanguard Cave Middle Paleolithic is dominated by marine mollusks and ungulates, specifically ibex and red deer, though other ungulates are also found in addition to small amounts of carnivore, seal, dolphin, bird, tortoise and fish (Stringer, et al. 2008). At Gorham’s Cave, lagomorphs are the most common prey in both the Middle and Upper Paleolithic layers, followed by birds in the Middle Paleolithic and cervids in the Upper Paleolithic, though again, other herbivores and carnivores are also present in small numbers. Stringer et al. (2008) notes little change in the faunal assemblages between the Middle and Upper Paleolithic layers, either in represented species, prey diversity, or butchery patterns. The Late Pleistocene faunal record of Portugal is less well-known than adjacent areas, but recent work has greatly expanded our understanding of the western-most expanse of the Mediterranean. One of the most striking features of the Portuguese record is the importance of lagomorphs in the diet, at least during the Upper Paleolithic (Hockett and Haws 2002). Middle Paleolithic faunal remains are fairly uncommon, but seem to focus on the procurement of large game species (see Hockett and Haws 2009). Robust data sets are available for some Upper Paleolithic sites, however. Calderão Cave in north-central Portugal preserves Mousterian through Neolithic remains, though the majority of the faunas associated with the Middle Paleolithic and Early Upper Paleolithic layers were probably accumulated by hyenas (Davis 2002). In the later Upper Paleolithic, a range of ungulates and carnivores occur in low frequencies, but the 334 proportion of rabbits is much greater (>95% in the Solutrean and Magdalenian layers) (Davis 2002). At Vale Boi in Algarve, there is evidence of the intensive use of faunal resources across the Upper Paleolithic. Limpets are abundant during the Gravettian but subsequently decline during the Solutrian as the coastal geography changed (Manne and Bicho 2009). Rabbits dominate the archaeofaunas in all Upper Paleolithic phases at the site, followed by red deer and horse, which fluctuate in frequency probably as a result of climate shifts. Significantly, there is extensive evidence for bone grease rendering, an extremely laborintensive activity, as early as the Gravettian (Manne and Bicho 2009; Manne, et al. 2005; Stiner 2003b). Though no trends are evident in small game exploitation patterns or grease rendering, it seems clear that by the Upper Paleolithic, human populations in southern Portugal experienced some form of continuous resource stress. Rabbits were likewise the main food source during the late Upper Paleolithic at Picareiro Cave (Bicho, et al. 2006; Hockett and Bicho 2000). A range of ungulate species are found at the site in low numbers, red deer and wild boar in particular, as well as small carnivores, fish and shellfish. During the late Magdalenian, Bicho et al. (2006) argue that the site was used as a hunting and carcass processing camp for small and large game, though rabbit is by far the dominant species. The authors also suggest that there is evidence of intensive marrow processing and bone grease rendering of large game animals at the site (Bicho, et al. 2006). In Portugal, it seems that Upper Paleolithic subsistence strategies differed from other areas of the Mediterranean, with the near-exclusive exploitation of lowranked prey and intensified processing of large game animals. On the other hand, quantitative 335 faunal data from the Middle Paleolithic and early Upper Paleolithic, which may indicate a more widespread use of large game, are largely lacking at this point. The main features of Late Pleistocene Mediterranean subsistence strategies are their variability through space and time in response to local prey availability. Some broader trends are apparent, however. Ungulate prey was preferred in all periods; evidence for the heavy exploitation of small game typically corresponds with declines in large game hunting. Small, sessile animals such as tortoises and marine mollusks are also exploited when they are available in the environment. The timing of the inclusion of low-ranked prey such as small, fast-moving hares, birds and fish varies regionally. In Iberia, for example, lagomorphs appear in diets quiet early, possibly in the Middle Paleolithic but definitely with the Upper Paleolithic. There are also some situations, such as when rabbit warrens are available, where collecting young lagomorphs is argued to be less labor-intensive and does not require specific technologies (Jones 2006). But the trend also involves hares in virtually every Mediterranean region, and this species does not live underground. At the end of the Pleistocene, diets included marine, terrestrial and avian faunas to a greater extent. In the most general terms, then, there is a picture of intensification throughout the Mediterranean Basin from the Late Pleistocene through the early Holocene, though understanding the dynamic relations between local environmental factors and prey availability is critically important and not always obvious. Intensification and Specialization There is the question of whether specialization and/or intensification occurred at Klissoura Cave 1. These issues are often framed in terms of differences between the Middle 336 and Upper Paleolithic. Clearly, as discussed above, diet breadth widened across the occupation sequence of the site and more low-ranked resources were included in the diet with time, though mostly later in the Upper Paleolithic. In this sense, resource depression would seem to have led to the intensified use of different species, particularly those taxa with high fertility rates. On the other hand, there is no evidence of increased carcass processing in later periods that would also be indicative of intensification. Perhaps resource depression evidenced by changing prey types at the site was a local phenomenon, but comparisons to other sites in southern Greece, including Franchthi Cave, would suggest otherwise. It nonetheless is noteworthy that the evidence for foraging intensification does not include increased marrow processing or bone grease rendering. The dominance of fallow deer, particularly in the Middle Paleolithic layers of Klissoura Cave 1 is also worth discussing. Some archaeologists have indicated that specialized hunting was a hallmark of the Upper Paleolithic, at least in western Europe (Costamagno 2003, 2004; Mellars 1973, 1989; 1991, but see Grayson and Delpech 2002, 2006). Indeed, specialized hunting as defined by monospecific assemblages is observed at many Upper Paleolithic sites, including those with reindeer (David and Enloe 1993; Mellars 1989, 1991, 2004) and craggy-terrain-loving ibex and chamois (Gamble 1997; PhocaCosmetatou 2003a, 2003b, 2004a; Straus 1977, 1987). However, faunal assemblages dominated by a single species are also found in the Middle Paleolithic (e.g., Adler, et al. 2006; Costamagno, et al. 2006; Gaudzinski 1995, 2006). The fundamental issue concerning the question of specialization is what the existence of monospecific assemblages actually means. 337 Monospecific or specialized assemblages contain a high proportion of a single taxon, often as much as 90-95% of a single prey type (Mellars 1973, 1989, 2004), sometimes with MNI values that approach nearly 100 animals (Gaudzinski 2006), though this is not a requirement. From a taphonomic perspective, even if humans were responsible for accumulating these faunas, several different hunting strategies can cause monospecific assemblages, often as a palimpsest in the archaeological record. Gaudzinski (2006) notes a few examples of different hunting strategies from the Middle Paleolithic: bovid-dominated assemblages from the long-term exploitation of gregarious animals at Mauran, La Borde and Wallertheim (Farizy, et al. 1994; Gaudzinski 1995; Jaubert, et al. 1990), the selection of solitary young rhinoceros over a long period of time at Taubach (Bratlund 1999, though rhinoceros only makes up 27% of this assemblage by NISP and 49% by MNI), and mass hunting scenarios, such as reindeer at Salzgitter Lebenstedt (Gaudzinski and Roebroeks 2000) or Les Pradelles (Costamagno, et al. 2006). Clearly, there are many different situations that result in monospecific assemblages. The behavioral implications of specialization must be considered, and historically this has been the source of many interpretive biases. Adler et al. (2006:90) rightly point out that similarly diverse faunal assemblages are described as reflecting “opportunistic” hunting or “increased diet breadth,” depending on whether they were accumulated by Neandertals or modern humans, respectively, and Neandertal sites dominated by a single species have “low diet breadth,” whereas a modern human monospecific assemblage indicates “specialization.” For some authors, high proportions of a single species at a site may indicate intentional selection of that species from all available prey (Mellars 1996), while others consider the communal aspect of mass kills and processing (Binford 1968; David and Enloe 1993). Stiner 338 (1992:447) points out that specialization may simply reflect short-term adjustments to seasonal or local food availability. At Klissoura Cave 1, fallow deer dominate the faunal assemblages, comprising 6091% of the specimens identifiable to taxon in the Middle Paleolithic, and 12-78% in the Upper Paleolithic and later layers. To be sure, there is considerable variation between the assemblages within the Middle and Upper Paleolithic, but the larger point is that there is a significant decline in fallow deer proportions over time. Fallow deer MNI values range between one and fourteen individuals, depending on the layer. The proportions of fallow deer are not over 90% except in the very earliest Middle Paleolithic layer. Based on the age structure of individuals in the assemblages, this assemblage probably does not represent a mass kill. The prevalence of fallow deer in most of the assemblages likely reflects a prime fallow deer habitat in the Klissoura Gorge, particularly at the times of year when the site was occupied the most frequently. Interestingly, fallow deer are not found to the southeast at Franchthi Cave, though red deer are common, so perhaps the Klissoura Gorge remained slightly drier and grassy during most periods. CONCLUSION The application of models from evolutionary ecology to understand the foraging behaviors of hominins at Klissoura Cave 1 yields important conclusions about prey choice, transport decisions, and butchering intensity and how these behaviors changed through the sequence. Deviations from the modeled expectations are likewise interesting, such as the lack of evidence for preferential transport of high-utility elements during periods when other lines 339 of evidence (e.g., hunting low ranked prey or increased marrow processing) indicate intensification was occurring at the site. Overall, trends in faunal exploitation at Klissoura Cave 1 fit fairly well with those found at other Late Pleistocene sites in the Mediterranean, but there are also differences brought on by factors of the local environment and unique cultural adaptations in Greece. The presence of a long, well-preserved sequence with Middle Paleolithic, Uluzzian, abundant Aurignacian and overlying Upper Paleolithic and Mesolithic deposits is unprecedented for southern Greece. One of the more striking features of the long faunal sequence at Klissoura Cave 1 is the relatively static nature of the Middle Paleolithic record, as compared to the highly variable Upper Paleolithic hunting, transport, and carcass butchery strategies. Trends in the latter largely reflect changes in site use and occupation intensity. Overarching trends through the sequence concern mainly the increased importance of small game, particularly fast-moving animals, with time, which begins as early as the Aurignacian. Ungulates dominate the assemblages until the Epigravettian and Mesolithic layers, when small game overtakes large taxa by NISP counts. This tendency generally parallels the situation in many parts of the Mediterranean, including other sites in Greece, where Mesolithic foragers turned their attention to low-ranked resources such as birds, hares, and especially fish. Human groups at Klissoura Cave1 did not uniformly contend with resource shortages by intensifying carcass processing, though this did occur elsewhere in the Mediterranean Basin (e.g., Portugal and the Levant), but the occupants of Klissoura Cave 1 did intensify marrow processing efforts in the Aurignacian (layer IV). It is possible that intensification strategies such as bone grease rendering never entered the repertoire of foragers in southern Greece, but 340 it should also be noted that grease rendering appeared elsewhere in Europe after the Aurignacian and therefore after the heaviest use of Klissoura Cave 1. Klissoura Cave 1 preserves a long sequence, and its location provides the opportunity to address two of the major geographical gradients in Mediterranean Europe. First, are eastwest trends in species availability and the expansion and establishment of modern human populations, and subsequent cultural diversification during the early Upper Paleolithic. Second are the ecological conditions or refugium status of one of three major European peninsulas (Iberia, Italy, and southern Greece) during the climatically variable Pleistocene. Faunal evidence indicates that Klissoura Cave 1 was used mostly in the winter, particularly during the Middle Paleolithic, possibly following ungulates into lowland pastures with milder temperatures. Assuming modern humans were the manufacturers of Aurignacian stone tool industries, there is no evidence to suggest that Neandertal populations persisted later than 40,000 BP in southern Greece, as opposed to the situation in Iberia. Rather, there is a spike in human activity at Klissoura Cave 1 not long after modern human populations are expected to have appeared in southern Greece. Site use dropped significantly after the early Upper Paleolithic. The faunal trends at Klissoura Cave 1 nonetheless suggest increases in human populations locally and regionally, and the taxing of high-value faunal resources. This caused foragers in southern Greece to turn to lower-ranked terrestrial and marine resources by the end of the Upper Paleolithic and Mesolithic, before the eventual adoption of agricultural lifeways with the onset of the Neolithic. 341 APPENDICES Appendix A: Hare survivorship and bone density values (values from Pavao and Stahl 1999). 342 Appendix B: Fallow deer survivorship and bone density values (values from Lyman 1994). 343 Appendix B - continued 344 Appendix B - continued 345 Appendix B - continued 346 Appendix C: Tooth and bone-based MNE values by layer. 347 Appendix D: Shaft and end-based MNE values by layer. 348 Appendix E: Middle Paleolithic hare elements. Horizontal line indicates lowest structural density for cervids. 349 Appendix F: NISP and MNE values by layer for all taxon. 350 Appendix F - continued 351 Appendix G: Biomass values for common taxa. 352 Appendix G - continued 353 Appendix G - continued 354 Appendix H: Utility indices for Rangifer, combined by anatomical region (values from Binford 1978). 355 Appendix I: Percent MAU for fallow deer by layer. 356 Appendix I - continued 357 Appendix I - continued 358 Appendix J: Mean FUI for fallow deer by layer. 359 Appendix J - continued 360 Appendix K: Fallow deer anatomical representation. 361 Appendix K - continued 362 Appendix L: Hare anatomical representation. 363 Appendix M: Partridge anatomical representation. 364 Appendix N: Anatomical representation for red deer and ibex in layer IV. 365 Appendix O: Tooth wear stages for ungulates by layer. 366 Appendix O - continued 367 Appendix P: Ungulate fusion stages (values from Reitz and Wing 2008 and Silver 1969). 368 Appendix P - continued 369 Appendix P - continued 370 WORKS CITED Abramsky, Z. and M. L. Rosenzweig 1984 Tilman's predicted productivity-diversity relationship shown by desert rodents. Nature 309:150-151. Adam, E. 1989 A Technological and Typological Analysis of Upper Palaeolithic Stone Industries of Epirus, Northwestern Greece. BAR International Series 512, Oxford. 1997 Upper Palaeolithic Stone Assemblage Structure in Epirus. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, edited by G. N. Bailey, pp. 481-496. vol. 2. McDonald Institute for Archaeological Research, Cambridge. 1999a Preliminary Presentation of the Upper Palaeolithic and Mesolithic Stone Industries of Theopetra Cave, Western Thessaly. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perles and K. Zachos, pp. 266270, T. B. S. a. Athens, general editor. Technical Print Services Ltd, Nottingham, Great Britain. 1999b The Upper Paleolithic Stone Industries of Epirus. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perles and K. Zachos, pp. 137-147, T. B. S. a. Athens, general editor. Technical Print Services Ltd, Nottingham, Great Britain. 2007 Looking out for the Gravettian in Greece. Paleo 19:145-158. Adam, E. and E. Kotjabopoulou 1997 The organic artifacts from Klithi. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 1: Excavation and intra-site analysis at Klithi, edited by G. N. Bailey, pp. 245-259. McDonald Institute of Archaeological Research, Cambridge. Adler, D. S. and G. Bar-Oz 2009 Seasonal Patterns of Prey Acquisition and Inter-group Competition During the Middle and Upper Palaeolithic of the Southern Caucasus. In The Evolution of Hominin Diets: Integrating Approaches to the Study of Palaeolithic Subsistence, edited by J.-J. Hublin and M. P. Richards, pp. 127-140. Springer Science + Business Media B.V. Adler, D. S., G. Bar-Oz, A. Belfer-Cohen and O. Bar-Yosef 371 2006 Ahead of the Game: Middle and Upper Palaeolithic Hunting Behaviors in the Southern Caucasus. Current Anthropology 47(1):89-118. Adouze, F. 1987 The Paris Basin in Magdalenian times. In The Pleistocene Old World: Regional Perspectives, edited by O. Soffer, pp. 183-200. Plenum, New York. Adouze, F. and J. G. Enloe 1991 Subsistence strategies and economy in the Magdalenian of the Paris Basin, France. In The Late Glacial in North-West Europe: Human Adaptation and Environmental Change at the End of the Pleistocene, edited by R. N. E. Barton, A. J. Roberts and D. A. Roe, pp. 63-71. vol. 77. Council for British Archaeology Research Report, London. Ahern, J. C. M., I. Karavanic, M. Paunovic, I. Jankovic and F. Smith 2004 New Discoveries and Interpretations of Hominid Fossils and Artifacts from Vindija Cave, Croatia. Journal of Human Evolution 46(1):27-67. Albert, R. M. 2010 Hearths and Plant uses during the Upper Palaeolithic period at Klissoura Cave 1 (Greece): The results from phytolith analyses. Eurasian Prehistory 7(2):71-85. Albert, R. M. and S. Weiner 2001 Study of phytoliths in prehistoric ash layers using a quantitative approach. In Phytoliths, Applications in Earth Sciences and Human History, edited by J. D. Meunier and F. Coline, pp. 251-266. A. A. Balkema Publishers. Alhaique, F., M. Bisconti, E. Castiglioni, C. Cilli, L. Fasani, G. Giacobini, R. Grifoni, A. Guerreschi, A. Iacopini, G. Malerbata, C. Peretto, A. Recchi, A. R. Ris, A. Ronchitelli, M. Rottoli, U. T. Hohenstein, C. Tozzi, P. Visentini and B. Wilkens 2004 Animal Resources and Subsistence Strategies. Coll. Anthropol. 28(1):23-40. Allen, J. R. M., U. Brandt, A. Barauer, H.-W. Hubbertens, B. Huntley, J. Keller, M. Kraml, A. Mackensen, J. Mingram, J. F. W. Negendank, N. R. Nowaczyk, H. Oberhänsli, W. A. Watts, S. Wulf and B. Zolitschka 1999 Rapid environmental changes in southern Europe during the last glacial period. Nature 400:740-743. Allen, J. R. M. and B. Huntley 2000 Weichselian palynological records from southern Europe: correlation and chronology. Quaternary International 73/74:111-125. Allen, J. R. M., W. A. Watts and B. Huntley 372 2000 Weichselian palynostratigraphy, palaeovegetation and palaeoenvironment; the record from Lago Grande di Monticchio, southern Italy. Quaternary International 73/74:91-110. Ammerman, A. J., N. Efstratiou and E. Adam 1999 First Evidence for the Palaeolithic in Aegean Thrace. In The Palaeolithic Archaeology of Greece and Adjacent Areas. Proceedings of the ICOPAG Conference, Ioannina, September 1994, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 211-214. British School at Athens Studies 3. British School at Athens, London. Andrews, P. 1990 Owls, caves and fossils. University of Chicago Press, Chicago. Arnold, E. N. and J. A. Burton 1992 A Field Guide to the Reptiles and Amphibians of Britain and Europe. Collins, London. Atici, L. 2009 Specialization & diversification: animal exploitation strategies in the terminal Pleistocene, Mediterranean Turkey. Before Farming 3(1):1-17. Bailey, G. N. 1997 Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, 2 Volume Set. McDonald Institute for Archaeological Research. 1999 The Palaeolithic archaeology and palaeogeography of Epirus with particular reference to the investigations of the Klithi rockshelter. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, September 1994, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 159-169. British School at Athens Studies 3, London. Bailey, G. N., P. Carter, C. Gamble and H. P. Higgs 1983 Asprochaliko and Kastritsa: Further Investigations of Palaeolithic Settlement and Economy in Epirus (North-West Greece). Proceedings of the Prehistoric Society 49:15-42. Bailey, G. N., V. Papaconstantinou and D. Sturdy 1992 Asprochaliko and Kokkinopilos: TL Dating and Reinterpretation of Middle Palaeolithic Sites in Epirus, North-West Greece. Cambridge Archaeological Journal 2:136-144. Bailey, G. N. and J. C. Woodward 373 1997 The Klithi Deposits: Sedimentology, Stratigraphy and Chronology. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 1: Excavation and intra-site analysis at Klithi, edited by G. N. Bailey, pp. 61-94. McDonald Institute for Archaeological Research, Cambridge. Bar-Matthews, M., A. Ayalon and A. Kaufman 1997 Late Quaternary Paleoclimate in the Eastern Mediterranean Region from Stable Isotope Analysis of Speleothems at Soreq Cave, Israel. Quaternary Research 47:155-168. Bar-Matthews, M., A. Ayalon, A. Kaufman and G. J. Wasserburg 1999 The Eastern Mediterranean Paleoclimate as a Reflection of Regional Events: Soreq Cave, Israel. Earth and Planetary Science Letters 166:85-95. Bar-Oz, G. 2004 Epipalaeolithic Subsistence Strategies in the Levant: A Zooarchaeological Perspective. The American School of Prehistoric Research (ASPR) Monograph Series, Brill Academic Publishers, Inc., Boston, MA. 2005 Epipaleolithic Subsistence Strategies in the Levant: A Zooarchaeological Perspective. Brill Academic Publishers, Inc., Boston. Bar-Oz, G. and N. Munro 2005 Gazelle bone fat processing in the Levantine Epipalaeolithic. Journal of Archaeological Science 32:223-239. 2007 Gazelle bone marrow yields and Epipaleolithic carcass exploitation strategies in the southern Levant. Journal of Archaeological Science 34:946-956. Bar-Yosef Mayer, D., B. Vandermeersch and O. Bar-Yosef 2009 Shells and ochre in Middle Paleolithic Qafzeh Cave, Israel: indications for modern behavior. Journal of Human Evolution 56:307-314. Bar-Yosef, O. 2000 The Middle and Early Upper Paleolithic in Southwest Asia and Neighboring Regions. In The Geography of Neandertals and Modern Humans in Europe and the Greater Mediterranean, edited by O. Bar-Yosef and D. Pilbeam, pp. 107-156. Peabody Museum Bulletin 8. Peabody Museum of Archaeology and Ethnology, Harvard University. 2007 The Dispersal of Modern Humans in Eurasia. In Rethinking the Human Revolution: New behavioural and biological perspectives on the origin and dispersal of modern humans, edited by P. Mellars, K. V. Boyle, O. Bar-Yosef and C. Stringer, pp. 199-217. McDonald Institute Monographs, Cambridge. 374 Bar-Yosef, O., A. Belfer-Cohen and D. S. Adler 2006 The Implications of the Middle-Upper Paleolithic Chronological Boundary in the Caucasus to Eurasian Prehistory. Anthropologie XLIV(1):49-60. Bar-Yosef, O. and J.-G. Bordes 2010 Who were the makers of the Châtelperronian culture? Journal of Human Evolution 59:586-593. Barlow, K. R. and D. Metcalfe 1996 Plant Utility Indices: Two Great Basin Examples. Journal of Archaeological Science 23:351-371. Bartram, L. E., E. M. Kroll and H. T. Bunn 1991 Variability in camp structure and bone food refuse patterning at Kua San hunter-gatherer camps. In The Interpretation of Archaeological Spatial Patterning, edited by E. M. Kroll and T. D. Price, pp. 77-148. Plenum Press, New York. Baskin, L. and K. Danell 2003 Ecology of Ungulates: A Handbook of Species in Eastern Europe and Northern and Central Asia. Springer-Verlag, Berlin, Heidelberg. Beck, C., A. K. Taylor, G. T. Jones, C. M. Fadem, C. R. Cook and S. A. Millward 2002 Rocks are heavy: transport costs and Paleoarchaic quarry behavior in the Great Basin. Journal of Anthropological Archaeology 21:481-507. Beckerman, S. 1983 Carpe Diem: An Optimal Foraging Approach to Bari Fishing and Hunting. In Adaptive Responses of Native Amazonians, edited by R. Hames and W. Vickers, pp. 269-299. Academic Press, New York. Behrensmeyer, A. K. 1978 Taphonomic and Ecologic Information from Bone Weathering. Paleobiology 4:150-160. Bennet, J. L. 1999 Thermal Alteration of Buried Bone. Journal of Archaeological Science 26(1):1-8. Bettinger, R. L., R. Malhi and H. McCarthy 1997 Central Place Models of Acorn and Mussel Processing. Journal of Archaeological Science 24:887-899. Bialor, P. A. and M. H. Jameson 1962 Palaeolithic in the Argolid. American Journal of Archaeology 66(2):181-182. 375 Bicho, N. F., J. Haws and B. Hockett 2006 Two sides of the same coin - rocks, bones and site function of Picariero Cave, central Portugal. Journal of Anthropological Archaeology 25:485-499. Binford, L. R. 1968 Post-Pleistocene Adaptations. In New Perspectives in Archaeology, edited by L. R. Binford and S. R. Binford, pp. 313-341. Aldine, Chicago. 1978 Nunamiut Ethnoarchaeology. Academic Press, New York. 1981 Bones: Ancent Men and Modern Myths. Academic Press, New York. 1984a Butchering, sharing, and the archaeological record. Journal of Anthropological Archaeology 3:235-257. 1984b Faunal Remains from Klasies River Mouth. Academic Press, New York. 1985 Human Ancestors: Changing Views of Their Behavior. Journal of Anthropological Archaeology 4:292-327. 1988 Etude taphonomique des restes fauniques de la Grotte Vaufrey, couche VIII. In La Grotte Vaufrey: Paleoenvironment, Chronologie, Activities Humaines, edited by J. P. Rigaud, pp. 535-564. Memoires de la Societe Prehistorique Francais, Paris. 1993 Bones for stones: Considerations of analogues for features found on the Central Russian Plain. In From Kostenki to Clovis: Upper Paleolithic-Paleo-Indian Adaptations, edited by O. Soffer and N. D. Praslov, pp. 101-124. Plenum, New York. Binford, L. R. and J. Bertram 1977 Bone frequencies and attritional processes. In For Theory Building in Archaeology, edited by L. R. Binford, pp. 77-156. Academic Press, New York. Bird, D. W. 1997 Behavioral Ecology and the Archaeological Consequences of Central Place Foraging among the Meriam. In Rediscovering Darwin: Evolutionary Theory and Archaeological Explanation, edited by C. M. Barton and G. A. Clark, pp. 291-308. American Anthropological Association, Arlington. Bird, D. W. and R. L. Bliege Bird 1997 Contemporary Shellfish Gathering Strategies among the Meriam of the Torres Strait Islands, Australia: Testing Predictions of a Central Place Foraging Model. Journal of Archaeological Science 24:39-63. 2000 The Ethnoarchaeology of Juvenile Foragers: Shellfish Strategies among Meriam Children. Journal of Anthropological Archaeology 19:461-476. 376 Bird, D. W. and J. F. O'Connell 2006 Behavioral Ecology and Archaeology. Journal of Archaeological Research 14:143-188. Bird, M. I., L. K. Ayliffe, L. K. Fifield, C. S. M. Turney, R. G. Cresswell, T. T. Barrows and B. David 1999 Radiocarbon dating of "old" charcoal using a wet oxidation, steppedcombustion technique. Radiocarbon 41:127-140. Blasco, M., E. Crespillo and J. M. Sanchez 1986-87 The growth dynamics of Testudo graeca L. (Reptilia: Testudinidae) and other data on its populations in the Iberian Peninsula. Israel Journal of Zoology 34:139-147. Blasco, R. 2008 Human consumption of tortoises at Level IV of Bolomor Cave (Valencia, Spain). Journal of Archaeological Science 35:2839-2848. Blasco, R. and J. Fernández Peris 2009 Middle Pleistocene bird consumption at Level XI of Bolomor Cave (Valencia, Spain). Journal of Archaeological Science 36:2213-2223. Blondel, J. and J. Aronson 1999 Biology and Wildlife of the Mediterranean Region. Oxford University Press, Oxford. Blumenschine, R. J. 1988 An Experimental Model of the Timing of Hominid Carnivore Influence on Archaeological Bone Assemblages. Journal of Archaeological Science 15:483-502. Blumenschine, R. J. and T. C. Magrigal 1993 Variability in long bone marrow yields of East African ungulates and its zooarchaeological implications. Journal of Archaeological Science 20:555-587. Blumenschine, R. J. and M. M. Selvaggio 1991 On the marks of marrow bone processing by hammerstones and hyenas: Their anatomical patterning and archaeological implications. In Cultural Beginnings, edited by J. D. Clark, pp. 17-32. Dr R. Hambelt GMBH, Bonn. Bocheński, Z. M., V. A. Korovin, A. E. Nekrasov and T. Tomek 1997 Fragmentation of bird bones in food remains of imperial eagles (Aquila heliaca). International Journal of Osteoarchaeology 7:165-171. Bocheński, Z. M. and T. Tomek 377 1994 Pattern of bird bone fragmentation in pellets of the Long-eared Owl Asio otus and its taphonomic implications. Acta zoologica cracoviensia 37:177-190. 2010 The Birds of Klissoura Cave 1: a Window into the Upper Palaeolithic Greece. Eurasian Prehistory 7(2):91-106. Bocheński, Z. M., T. Tomek, T. Boev and I. Mitev 1993 Patterns of bird bone fragmentation in pellets of the tawny owl (Strix aluco) and the eagle owl (Bubo bubo) and their taphonomic implications. Acta zoologica cracoviensia 36:313-328. Bond, G., W. Broecker, S. Johnsen, J. McManus, L. Labeyrie, J. Jouzel and G. Bonani 1993 Correlations between climate records from North Atlantic sediments and Greenland ice. Nature 365:143-147. Bond, G., H. Heinrich, W. Broecker, L. Labeyrie, J. McManus, J. Andrews, S. Huon, R. Jantschik, S. Clasen, C. Simet, K. Tedesco, M. Klas, G. Bonani and S. Ivy 1992 Evidence for massive discharges of icebergs into the North Atlantic ocean during the last glacial period. Nature 360:245-249. Bond, G. and R. Lotti 1995 Iceberg Discharges Into the North Atlantic on Millennial Time Scales During the Last Glaciation. Science 267:1005-1010. Boyle, K. V. 2000 Reconstructing Middle Palaeolithic Subsistence Strategies in the South of France. International Journal of Osteoarchaeology 10:336-356. Brain, C. K. 1980 Some criteria for the recognition of bone-collecting agencies in African Caves. In Fossils in the making, edited by A. K. Behrensmeyer and A. P. Hill, pp. 107-130. University of Chicago Press, Chicago. 1981 The Hunters or the Hunted? An Introduction to African Cave Taphonomy. University of Chicago Press, Chicago. Bratlund, B. 1999 Taubach revisited. Jahrbuch des Römisch-Germanischen Zentralmuseums Mainz 46:67-174. Braza, R., C. San Joze and A. Blom 1988 Birth measurements, paturition dates and progeny sex ratio of Dama dama in Donana, Spain. Journal of Mammalogy 69:607-610. 378 Briggs, A. W., J. M. Good, R. E. Green, J. Krause, T. Maricic, U. Stenzel, C. Lalueza-Fox, P. Rudan, D. Brajkovic, Z. Kucan, I. Gusic, R. W. Schmitz, V. B. Doronichev, L. V. Golovanova, M. de la Rasilla, J. Fortea, A. Rosas and S. Paabo 2009 Targeted Retrieval and Analysis of Five Neandertal mtDNA Genomes. Science 325(5938):318-321. Brink, J. W. 1997 Fat content in leg bones of Bison bison, and applications to archaeology. Journal of Archaeological Science 24(259-274). Broughton, J. M. 1994a Declines in mammalian foraging efficiency during the Late Holocene, San Francisco Bay, California. Journal of Anthropological Archaeology 13:371-401. 1994b Late Holocene resource intensification in the Sacramento Valley, California: the vertebrate evidence. Journal of Archaeological Science 21:501-514. 1997 Widening diet breadth, declining foraging efficiency, and prehistoric harvest pressure. Ichthyofaunal evidence from the Emeryville Shellmound, California. Antiquity 71:845-862. 1999 Resource Depression and Intensification During the Late Holocene, San Francisco Bay: Evidence from the Emeryville Shellmound. University of California Press, Berkeley. 2002 Prey spatial structure and behavior affect archaeological tests of optimal foraging models: Examples from the Emeryville Shellmound vertebrate fauna. World Archaeology 34:60-83. Broughton, J. M. and D. K. Grayson 1993 Diet breadth, adaptive change, and the White Mountain Faunas. Journal of Archaeological Science 20:331-336. Broughton, J. M. and J. F. O'Connell 1999 On evolutionary ecology, selectionist archaeology and behavioral archaeology. American Antiquity 64:153-165. Brown, J. H. 1973 Species diversity of seed-eating desert rodents in sand dune habitats. Ecology 54:775-787. 1975 Geographical ecology of desert rodents. In Ecology and Evolution of Communities, edited by M. L. Cody and J. M. Diamond, pp. 315-341. Belknapp Press of Harvard University Press, Cambridge. 379 Bunn, H. T., L. E. Bartram and E. M. Kroll 1988 Variability in Bone Assemblage Formation from Hadza Hunting, Scavenging, and Carcass Processing. Journal of Anthropological Archaeology 7:412-457. Bunn, H. T. and E. M. Kroll 1986 Systematic butchery by Plio/Pleistocene hominids at Olduvai Gorge, Tanzania. Current Anthropology 27:431-452. Burger, O., M. J. Hamilton and R. Walker 2005 The prey as patch model: optimal handling of resources with diminishing returns. Journal of Archaeological Science 32:1147-1158. Burton, J. A. 1991 Field Guide to the Mammals of Britain and Europe. Kingfisher Books, Grisewood & Dempsey Ltd., London. Butler, V. L. 2000 Resource depression on the Northwest Coast of North America. Antiquity 74:649-661. 2001 Changing Fish Use on Mangaia, Southern Cook Islands: Resource Depression and the Prey Choice Model. International Journal of Osteoarchaeology 11:88-100. Butler, V. L. and S. K. Campbell 2004 Resource Intensification and Resource Depression in the Pacific Northwest of North America: A Zooarchaeological Review. Journal of World Prehistory 18(4):327-405. Byers, D. A. and A. 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. Cacho, I., J. O. Grimalt, C. Pelejero, M. Canals, F. J. Sierro, J. A. Flores and N. J. Shackleton 1999 Dansgaard-Oeschger and Heinrich event imprints in Alboran Sea paleotemperatures. Paleoceanography 14(6):698-705. Cain, C. R. 2005 Using burned animal bone to look at Middle Stone Age occupation and behavior. Journal of Archaeological Science 32:873-884. Cannon, M. D. 2000 Large mammal 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. 380 2003 A model of central place forager prey choice and an application to faunal remains from the Mimbres Valley, New Mexico. Journal of Anthropological Archaeology 22:1-25. Cannon, M. D. and D. J. Meltzer 2008 Explaining variability in Early Paleoindian foraging. Quaternary International 191:5-17. Capaldo, S. D. 1997 Experimental determinations of carcass processing by Plio-Pleistocene hominids and carnivores at FLK 22 (Zinjanthropus), Olduvai Gorge, Tanzania. Journal of Human Evolution 33:555-597. Carbonell, E., M. Vaquero, J. Maroto, J. M. Rando and C. Mallol 2000 A Geographic Perspective on the Middle to Upper Paleolithic Transition in the Iberian Peninsula. In The Geography of Neandertals and Modern Humans in Europe and the Greater Mediterranean, edited by O. Bar-Yosef and D. Pilbeam, pp. 5-34. Peabody Museum Bulletin 8. Peabody Museum of Archaeology and Ethnology, Harvard University. Chapman, D. I. and N. G. Chapman 1975 Fallow Deer, their history, distribution and biology. Terence Dalton Ltd., Lavenham, Suffolk. Charnov, E. L. 1976 Optimal Foraging, the Marginal Value Theorem. Theoretical Population Biology 9:129-136. Charnov, E. L., G. H. Orians and K. Hyatt 1976 Ecological Implications of Resource Depression. The American Naturalist 110(972):247-259. Chase, P. G. 1985 On the use of Binford's utility indices in the analysis of archaeological sites. P.A.C.T. 11:287-302. 1989 How Different was Middle Palaeolithic Subsistence? A Zooarchaeological Perspective on the Middle to Upper Palaeolithic Transition. In The Human Revolution: Behavioral and Biological Perspectives on the Origins of Modern Humans, edited by P. Mellars and C. Stringer, pp. 321-337. Edinburgh University Press, Edinburgh. Church, R. R. and R. L. Lyman 381 2003 Small fragments make small differences in efficiency when rendering grease from fractured artiodactyl bones by boiling. Journal of Archaeological Science 30:1077-1084. Churchill, S. E. and B. Smith 2000 Makers of the Early Aurignacian of Europe. Yearbook of Physical Anthropology 43:61-115. Clark, G. A. 2002 Neandertal Archaeology: Implications for our Origins. American Anthropologist 104(1):50-67. Cochard, D. and J. P. Brugal 2004 Importance des fonctions de sites dans les accumulations paléolithiques de léporidés. In Petits Animaux et Sociétés Humaines, du Complément Alimentaire aux Ressources Utilitaires, edited by J. P. Brugal and J. Desse, pp. 283-296. APDCA, Antibes. Colonese, A. C., M. A. Mannino, D. Bar-Yosef Mayer, D. Fa, J. C. Finlayson, D. Lubell and M. C. Stiner 2011 Marine mollusc exploitation in Mediterranean prehistory: An overview. Quaternary International 239:86-103. Conard, N. J. and M. Bolus 2003 Radiocarbon dating the appearance of modern humans and timing of cultural innovations in Europe: new results and new challenges. Journal of Human Evolution 44:331-371. Costamagno, S. 2003 Exploitation de la grande faune au Magdalénien dans le sud de la France. In Mode de Vie au Magdalénien: Les Apports de l'Archéozoologie, edited by S. Costamagno and V. Laroulandie, pp. 73-88. BAR International Series, Oxford. 2004 Si les Magdaléniens du sud de la France n'étient pas des chasseurs spécialisés, qu'étaient-ils? In Approches Fonctionnelles en Préhistoire, edited by P. Bodu and C. Constantin, pp. 361-369. Société Préhistorique français, Paris. Costamagno, S., C. Griggo and V. Mourre 1998 Approche expérimentale d'un problème taphonomique: utilisation de combustible osseux aux Paléolithique. Préhistoire Europeenne 13:167-194. Costamagno, S. and V. Laroulandie 2004 L'exploitation des petits vertébrés dans les Pyrénées françaises du Paléolithique au Mésolithique: un inventaire taphonomique et archéozoologique. In Petits Animaux et Sociétés Humaines: du Complément Alimentaire aux Ressources 382 Utilitaires, edited by J.-P. Brugal and J. Desse, pp. 403-426. Éditions APDCA, Antibes. Costamagno, S., M. Liliane, B. Cedric, V. Bernard and M. Bruno 2006 Les Pradelles (Marillac-le-France, France): a Mousterian Reindeer Hunting Camp? Journal of Anthropological Archaeology 25:466-484. Costamagno, S., I. Théry-Parisot, J. P. Brugal and R. Guibert 2005 Taphonomic consequences of the use of bones as fuel. Experimental data and archaeological applications. In Biosphere to Lithosphere: New studies in vertebrate taphonomy, edited by T. O'Connor, pp. 51-62. Oxbow Books, Oxford. Cramp, S. 1980 Handbook of the Birds of Europe, the Middle East and North Africa. Oxford University Press, Oxford. Cruz-Uribe, K. and R. G. Klein 1998 Hyrax and Hare Bones from Modern South African Eagle Roosts and the Detection of Eagle Involvement in Fossil Bone Assemblages. Journal of Archaeological Science 25:135-147. Cullen, T. 1995 Mesolithic mortuary ritual at Franchthi Cave, Greece. Antiquity 69:270-289. Currey, J. 1984 The Mechanical Adaptations of Bones. Princeton University Press, Princeton, New Jersey. d'Errico, F., M. Julien, D. Liolios, M. Vanhaeren and D. Baffier 2003 Many awls in our argument. Bone tool manufacture and use in the Châtelperronean and Aurignacian levels of the Grotte du Renne at Arcy-sur-Cure. In The Chronology of the Aurignacian and of the Transitional Technocomplexes. Trabalhos de Arqueologia 33, edited by J. Zilhão and F. d'Errico, pp. 247-270. American School of Prehistoric Research/Instituto Portuges de Arqueologia, Lisbon. d'Errico, F. and V. Laroulandie 2000 Bone technology at the Middle-Upper Paleolithic transition. The case of the worked bones from Buran-Kaya III level C (Crimea, Ukraine). In Neanderthals and Modern Humans - Discussing the Transition: Central and Eastern Europe from 50,000-30,000 B.P., edited by J. Orschiedt and G. C. Weniger, pp. 227-242. Neanderthal Museum, Mettmann. d'Errico, F. and M. F. Sanchez Goñi 2003 Neanderthal extinction and the millenia scale climatic variability of OIS 3. Quaternary Science Reviews 22:769-788. 383 d'Errico, F., J. Zilhão, M. Julien, D. Baffier and J. Pelegrin 1998 Neandertal Acculturation in Western Europe? A Critical Review of the Evidence and Its Interpretation. Current Anthropology 39:S1-S44. Dakaris, S. I., E. S. Higgs, R. W. Hey, H. Tippett and P. Mellars 1964 The Climate, Environment and Industries of Stone Age Greece: Part I. Proceedings of the Prehistoric Society XXX:199-244. Dansgaard, W., S. J. Johnsen, H. B. Clausen, D. Dahl-Jensen, N. S. Gundestrup, C. U. Hammer, C. S. Hvidberg, J. P. Seteffensen, A. E. Sveinbjörnsdottir, J. Jouzel and G. Bond 1993 Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364:218-220. Darlas, A. 1995 L'industrie lithique trouvée avec le squelette de femme LAO 1/S 3 de la grotte d'Apidima (Mani - Grèce). Acta Anthropologica (Athens) 1:59-64. 1999 Palaeolithic Research in Western Achaia. In The Palaeolithic Archaeology of Greece and Adjacent Areas. Proceedings of the ICOPAG Conference, Ioannina, September 1994, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 303-310. British School at Athens Studies 3. British School at Athens, London. 2007 Le Moustérien de Grèce è la lumière des récentes recherches. L'anthropologie 111:346-366. Darlas, A. and H. de Lumley 1999 Palaeolithic Research in Kalamakia Cave, Areopolis, Peloponnese. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perles and K. Zachos, pp. 293-302. Technical Print Services Ltd, Nottingham, Great Britain. 2004 La Grotte de Kalamakia (Areopolis, Grèce). Sa Contribution à la Connaissance du Paléolithique Moyen de Grèce. In Actes de la 5e Section: Le Paléolithique Moyen, pp. 225-233. BAR International Series 1239. Darlas, A. and E. Psathi 2008 Le Paléolithique Supérieur dans la Péninsule du Mani (Péloponnèse, Grèce). In The Palaeolithic of the Balkans: Proceedings of the XV World COngress (Lisbon, 4-9 September 2006), edited by A. Darlas and D. Milhailovic, pp. 51-59. BAR International Series 1819, Oxford. David, F. and J. G. Enloe 384 1993 L'exploitation des animaux sauvages de la fin du Paléolithique moyen au Magdalénien. In Exploitation des animaux sauvages a travers le temps, pp. 29-47. XIIIe Recontres Internationales d'Archéologie et d'Histoire d'Antibes. Éditions APDCA, Juan-les-Pins. Davis, S. J. 1977 The ungulate remains from Kebara Cave. In Moshé Stekelis Memorial Volume, edited by B. Arensburg and O. Bar-Yosef, pp. 150-163. Eretz-Israel: Archaeological, Historical and Geographical Studies 13. The Israel Exploration Society, Jerusalem. 1983 The age profiles of gazelles predated by ancient man in Israel: possible evidence for a shift from seasonality to sedentism in the Natufian. Paléorient 9:55-62. 1985 A taphonomic approach to experimental bone fracturing and applications to several South African Pleistocene sites. Ph.D. Dissertation, State University of New York and Binghamton. 1987 The Archaeology of Animals. Yale University Press, London. 2002 The mammals and birds from the Gruta do Caldierão, Portugal. Revista Portuguesa de Arqueologia 5(2):29-98. De Vivo, B., G. Rolandi, P. B. Gans, A. Calvert, W. A. Bohrson, F. J. Spera and H. E. Belkin 2001 New constraints on the pyroclastic eruptive history of the Campanian volcanic Plain (Italy). Mineralogy and Petrology 73:47-65. Delpech, F. 1998 Comment on Marean and Kim, Mousterian large-mammal remains from Kobeh Cave: Behavioral implications. Current Anthropology 39(594-595). Deniz, E. and S. Payne 1982 Eruption and Wear in the Mandibular Dentition as a Guide to Aging Turkish Angora Goats. In Ageing and Sexing Animal Bones from Archaeological Sites, edited by B. Wilson, C. Grigson and S. Payne, pp. 155-205. BAR Series 109. British Archaeological Reports, Oxford. Denys, C. 2002 Taphonomy and Experimentation. Archaeometry 44(3):469-484. Dousougli, A. 1999 Palaeolithic Leukas. In The Palaeolithic Archaeology of Greece and Adjacent Areas. Proceedings of the ICOPAG Conference, Ioannina, September 1994, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 288-292. British School at Athens Studies 3. British School at Athens, London. 385 Duarte, C., J. Mauricio, P. B. Pettitt, P. Souto, E. Trinkhaus, H. van der Plicht and J. Zilhão 1999 The Early Upper Paleolithic human skeleton from the Abrigo do Lagar Velho (Portugal) and modern human emergence in Iberia. PNAS 96:7604-7609. Dwyer, P. D. 1986 A Hunt in New Guinea: Some Difficulties for Optimal Foraging Theory. Man, New Series 20(22):243-253. Egeland, C. P. and R. M. Byerly 2005 Application of Return Rates to Large Mammal Butchery and Transport among Hunter-gatherers and its Implications for Plio-Pleistocene Hominid Carcass Foraging and Site Use. Journal of Taphonomy 3(3):135-157. Elefanti, P. 2003 Subsistence Strategies in Greece During the Upper Palaeolithic from the Perspective of Lithic Technology. BAR International Series 1130. Biddles Ltd., England. Elefanti, P., E. Panagopoulou and P. Karkanas 2008 The Transition from the Middle to the Upper Paleolithic in the Southern Balkans: the Evidence from the Lakonis I Cave, Greece. Eurasian Prehistory 5(2):8595. Emlen, J. M. 1966 The Role of Time and Energy in Food Preference. American Naturalist 100:611-617. Enloe, J. G. 1991 Subsistence Organization in the Upper Paleolithic: Carcass Refitting and Food Sharing at Pincevent. Ph. D. Dissertation, University of New Mexico. 1997 Seasonality and age structure in remains of Rangifer tarandus: Magdalenian hunting strategy at Verberie. Anthropozoologica 25-26:95-102. 2003a Acquisition and processing of reindeer in the Paris Basin. In Zooarchaeological Insights into Magdalenian Lifeways, edited by S. Costamagno and V. Laroulandie, pp. 23-31. BAR International Series 1144, Oxford. 2003b Food sharing past and present: archaeological evidence for economic and social interactions. Before Farming 1(1):1-23. Enloe, J. G. and F. David 386 1989 Le remontage des os par individus: le partage du renne chez les Magdaléniens de Pincevent (La Grande Paroisse, Seine-et-Marne). Bulletin de la Société Préhistorique Française 86(9):275-281. 1992 Food Sharing in the Paleolithic: Carcass Refitting at Pincevent. In Piecing Together the Past: Applications of Refitting Studies in Archaeology, edited by J. L. Hofman and J. G. Enloe, pp. 296-315. BAR International Series 578, Oxford. Facorellis, Y., N. Kyparissi-Apostolika and Y. Maniatis 2001 Cave of Theopetra, Kalambaka: Radiocarbon Evidence for 50,000 Years of Human Presence. Radiocarbon 43(2B):1029-1048. Faith, J. T. 2007 Changes in reindeer body part representation at Grotte XVI, Dordogne, France. Journal of Archaeological Science 34:2003-2011. Faith, J. T. and A. D. Gordon 2007 Skeletal element abundances in archaeofaunal assemblages: economic utility, sample size, and assessment of carcass transport strategies. Journal of Archaeological Science 34(6):872-882. Farizy, C., F. David and J. Jaubert 1994 Hommes et Bisons du Paléolithique Moyen á Mauran (Haute Garonne). CNRS Editions, Paris. Farrand, W. R. 2000 Depositional History of Franchthi Cave: Sediments, Stratigraphy, and Chronology. Excavations at Franchthi Cave, Greece, Fascicle 12. Indiana University Press, Bloomington and Indianapolis. Feldhamer, G. A., K. C. Farris-Renner and C. M. Barker 1988 Dama dama. Mammalian Species 317:1-8. Felsch, R. C. S. 1973 Die Höhle von Kephalari, eine jungpaläolithische Siedlung in der Argolis. Athens Annals of Archaeology 6:13-27. Fernandez-Jalvo, Y., B. Sanchez-Chillon, P. Andrews, S. Fernandez-Lopez and L. Alcala Martinez 2002 Morphological Taphonomic Transformations of Fossil Bones in Continental Environments, and Repercussions on their Chemical Composition. Archaeometry 44(3):353-361. Finlayson, J. C., F. Giles-Pacheco, J. Rodríguez-Vidal, D. A. Fa, J. M. G. López, A. S. Pérez, G. Finlayson, E. Allue, J. B. Preysler, I. Cáceres, J. S. Carrión, Y. F. Jalvo, C. P. Gleed- 387 Owen, F. J. Espejo, P. López, J. A. L. Sáez, J. A. R. Cantal, A. S. Marco, F. G. Guzman, K. Brown, N. Fuentes, C. A. Valarino, A. Villalpando, C. B. Stringer, F. M. Ruiz and T. Sakamoto 2006 Late survival of Neanderthals at the southernmost extreme of Europe. Letters to Nature 443:850-853. Fiore, I., M. Gala and A. Tagliacozzo 2004 Ecology and Subsistence Strategies in the Eastern Italian Alps during the Middle Palaeolithic. International Journal of Osteoarchaeology 14:273-286. Fisher, J. W., Jr. 1995 Bone surface modifications in zooarchaeology. Journal of Archaeological Method and Theory 2(1):7-68. Fletcher, W. J. and M. F. Sánchez Goñi 2008 Orbital- and sub-orbital-scale climate impacts on vegentation of the western Mediterranean basin over the last 48,000 yr. Quaternary Research 70:451-464. Galanidou, N. 1997 Lithic Refitting and Site Structure at Kastritsa. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 2: Klithi in its local and regional setting, edited by G. N. Bailey, pp. 497-520. McDonald Institute Monographs, Cambridge. 1999 Regional settlement and intra-site spatial patterns in Upper Palaeolithic Epirus. In The Palaeolithic Archaeology of Greece and Adjacent Areas. Proceedings of the ICOPAG Conference Ioannina, September 1994, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 148-158. British School at Athens Studies 3, London. 2003 Reassessing the Greek Mesolithic: the pertinence of the Markovits collections. In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 99-112. British School at Athens Studies 10, London. Galanidou, N. and P. C. Tzedakis 2001 New AMS dates from Upper Palaeolithic Kastritsa. Proceedings of the Prehistoric Society 67:271-278. Galanidou, N., P. C. Tzedakis, I. T. Lawson and M. R. Frogley 2000 A Revised Chronological and Palaeoenvironmental Framework for the Kastritsa Rockshelter, Northwest Greece. Antiquity 74:349-355. Gambier, D. 1989 Fossil Hominids of the Early Upper Paleolithic (Aurignacian) of France. In The Human Revolution: Behavioural and Biological Persprctives on the Origins of 388 Modern Humans, edited by P. Mellars and C. Stringer, pp. 194-211. Edinburgh Univeristy Press, Edinburgh. Gamble, C. 1986 The Palaeolithic Settlement of Europe. Cambridge University Press, Cambridge. 1997 The Animal Bones from Klithi. In Klithi: Paleolithic Settlement and Quaternary Landscapes in Northwest Greece. Volume 1: Excavation and intra-site analysis at Klithi, edited by G. N. Bailey, pp. 207-244, Cambridge. 1999 Faunal Exploitation at Klithi. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perles and K. Zachos, pp. 179-187, T. B. S. a. Athens, general editor. Technical Print Services Ltd, Nottingham, Great Britain. Gardeisen, A., K. Trantalidou and A. Darlas 1999 Faunal remains from Kalamakia cave (Peloponnese, Greece). In Human Population Origins in the Circum-Mediterranean Area: Adaptations of the HunterGatherer Groups to Environmental Modifications, edited by A. R. Cruz, S. Milliken, L. Oosterbeek and C. Peretto, pp. 111-121. Instituto Politecnico de Tomar, Tomar. Gaudzinski-Windheuser, S. and L. Niven 2009 Hominin Subsistence Patterns During the Middle and Late Paleolithic in Northwestern Europe. In The Evolution of Hominin Diets: Integrating Approaches to the Study of Palaeolithic Subsistence, edited by J.-J. Hublin and M. P. Richards, pp. 99-111. Springer Science + Business Media B.V. Gaudzinski, S. 1995 Wallertheim Revisited: a Re-analysis of the Fauna from the Middle Palaeolithic site of Wallertheim (Rheinhessen/Germany). Journal of Archaeological Science 22:51-66. 2006 Monospecific or Species-Dominated Faunal Assemblages during the Middle Paleolithic in Europe. In Transitions Before the Transition: Evolution and Stability in the Middle Paleolithic and Middle Stone Age, edited by E. Hovers and S. L. Kuhn, pp. 137-147. Springer Science and Business Media, Inc., New York. Gaudzinski, S. and W. Roebroeks 2000 Adults only: reindeer hunting at the Middle Paleolithic site SalzgitterLebenstedt, northern Germany. Journal of Human Evolution 38:491-521. Geraga, M., S. Tsaila-Monopolis, C. Ioakim, G. Papatheodorou and G. Ferentinos 2005 Short-term climate changes in the southern Aegean Sea over the last 48,000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 220:311-332. 389 Gifford-Gonzalez, D. P. 1989 Ethnographic analogues for interpreting modified bones: some cases from East Africa. In Bone modificaton, edited by R. Bonnichsen and M. H. Sorg, pp. 179246. University of Main Center for the Study of the First Americans, Orono. Gifford, D. P. 1981 Taphonomy and paleoecology: a critical review of archaeology's sister disciplines. In Advances in archaeological method and theory, edited by M. B. Schiffer, pp. 365-438. vol. 4. Academic Press, New York. Gowlett, J. and P. Carter 1997 The Basal Mousterian of Asprochaliko Rockshelter, Louros Valley. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 2: Klithi in its local and regional setting, edited by G. N. Bailey, pp. 441-458. McDonald Institute Monographs, Cambridge. Grant, A. 1982 The use of tooth wear as a guide to the age of domestic animals. In Aging and Sexing Animal Bones from Archaeological Sites, edited by B. Wilson, C. Grigson and S. Payne, pp. 91-108. vol. 109. British Archaeological Reports, Oxford. Graves, H. B. 1984 Behavior and Ecology of Wild and Feral Swine (Sus scrofa). Journal of Animal Science 58(2):482-492. Grayson, D. K. 1984 Quantitative Zooarchaeology. Academic Press, Orlando. 1988 Danger Cave, Last Supper Cave, and Hanging Rock Shelter: the faunas. American Museum of Natural History Anthropological Papers 66(1). 1989 Bone Transport, Bone Destruction, and Reverse Utility Curves. Journal of Archaeological Science 16:643-652. 1998 Moisture History and Small Mammal Community Richness during the Latest Pleistocene and Holocene, Northern Bonneville Basin, Utah. Quaternary Research 49:330-334. 2000 The Homestead Cave Mammals. In Late Quaternary Paleoecology in the Great Basin, edited by D. B. Madsen. Utah Geological Survey Bulletin 130, Salt Lake City. Grayson, D. K. and M. D. Cannon 390 1999 Human paleoecology and foraging theory in the Great Basin. In Current Models in Great Basin Anthropology, edited by C. Beck. University of Utah, Salt Lake City. Grayson, D. K. and F. Delpech 1994 The Evidence for Middle Paleolithic Scavenging from Couche VIII, Grotte Vaufrey (Dordogne, France). Journal of Archaeological Science 21:359-375. 1998 Changing Diet Breadth in the Early Upper Paleolithic of Southwestern France. Journal of Archaeological Science 25:1119-1129. 2002 Specialized Early Upper Palaeolithic Hunters in Southwestern France? Journal of Archaeological Science 29:1439-1449. 2003 Ungulates and the Middle-to-Upper Paleolithic transition at Grotte XVI (Dordogne, France). Journal of Archaeological Science 30:1633-1648. 2006 Was there Increasing Dietary Specialization Across the Middle-to-Upper Paleolithic Transition in France? In When Neandertals and Modern Humans Met, edited by N. J. Conard, pp. 377-417, Tubingen. Grayson, D. K., F. Delpech, J. P. Rigaud and J. F. Simek 2001 Explaining the Development of Dietary Dominance by a Single Ungulate Taxon at Grotte XVI, Dordogne, France. Journal of Archaeological Science 28:115125. Green, R. E., J. Krause, A. W. Briggs, T. Maricic, U. Stenzel, M. Kircher, N. Patterson, H. Li, W. Zhai, M. H.-Y. Fritz, N. F. Hansen, E. Y. Durand, A.-S. Malaspinas, J. D. Jensen, T. Marques-Bonet, C. Alkan, K. Prüfer, M. Meyer, H. Burbano, J. M. Good, R. Schulz, A. Aximu-Petri, A. Butthof, B. Höber, B. Höffner, M. Siegemund, A. Weihmann, C. Nusbaum, E. S. Lander, C. Russ, N. Novod, J. Affourtit, M. Egholm, C. Verna, P. Rudan, D. Brajkovic, Ž. Kucan, I. Gušic, V. B. Doronichev, L. V. Golovanova, C. Lalueza-Fox, M. de la Rasilla, J. Fortea, A. Rosas, R. W. Schmitz, P. L. F. Johnson, E. E. Eichler, D. Falusch, E. Birney, J. C. Mullikin, M. Slatkin, R. Nielsen, J. Kelso, M. Lachmann, D. Reich and S. Pääbo 2010 A Draft Sequence of the Neandertal Genome. Science 328:710-722. Grootes, P. M., M. Stuiver, J. W. C. White, S. J. Johnsen and J. Jouzel 1993 Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature 366:552-554. Hahn, J. 1984 Forschungen zur allgemeinen und vergleichenden Archäologie: Südeuropa und Nordafrika. C.H. Beck, München. Hailey, A., J. Wright and E. Steer 391 1988 Population Ecology and Conservation of Tortoises: the Effects of Disturbance. Herpetological Journal 1:294-301. Hall, S. J. G. 2008 A comparative analysis of the habitat of the extinct aurochs and other prehistoric mammals in Britain. Ecography 31:187-190. Hames, R. and W. Vickers 1982 Optimal diet breadth theory as a model to explain variability in Amazonian hunting. American Ethnologist 9:358-378. Handrinos, G. and T. Akriotis 1997 The Birds of Greece. Christopher Helm Ltd., A & C Black Ltd., London. Hansen, J. M. 1978 The earliest seed remains from Greece: Palaeolithic through Neolithic at Franchthi Cave. Berichte der Deutschen Botanischen Gesellschaft 91:39-46. 1991 The Palaeoethnobotany of Franchthi Cave. Excavations at Franchthi Cave, Greece Fascicle 7. Indiana University Press, Bloomington & Indianapolis. Harrison, S. and G. Digerfeldt 1993 European Lakes as Palaeohydrological and Palaeoclimatic Indicators. Quaternary Science Reviews 12:233-248. Harvati, K., E. Panagopoulou and P. Karkanas 2003 First Neandertal Remains from Greece: the Evidence from Lakonis I. Journal of Human Evolution 21:359-375. Harvati, K., E. Panagopoulou and C. Runnels 2009 The Paleoanthropology of Greece. Evolutionary Anthropology 18:131-143. Hawkes, K. 1990 Why do men hunt? Benefits for risky choices. In Risk and Uncertainty in Tribal and Peasant Economies, edited by E. Cashdan, pp. 145-166. Westview Press, Boulder. 1991 Showing off: Tests of a hypothesis about men's foraging goals. Ethology and Sociobiology 12:29-54. Hawkes, K. and R. L. Bliege Bird 2002 Showing off, handicap signaling, and the evolution of men's work. Evolutionary Anthropology 11:58-67. Hawkes, K., K. Hill and J. F. O'Connell 392 1982 Why hunters gather: optimal foraging and the Ache of eastern Paraguay. American Ethnologist 9:379-398. Hawkes, K. and J. F. O'Connell 1981 Affluent hunteres? Some comments in light of the Alyware case. American Anthropologist 83:622-626. 1985 Optimal foraging models and the case of the !Kung. American Anthropologist 87:401-404. Heinzel, H., R. Fitter and J. Parslow 1992 Birds of Britain and Europe (with North Africa and the Middle East). Butler & Tanner Ltd., London. Hennig, G. J., W. Herr, E. Weber and N. I. Xirotiris 1981 ESR-dating of the fossil hominid cranium from Petralona Cave, Greece. Nature 292:533-536. Higgs, E. S. and C. Vita-Finzi 1966 The Climate, Environment and Industries of Stone Age Greece: Part II. Proceedings of the Prehistoric Society XXXII:1-29. Higgs, E. S., C. Vita-Finzi, D. R. Harris, A. E. Fagg and S. Bottema 1967 The Climate, Environment and Industries of Stone Age Greece: Part III. Proceedings of the Prehistoric Society XXXIII:1-29. Higham, T., F. Brock, M. Peresani, A. Broglio, R. Wood and K. Douka 2009 Problems with radiocarbon dating the Middle to Upper Palaeolithic transition in Italy. Quaternary Science Reviews 28:1257-1267. Higham, T., R. Jacobi, M. Julien, F. David, L. Basell, R. Wood, W. Davies and C. Bronk Ramsey 2010 Chronology of the Grotte du Renne (France) and implications for the context of ornaments and human remains within the Châtelperronian. Proceedings of the National Academy of Sciences 107(47):20234-20239. Hill, K. and K. Hawkes 1983 Neotropical hunting among the Ache of eastern Paraguay. In Adaptive Responses of Native Amazonians, edited by R. Hames and W. Vickers, pp. 139-188. Academic Press, New York. Hill, K., H. Kaplan, K. Hawkes and A. M. Hurtado 1987 Foraging Decisions Among Ache Hunter-Gatherers: New Data and Implications for Optimal Foraging Models. Ethology and Sociobiology 8:1-36. 393 Hill, M. E., Jr. 2007 Causes of Regional and Temporal Variation in Paleoindian Diet in Western North America. Doctoral Dissertation, University of Arizona. Hillson, S. 2005 Teeth. Cambridge University Press, Cambridge. Hockett, B. 1989 The concept of "carrying range:" a method for determining the role played by woodrats in contributing bones to archaeological sites. Nevada Archaeologist 7:2835. 1994 A descriptive reanalysis of the leporid bones from Hogup Cave, Utah. Journal of California and Great Basin Anthropology 16(1):106-117. 1996 Corroded, Thinned and Polished Bones Created by Golden Eagles (Aquila chrysaetos): Taphonomic Implications for Archaeological Interpretations. Journal of Archaeological Science 23:587-591. 1999 Taphonomy of a Carnivore-Accumulated Rabbit Bone Assemblage from Picareiro Cave, Central Portugal. Journal of Iberian Archaeology 1:251-257. Hockett, B. and N. F. Bicho 2000 The Rabbits of Picareiro Cave: Small Mammal Hunting During the Late Upper Palaeolithic in the Portuguese Estremadura. Journal of Archaeological Science 27:715-723. Hockett, B. and J. Haws 2002 Taphonomic and Methodological Perspectives of Leporid Hunting During the Upper Paleolithic of the Western Mediterranean Basin. Journal of Archaeological Method and Theory 9(3):269-302. 2009 Continuity in animal resource diversity in the Late Pleistocene human diet of Central Portugal. Before Farming [online version] 2(2):1-14. Hoffecker, J. F., G. Baryshnikov and O. Potapova 1991 Vertebrate remains from the Mousterian site of Il'skaya I (northern Caucasus, USSR): New analysis and interpretation. Journal of Archaeological Science 18:113147. Hoffman, R. and C. Hays 1987 The eastern wood rat (Neotoma floridana) as a taphonomic factor in archaeological sites. Journal of Archaeological Science 14:325-337. Hovers, E. and A. Belfer-Cohen 394 2006 Now You See It, Now You Don't: Modern Human Behavior in the Middle Paleolithic. In Transitions Before the Transition, edited by S. L. Kuhn and E. Hovers, pp. 295-305. Springer, New York. Hublin, J.-J., F. Spoor, F. Zonneveld and S. Condemi 1996 A Late Neanderthal associated with Upper Palaeolithic artefacts. Nature 381:224-226. Hughes, P. D., J. C. Woodward and P. L. Gibbard 2006 Late Pleistocene glaciers and climate in the Mediterranean. Global and Planetary Change 50:83-98. Huntley, B. and J. R. M. Allen 2003 Glacial Environments III: Palaeo-vegetation Patterns in Last Glacial Europe. In Neanderthals and Modern Humans in the European Landscape During the Last Glaciation: Archaeological Results of the Stage 3 Project, edited by W. Davies and T. H. Van Andel, pp. 79-102. McDonald Institute for Archaeological Research, Cambridge. Hurtado, A. M., K. Hawkes, K. Hill and H. Kaplan 1985 Female Subsistence Strategies Among Ache Huner-Gatherers of Eastern Paraguay. Human Ecology 13(1):1-28. Huxtable, J., J. Gowlett, G. N. Bailey, P. Carter and V. Papaconstantinou 1992 Thermoluminescence Dates and a New Analysis of the Early Mousterian from Asprochaliko. Current Anthropology 33(1):109-114. Ikeya, M. 1980 Dating of carbonates at Petralona Cave. Anthropos (Athens) 7:143-151. Ioannidou, E. 2003 Taphonomy of Animal Bones: Species, Sex, Age and Breed Variability of Sheep, Cattle and Pig Bone Density. Journal of Archaeological Science 30:355-365. Jacobsen, T. W. 1981 Franchthi Cave and the beginning of settled village life in Greece. Hesperia 50:303-319. Jacobsen, T. W. and W. R. Farrand 1987 Franchthi Cave and Paralia: Maps, Plans and Sections. Excavations at Franchthi Cave, fasc. 1. Indiana University Press, Indiana. James, S. R. 1990 Monitoring Archaeofaunal Changes During the Transition to Agriculture in the American Southwest. Kiva 56:25-43. 395 Janis, C. 2008 An Evolutionary History of Browsing and Grazing Ungulates. In The Ecology of Browsing and Grazing, edited by I. J. Gordon and H. H. T. Prins, pp. 21-45. Ecological Studies 195. Springer. Jaubert, J., M. Lorblanchet, H. Laville, R. Slott-Moller, A. Turq and J.-P. Brugal 1990 Les Chasseurs d'Aurochs de La Borde: un Site du Paléolithique Moyen. Maison des Sciences de l'Homme, DAF No. 27, Paris. Jochim, M. 1988 138. Optimal foraging and the division of labor. American Anthropologist 90:130- Johnsgard, P. A. 1991 Bustards, Hemipodes, and Sandgrouze: Birds of Dry Places. Oxford University Press, Oxford. Johnson, E. 1985 Current developments in bone technology. In Advances in Archaeological Method and Theory, edited by M. B. Schiffer, pp. 157-235. vol. 8. Academic Press, New York. Jones, E. 2004 Dietary Evenness, Prey Choice, and Human-Environmental Interractions. Journal of Archaeological Science 31(307-317). 2006 Prey choice, mass collecting, and the wild European rabbit (Oryctolagus cuniculus). Journal of Anthropological Archaeology 25:275-289. 2007 Subsistence change, landscape use, and changing site elevation at the Pleistocene-Holocene transition in the Dordogne of southwestern France. Journal of Archaeological Science 34:344-353. 2009 Climate change, patch choice, and intensification at Pont d'Ambon (Dordogne, France) during the Younger Dryas. Quaternary Research 72(3):371-376. Jones, K. T. and D. B. Madsen 1989 Calculating the Cost of Resource Transportation: a Great Basin Example. Current Anthropology 30(4):529-534. Jones, K. T. and D. Metcalfe 1988 Bare bones archaeology: bone marrow indices and efficiency. Journal of Archaeological Science 15:415-423. 396 Joseph, S. 2000 Anthropological evolutionary ecology: A critique. Journal of Ecological Anthropology 4:6-30. Kaczanowska, M., J. K. Kozlowski, K. Sobczyk and J. Wilczyński 2010 Upper Palaeolithic Human Occupations and Material Culture. Eurasian Prehistory 7(2):133-285. Karali-Giannakopoulou, I. 1995 Preliminary report on malacological material found in Apidima (Laconia). Acta Anthropologica (Athens) 1:159-163. Karkanas, P. 2001 Site Formation Processes in Theopetra Cave: A Record of Climatic Change during the Late Pleistocene and Early Holocene in Thessaly, Greece. Geoarchaeology 16(4):373-399. 2010 Geology, stratigraphy and site formation processes of the Upper Palaeolithic and later sequence in Klissoura Cave 1. Eurasian Prehistory 7(2):15-36. Karkanas, P., M. Koumouzelis, J. K. Kozlowski, V. Sitlivy, K. Sobczyk, F. Berna and S. Weiner 2004 The Earliest Evidence for Clay Hearths: Aurignacian Features in Klisoura Cave 1, Southern Greece. Antiquity 78:513-525. Karkanas, P., N. Kyparissi-Apostolika, O. Bar-Yosef and S. Weiner 1999 Mineral Assemblages in Theopetra, Greece: A Framework for Understanding Diagenesis in a Prehistoric Cave. Journal of Archaeological Science 26:1171-1180. Kelly, R. L. 1995 The Foraging Spectrum: Diversity in Hunter-Gatherer Lifeways. Smithsonian Institution Press, Washington. Kersten, A. M. P. 1987 Age and sex composition of Epipalaeolithic fallow deer and wild goat from Ksar 'Akil. Palaeohistoria 29:119-131. Klein, R. G. and K. Cruz-Uribe 1984 The Analysis of Animal Bones from Archaeological Sites. The University of Chicago Press, Chicago. 1996 Exploitation of large bovids and seals at Middle and Later Stone Age Sites in South Africa. Journal of Human Evolution 31:315-334. Kooyman, B. 397 1990 Moa procurement: communal or individual hunting? In Hunters of the Recent Past, edited by L. B. Davis and B. O. K. Reeves, pp. 327-351. Unwin Hyman, London. Kot, M. A. 2009 Analysis of Clay Hearths from Site Klissoura 1 in Greece. Unpublished report. Kotjabopoulou, E. 2001 Patterned Fragments and Fragments of Patterns: Upper Palaeolithic Rockshelter Faunas from Epirus Northwestern Greece. Doctoral Dissertation, University of Cambridge. Kotjabopoulou, E. and E. Adam 2004 People, Mobility and Ornaments in Upper Palaeolithic Epirus, NW Greece. In La Spiritualité: Actes du colloque de la commission 8 de l'UISPP (Paléolithique supérieur), Liege, 10-12 décembre 2003, edited by M. Otte, pp. 37-53. ERAUL 106, Liege. Kotjabopoulou, E., E. Panagopoulou and E. Adam 1997 The Boïla Rockshelter: a Preliminary Report. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 2: Klithi in its local and regional setting, edited by G. N. Bailey, pp. 427-437. McDonald Institute for Archaeological Research, Cambridge. 1999 The Boila Rockshelter: Further Evidence of Human Activity in the Voidomatis Gorge. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perles and K. Zachos, pp. 197-210, T. B. S. a. Athens, general editor. Technical Print Services Ltd, Nottingham, Great Britain. Kotthoff, U., U. C. Müller, J. Pross, G. Schmiedl, I. T. Lawson, B. van de Schootbrugge and H. Schulz 2008 Lateglacial and Holocene vegetation dynamics in the Aegean region: an integrated view based on pollen data from marine and terrestrial archives. The Holocene 18(7):1019-1032. Koumouzelis, M., B. Ginter, J. K. Kozlowski, M. Pawlikowski, O. Bar-Yosef, R. M. Albert, M. Litynska-Zajac, E. Stworzewicz, P. Wojtal, G. Lipecki, T. Tomek, Z. M. Bocheński and A. Pazdur 2001 The Early Upper Palaeolithic in Greece: The Excavations in Klisoura Cave. Journal of Archaeological Science 28:515-539. Koumouzelis, M., J. K. Kozlowski, C. Escutenaire, V. Sitlivy, K. Sobczyk, H. Valladas, N. Tisnerat-Laborde, P. Wojtal and B. Ginter 398 2001 La fin du Paléolithique moyen et le début du Paléolithique supérieur en Grèce: la séquence de la Grotte 1 de Klissoura. L'anthropologie 105:469-504. Koumouzelis, M., J. K. Kozlowski and B. Ginter 2003 Mesolithic finds from Cave 1 in the Klisoura Gorge, Argolid. In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 113-122. British School at Athens Studies 10, London. Koumouzelis, M., J. K. Kozlowski and M. Kaczanowska 2004 End of the Paleolithic in the Argolid (Greece): Excavations in Cave 4 and Cave 7 in the Klisoura Gorge. Eurasian Prehistory 2(2):33-56. Koumouzelis, M., J. K. Kozlowski, M. Nowack, K. Sobczyk, M. Kaczanowska, M. Palikowski and A. Pazdur 1996 Prehistoric settlement in the Klisoura Gorge, Argolid, Greece (excavation 1993, 1994). Préhistoire Européenne 8:143-173. Kozlowski, J. K. 2000 The Problem of Cultural Continuity between the Middle and Upper paleolithic in Central Europe. In The Geography of Neandertals and Modern Humans in Europe and the Greater Mediterranean, edited by O. Bar-Yosef and D. Pilbeam, pp. 77-105. Peabody Museum Bulletin 8. Peabody Museum of Archaeology and Ethnology, Harvard University. 2007 The Significance of Blade Technologies in the Period 50-35 kya BP for the Middle-Upper Palaeolithic Transition in Central and Eastern Europe. In Rethinking the Human Revolution: New behavioural and biological perspectives on the origin and dispersal of modern humans, edited by P. Mellars, K. V. Boyle, O. Bar-Yosef and C. Stringer, pp. 317-328. McDonald Institute Monographs, Cambridge. Kreutzer, L. A. 1992 Bison and Deer Bone Mineral Densities: Comparisons and Implications for the Interpretation of Archaeological Faunas. Journal of Archaeological Science 19:271-294. Krings, M., A. Stone, R. W. Schmitz, H. Krainitzki, M. Stoneking and S. Paabo 1997 Neandertal DNA Sequences and the Origin of Modern Humans. Cell 90(1):19-30. Kuhn, S. L. and A. Biette 2000 The Late Middle and Early Upper Paleolithic in Italy. In The Geography of Neandertals and Modern Humans in Europe and the Greater Mediterranean, edited by O. Bar-Yosef and D. Pilbeam, pp. 49-76. Peabody Museum Bulletin 8. Peabody Museum of Archaeology and Ethnology, Harvard University. 399 Kuhn, S. L., J. Pigati, P. Karkanas, M. Koumouzelis, J. K. Kozlowski and M. Ntinou 2010 Radiocarbon Dating Results for the Early Upper Paleolithic of Klissoura Cave 1. Eurasian Prehistory 7(2):37-46. Kuhn, S. L. and M. C. Stiner 2001 The antiquity of hunter-gatherers. In Another Day, Another Camp: An Interdisciplinary View of Hunter-gatherers, edited by C. Panter-Brick, R. H. Layton and P. A. Rowley-Conwy, pp. 99-142. Cambridge University Press, Cambridge. 2007 Body Ornamentation as Information Technology: Towards an Understanding of the Significance of Early Beads. In Rethinking the Human Revolution: New behavioural and biological perspectives on the origin and dispersal of modern humans, edited by P. Mellars, K. V. Boyle, O. Bar-Yosef and C. Stringer, pp. 45-54. McDonald Institute Monographs, Cambridge. Kuhn, S. L., M. C. Stiner, E. Güleç, I. Özer, H. Yılmaz, I. Baykara, A. Açıkkol, P. Goldberg, K. Martínez Molina, E. Ünay and F. Suata-Alpaslan 2009 The Early Upper Paleolithic occupations at Üçağızlı Cave (Hatay, Turkey). Journal of Human Evolution 56:87-113. Kuhn, S. L., M. C. Stiner, D. S. Reese and E. Gulec 2001 Ornaments of the Earliest Upper Paleolithic: New Insights from the Levant. PNAS 98(13):7641-7646. Kurtén, B. and A. N. Poulianos 1977 New stratigraphic and faunal material from Petralona Cave, with special reference to the Carnivora. Anthropos (Athens) 4:47-130. Kyparissi-Apostolika, N. 1999 The Palaeolithic Deposits of Theopetra Cave in Thessaly. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perles and K. Zachos, pp. 232-239, T. B. S. a. Athens, general editor. Technical Print Services Ltd, Nottingham, Great Britain. Lahr, M. M. and R. A. Foley 2003 Demography, Dispersal and Human Evolution in the Last Glacial Period. In Neanderthals and Modern Humans in the European Landscape During the Last Glaciation: Archaeological Results of the Stage 3 Project, edited by T. H. van Andel and W. Davies, pp. 241-256. McDonald Institute for Archaeological Research, Cambridge, England. Lam, Y. M., X. Chen, C. W. Marean and C. J. Frey 400 1998 Bone Density and Long Bone Representation in Archaeological Faunas: Comparing Results from CT and Photon Densitometry. Journal of Archaeological Science 25:559-570. Lam, Y. M., X. Chen and O. M. Pearson 1999 Intertaxonomic Variability in Patterns of Bone Density and the Differential Representation of Bovid, Cervid, and Equid Elements in the Archaeological Record. American Antiquity 64(2):343-362. Lambert, M. R. K. 1982 Studies on the growth, structure, and abundance of the Mediterranean spurthighed tortoise, Testudo graeca, in field populations. Journal of Zoology, London 196:165-189. Landals, A. 1990 The Maple Leaf site: implications of the analysis of small-scale bison kills. In Hunters of the Recent Past, edited by L. B. Davis and B. O. K. Reeves, pp. 122-151. Unwin Hyman, London. Lane, S. J., J. C. Alonso and C. A. Martín 2001 Habitat preferences of great bustard Otis tarda flocks in the arable steppes of central Spain: are potentially suitable areas unoccupied? Journal of Applied Ecology 38:193-203. Latham, J., B. W. Staines and M. L. Gorman 1999 Comparative feeding ecology of red (Cervus elaphus) and roe deer (Capreolus capreolus) in Scottish plantation forests. Journal of Zoology, London 247:409-418. Lawson, I. T., S. Al-Omari, P. C. Tzedakis, C. Bryant and K. Christaniss 2005 Lateglacial and Holocene vegetation history at Nisi Fen and the Boras mountains, northern Greece. The Holocene 15:873-887. Lawson, I. T., M. R. Frogley, C. Bryant and R. C. Preece 2004 The Lateglacial and Holocene environmental history of the Ioannina basin, north-west Greece. Quaternary Science Reviews 23:1599-1625. Lax, E. 1995 Quaternary Faunal Remains from the Cave Site of Apidima (Laconia, Greece). Acta Anthropologica (Athens) 1:127-156. Lee, R. 1979 The !Kung San: Men, Women, and Work in a Foraging Society. Cambridge University Press, Cambridge. Leonard, R. D. 401 1989 Anasazi Faunal Exploitation: Prehistoric Subsistence on Northern Black Mesa, Arizona. Center for Archaeological Investigations Occasional Paper No. 13. Southern Illinois University, Carbondale. Levins, R. 1968 Evolution in Changing Environments: Some Theoretical Explorations. Princeton University Press, Princeton, NJ. Ligoni, E. and M. Papagrigorakis 1995 Odontological examination of skeleton LAO I/S 3. Acta Anthropologica (Athens) 1:57-58. Lloveras, L., M. Moreno-Garcia and J. Nadal 2008a The Eagle Owl (Bubo bubo) as a Leporid Remains Accumulator: Taphonomic Analysis of Modern Rabbit Remains Recovered from Nests of this Predator. International Journal of Osteoarchaeology 19(5):573-592. 2008b Taphonomic analysis of leporid remains obtained from modern Iberian lynx (Lynx pardinus) scats. Journal of Archaeological Science 35(1):1-13. 2008c Taphonomic study of leporid remains accumulated by the Spanish Imperial Eagle (Aquila adalberti). Geobios 41(1):91-100. Lowe, V. P. 1967 Teeth as indicators of age, with special reference to red deer (Cervus elaphus) of known age from Rhum. Journal of Zoology, London 152:137-153. Lupo, K. 1998 Experimentally Derived Extraction Rates for Marrow: Implication for Body Part Exploitation Strategies of Plio-Pleistocene Hominid Scavengers. Journal of Archaeological Science 25:657-675. 2007 Evolutionary Foraging Models in Zooarchaeological Analysis: Recent Applications and Future Challenges. Journal of Archaeological Research 15:143-189. Lupo, K. and D. M. Schmitt 1997 Experiments in bone boiling: Nutritional returns and archaeological reflections. Anthropozoologica 25-26:137-144. 2002 Upper Paleolithic net-hunting, small prey exploitation and women's work effort: A view from the ethnographic and ethnoarchaeological record of the Congo Basin. Journal of Archaeological Method and Theory 9:147-179. 402 2005 Small prey hunting technology and zooarchaeological measures of taxonomic diversity and abundance: Ethnoarchaeological evidence from Central African forest foragers. Journal of Anthropological Archaeology 24:335-353. Lupo, K. D. 2006 What Explains the Carcass Field Processing and Transport Decisions of Contemporary Hunter-Gatherers? Measures of Economic Anatomy and Zooarchaeological Skeletal Part Representaton. Journal of Archaeological Method and Theory 13(1):19-66. Lyman, R. L. 1982 The Taphonomy of Vertebrate Archaeofaunas: Bone Density and Differential Survivorship of Fossil Classes. Ph.D. Dissertation, University of Washington. 1984 Bone Density and Differential Survivorship of Fossil Classes. Journal of Anthropological Archaeology 3:259-299. 1985 Bone Frequencies: Differential Transport, In Situ Destruction, and the MGUI. Journal of Archaeological Science 12:221-236. 1992 Anatomical Considerations of Utility Curves in Zooarchaeology. Journal of Archaeological Science 19:7-22. 1994 Vertebrate Taphonomy. Cambridge Manuals in Archaeology. Cambridge University Press, Cambridge. 2008 Quantitative Paleozoology. Cambridge Manuals in Archaeology. Cambridge University Press, New York. Lyman, R. L. and G. L. Fox 1989 A Critical Evaluation of Bone Weathering as an Indication of Bone Assemblage Formation. Journal of Archaeological Science 16:293-317. Lyritzis, Y. and Y. Maniatis 1989 ESR experiments on Quaternary calcites and bones for dating purposes. Journal of Radioanalytical Nuclear Chemistry (Budapest) 129:3-21. 1995 ESR experiments on calcites and bones for dating purposes. Acta Anthropologica (Athens) 1:65-92. MacArthur, R. H. and E. R. Pianka 1966 On optimal use of a patchy environment. American Naturalist 100:603-609. Macklin, M. G., I. C. Fuller, J. Lewin, G. S. Maas, D. G. Passmore, J. Rose, J. C. Woodward, S. Black, R. H. B. Hamlin and J. S. Rowan 403 2002 Correlation of Fluvial Sequences in the Mediterranean Basin of the Last 200 ka and their Relationship to Climate Change. Quaternary Science Reviews 21:16331641. Madrigal, T. C. and J. Z. Holt 2002 White-Tailed Deer Meat and Marrow Return Rates and their Application to Eastern Woodlands Archaeology. American Antiquity 67(4):745-759. Madsen, D. B. and J. E. Kirkman 1988 Hunting Hoppers. American Antiquity 53(3):593-604. Madsen, D. B. and D. N. Schmitt 1998 Mass Collecting and the Diet Breadth Model: a Great Basin Example. Journal of Archaeological Science 25(445-455). Magnell, O. 2006 Tooth Wear in Wild Boar (Sus scrofa). In Recent Advances in Ageing and Sexing Animal Bones, edited by D. Ruscillo, pp. 189-203. Oxbow Books, Oxford. Maguire, J. M., D. Pemberton and M. H. Collett 1980 The Makapansgat Limeworks Grey Breccia: hominids, hyaenas, hystricids or hillwash? Paleontologia Africana 23:75-98. Manne, T. and N. F. Bicho 2009 Val Boi: rendering new understandings of resource intensification and diversification in southwestern Iberia. Before Farming [online version] 2(1):1-21. Manne, T., M. C. Stiner and N. F. Bicho 2005 Evidence for Bone Grease Rendering during the Upper Paleolithic at Vale Boi (Algarve, Portugal). In Proceedings of the IV Congresso de Arqueologia Peninsular, Session 4, edited by N. F. Bicho. Centro de Estudos, Faro. Manolis, S. K., L. C. Aiello, R. Henessy and N. Kyparissi-Apostolika 2000 The Middle Palaeolithic Footprints from Theopetra Cave (Thessaly, Greece). In Theopetra Cave: Twelve years of excavation of research 1987-1998: Proceedings of the International Conference, Trikala, 6-7 November 1998, edited by N. KyparissiApostolika, pp. 87-94. Institute for Aegean Prehistory, Athens. Marean, C. W. and S. Y. Kim 1998 Mousterian large-mammal remains from Kobeh Cave: Behavioral implications. Current Anthropology 39:645-658. Marean, C. W. and L. M. Spencer 1991 Impact of Carnivore Ravaging on Zooarchaeological Measures of Element Abundance. American Antiquity 56(4):645-658. 404 Marean, C. W., L. M. Spencer, R. J. Blumenschine and S. D. Capaldo 1992 Captive Hyaena Bone Choice and Destruction, the Schlepp Effect and Olduvai Archaeofaunas. Journal of Archaeological Science 19:101-121. Marín Arroyo, A. B. 2009 The use of optimal foraging theory to estimate Late Glacial site catchment areas from a central place: The case of eastern Cantabria, Spain. Journal of Anthropological Archaeology 28:27-36. Marshall, L. G. 1989 Bone modification and "the laws of burial". In Bone Modifications, edited by R. Bonnichsen and M. H. Sorg, pp. 7-24. University of Maine Center for the Study of the First Americans, Orno. Martin, J. F. 1983 Optimal foraging theory: A review of some models and their applications. American Anthropologist 85:612-629. Martinson, D. G., N. G. Pisias, J. D. Hays, J. Imbrie, T. C. Moore Jr. and N. J. Shackleton 1987 Age Dating and the Orbital Theory of the Ice Ages: Development of a HighResolution 0 to 300,000-Year Chronostratigraphy. Quaternary International 27:1-29. McBrearty, S. and A. S. Brooks 2000 The revolution that wasn't: a new interpretation of the origin of modern human behavior. Journal of Human Evolution 39:453-563. Meignen, L., O. Bar-Yosef, P. Goldberg and S. Weiner 2001 Le Feu au Paléolithique Moyen: Recherches sur les Structures de Combustion et le Statut des Foyers. L'Exemple du Proche-Orient. Paléorient 26(2):9-22. Mellars, P. 1973 The Character of the Middle-Upper Paleolithic Transition in Southwest France. In The Explanation of Culture Change, edited by A. C. Renfrew. Duckworth, London. 1989 Major Issues in the Emergence of Modern Humans. Current Anthropology 30:349-385. 1991 Cognitive Changes and the Emergence of Modern Humans in Europe. Cambridge Archaeological Journal 1(1):63-76. 1996 The Neanderthal Legacy. Princeton University Press, Princeton, NJ. 1999 The Neandertal Problem: Continued. Current Anthropology 40(3):352-355. 405 2004 Reindeer specialization in the early Upper Palaeolithic: the evidence from south west France. Journal of Archaeological Science 31:613-617. Mentzer, S. M. 2009 Bone as a Fuel Source: The Effects of Initial Fragment Size Distribution. In Fuel Management during the Palaeolithic and Mesolithic Periods: New tools, new interpretations, edited by I. Théry-Parisot, S. Costamagno and A. Henry. BAR International Series 1914, Oxford. Meserve, P. L. and W. E. Glanz 1978 Geographical ecology of small mammals in the northern Chilean arid zone. Journal of Biogeographcy 5:135-148. Meshveliani, T., O. Bar-Yosef and A. Belfer-Cohen 2004 The Upper Paleolithic of eastern Georgia. In The Early Upper Paleolithic Beyond Western Europe, edited by P. J. Brantingham, S. L. Kuhn and K. W. Kerry, pp. 129-143. University of California Press, Berkeley. Metcalfe, D. and K. R. Barlow 1992 A Model for Exploring the Optimal Trade-off between Field Processing and Transport. American Anthropologist 94(2):340-356. Metcalfe, D. and K. T. Jones 1988 A Reconsideration of Animal Body-Part Utility Indices. American Antiquity 53(3):486-504. Milojčić, V., J. Boessneck, O. Jung and H. Schneider 1965 Paläolithikum um Larissa in Thessalien, Bonn. Miracle, P. 2005 Late Mousterian Subsistence and Cave Use in Dalmatia: the Zooarchaeology of Mujina Pecina, Croatia. International Journal of Osteoarchaeology 15:84-105. Mitchell-Jones, A. J., G. Amori, Bogdanowicz, B. Kryštufek, P. J. H. Reijnders, F. Spitzenberger, M. Stubbe, J. M. B. Thissen, V. Vohralík and J. Zima 1999 The Atlas of European Mammals. Academic Press, London and San Diego. Morales, M. B., J. C. Alonso and J. Alonso 2002 Annual productivity and individual female reproductive success in a Great Bustard Otis tarda population. Ibis 144:293-300. Morlan, R. E. 1994 Bison bone fragmentation and survivorship: a comparative method. Journal of Archaeological Science 21:797-807. 406 Moss, E. H. 1997 Lithic Use-Wear Analysis. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 1: Excavation and intra-site analysis at Klithi, edited by G. N. Bailey, pp. 193-206. McDonald Institute for Archaeological Research, Cambridge. Moundrea-Agrafioti, A. 2003 Mesolithic fish hooks from the Cave of Cyclope, Youra. In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 131-141. British School at Athens Studies 10, London. Munro, N. 1999 Small Game as Indicators of Sedentization During the Natufian Period at Hayonim Cave in Israel. In Zooarchaeology of the Pleistocene/Holocene Boundary, edited by J. C. Driver, pp. 37-45. British Archaeological Reports. 2001 A Prelude to Agriculture: Game Use and Occupation Intensity During the Natufian Period in the Southern Levant. Unpublished Ph.D. Dissertation, University of Arizona. 2004 Zooarchaeological Measures of Hunting Pressure and Occupation Intensity in the Natufian. Current Anthropology 45:S5-S33. 2009a Epipaleolithic Subsistence Intensification in the Southern Levant: The Faunal Evidence. In The Evolution of Hominin Diets: Integrating Approaches to the Study of Palaeolithic Subsistence, edited by J.-J. Hublin and M. P. Richards, pp. 141-155. Springer Science + Business Media B.V. 2009b Integrating inter- & intra-site analyses of Epipalaeolithic faunal assemblages from Israel. Before Farming 1(4):1-18. Munro, N. and G. Bar-Oz 2005 Gazelle bone fat processing in the Levantine Epipaleolithic. Journal of Archaeological Science 32:223-239. Mylona, D. 2003 The exploitation of fish resources in the Mesolithic Sporades: fish remains from the Cave of Cyclope, Youra. In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 181-188. British School at Athens Studies 10, London. Nagaoka, L. 407 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. 2002a The effects of resource depression on foraging efficiency, diet breadth, and patch use in southern New Zealand. Journal of Anthropological Archaeology 21:419442. 2002b Explaining Subsistence Change in Southern New Zealand Using Foraging Theory Models. World Archaeology 34:84-102. 2005 Declining foraging efficiency and moa carcass exploitation in southern New Zealand. Journal of Archaeological Science 32:1328-1338. Nakazawa, Y., L. G. Straus, M. R. Gonzales-Morales, D. C. Solana and J. C. Saiz 2009 On stone-boiling technology in the Upper Paleolithic: behavioral implications from an Early Magdalenian hearth in El Miron Cave, Cantabria, Spain. Journal of Archaeological Science 36:684-693. Newton, S. 1999 Theopetra Cave and the Palaeolithic-Mesolithic transition in southern Europe. In Zooarchaeology in Greece: Recent Advances, edited by E. Kotjabopoulou, Y. Hamilakis, P. Halstead and C. Gamble, pp. 115-122. Oxbow Books, Cambridge. Nielsen-Marsh, C. and R. E. M. Hedges 2000 Patterns of Diagenesis in Bone I: The Effects of Site Environments. Journal of Archaeological Science 27:1139-1150. Niven, L. 2007 From carcass to cave: Large mammal exploitation during the Aurignacian at Vogelherd, Germany. Journal of Human Evolution 53:362-382. Noddle, B. A. 1974 Ages of Epiphyseal Closure in Feral and Domestic Goats and Ages of Dental Eruption. Journal of Archaeological Science 1(2):195-204. Nowack, R. M. 1999 Walker's Mammals of the World. 6th ed. The Johns Hopkins University Press, Baltimore. Nowell, G. M. and M. S. A. Horstwood 2009 Comments on Richards et al., Journal of Archaeological Science 35, 2008 "Strontium isotope evidence of Neanderthal mobility at the site of Lakonis, Greece using laser-ablation PIMMS". Journal of Archaeological Science 36:1334-1341. 408 Ntinou, M. 2010 Wood Charcoal Analysis at Klissoura Cave 1 (Prosymna, Peloponese): The Upper Palaeolithic Vegetation. Eurasian Prehistory 7(2):47-69. O'Connell, J. F. and K. Hawkes 1984 Food Choice and Foraging Sites among the Alyawara. Journal of Anthropological Research 40(4):504-535. O'Connell, J. F., K. Hawkes and N. Blurton Jones 1988 Hadza Hunting, Butchering, and Bone Transport and Their Archaeological Implications. Journal of Anthropological Research 44:113-161. 1990 Reanalysis of Large Mammal Body Part Transport Among the Hadza. Journal of Archaeological Science 17:301-316. O'Connell, J. F. and B. Marshall 1989 Analysis of Kangaroo Body Part Transport among the Alyawara of Central Australia. Journal of Archaeological Science 16:393-405. Okuda, M., Y. Yasuda and T. Setoguchi 1999 Latest Pleistocene and Holocene pollen records from Lake Kopais, Southeast Greece. Journal of the Geological Society of Japan 105(6):450-455. Orians, G. H. and N. E. Pearson 1979 On the Theory of Central Place Foraging. In Analysis of Ecological Systems, edited by D. J. Horn, R. D. Mitchell and G. R. Stairs, pp. 154-177. The Ohio State University Press, Columbus. Orlando, L., M. Mashkour, A. Burke, C. J. Douady, V. Eisenmann and C. Hänni 2006 Geographic distribution of an extinct equid (Equus hydruntinus: Mammalia, Equidae) revealed by morphological and genetic analysis of fossils. Molecular Ecology 15:2083-2093. Outram, A. K. 2001 A New Approach to Identifying Bone Marrow and Grease Exploitation: Why the "Indeterminate" Fragments should not be Ignored. Journal of Archaeological Science 28:401-410. Ovchinnikov, I. V., A. Gotherstrom, G. P. Romanova, V. M. Kharitonov, K. Liden and W. Goodwin 2000 Molecular analysis of Neanderthal DNA from the northern Caucasus. Nature 404:490-493. Palma di Cesnola, A. 409 1966 Il Paleolitico superiore arcaico (facies uluzziana) della Grotta del Cavallo, Lecce (continuazione). Rivista di Scienze Prehistoriche 21:3-59. Panagopoulou, E. 2000 The Middle Palaeolithic Assemblages of Theopetra Cave: Technological Evolution in the Upper Pleistocene. In Theopetra Cave: Twelve years of excavation and research 1987-1998, edited by N. Kyparissi-Apostolika, pp. 139-162. The Institute for Aegean Prehistory, Athens (in Greek). Panagopoulou, E., P. Karkanas, G. Tsartsidou, E. Kotjabopoulou, K. Harvati and M. Ntinou 2002-2004 Late Pleistocene Archaeological and Fossil Human Evidence from Lakonis Cave, Southern Greece. Journal of Field Archaeology 29:323-349. Papaconstantinou, V. and D. Vassilopoulou 1997 The Middle Palaeolithic Industries of Epirus. In Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, edited by G. N. Bailey, pp. 459-480. vol. 2. McDonald Institute for Archaeological Research, Cambridge. Papagianni, D. 2000 Middle Palaeolithic Occupation and Technology in Northwestern Greece. The Evidence from Open-Air Sites. BAR International Series 882, Oxford. Pavao, B. and P. Stahl 1999 Structural Density Assays of Leporid Skeletal Elements with Implications for Taphonomic, Actualistic and Archaeological Research. Journal of Archaeological Science 26:53-66. Pawlikowski, M., M. Koumouzelis, B. Ginter and J. K. Kozlowski 2000 Emerging Ceramic Technology in Structured Aurignacian Hearths at Klisoura Cave 1 in Greece. Archaeology, Ethnology and Anthropology of Eurasia 4:19-29. Payne, S. 1975 Faunal Change at Franchthi Cave from 20,000 B.C. to 3,000 B.C. In Archaeozoological Studies: Papers of the Archaeozoological Conference 1974, Held at the Biologisch-Archaeologisch Instituut of the State University of Groningen, edited by A. T. Clason, pp. 120-131. American Elsiver Publishing Company, New York. 1982 Faunal evidence for environmental/climatic change at Franchthi Cave (Southern Argolid, Greece), 25,000 B.P. - 5000 B.P. Preliminary Results. In Palaeoclimates, Palaeoenvironments and Human Communities in the Eastern Mediterranean Region in Later Prehistory, edited by J. L. Bintliff and W. van Zeist, pp. 133-137. British Archaeological Reports, Oxford. 410 Peresani, M., M. Cremashchi, F. Faerraro, C. Falguères, J. J. Bahain, G. Gruppioni, E. Sibilia, G. Quarta, L. Calcagnile and J. M. Dolo 2008 Age of the final Middle Palaeolithic and Uluzzian levels at Fumane Cave, Northern Italy, using 14C, ESR, 234U/230Th and thermoluminescence methods. Journal of Archaeological Science 35:2986-2996. Perlès, C. 1987 Les industries lithiques taillées de Franchthi (Argolide, Grèce). Tome I: Présentation générale et industries paléolithiques. Excavations at Franchthi Cave Fascicle 3. Indiana University Press, Bloomington and Indianapolis. 1991 Les Industries lithiques taillées de Franchthi (Argolide, Grèce). Excavations at Franchthi Cave, Fascicle 5. Indiana University Press, Bloomington and Indianapolis. 1999 Long-term perspectives on the occupation of the Franchthi Cave: continuity and discontinuity. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 311-318. British School at Athens Studies 3, London. Perlès, C. and M. Vanhaeren 2010 Black Cyclope neritea Marine Shell Ornaments in the Upper Palaeolithic and Mesolithic of Franchthi Cave, Greece: Arguments for Intentional Heat Treatment. Journal of Field Archaeology 35(3):298-309. Petit-Maire, N., B. Vrielinck, J.-P. Bracco, J.-P. Brugal, P.-F. Burollet, G. Coude-Gaussen, G. Jalut, G. Lericolais and v. Vliet-Lanoë 2005 The Mediterranean Basin: The Last Two Climatic Extremes: Explanatory Notes of the Maps. Maison Méditerranéenne des Sciences de l’Homme and Agence Nationale pour le Gestion des Déchets Radioactifs Parc de la Croix Blanche, Groupe Horizon, Gémenos, France. Phoca-Cosmetatou, N. 2003a Ibex Exploitation: the Case of Klithi or the Case of the Upper Palaeolithic? In Zooarchaeology in Greece: Recent Advances, edited by E. Kotjabopoulou, Y. Hamilakis, P. Halstead, C. Gamble and P. Elefanti, pp. 161-173. vol. Studies 9. British School at Athens. 2003b Subsistence Changes During the Late Glacial? The Example of Ibex Exploitation in Southern Europe. In Le rôle de l'environment dans les comportements des chasseurs-cueilleurs préhistoriques, edited by M. Patou-Mathis and H. Bocherens, pp. 39-54. BAR International Series 1105, Oxford. 411 2004a Site Function and the 'Ibex-Site Phenomenon': Myth or Reality? Oxford Journal of Archaeology 23(3):217-242. 2004b A Zooarchaeological Reassessment of the Habitat and Ecology of the Ibex (Capra ibex). In The Future from the Past: Archaeozoology in Wildlife Conservation and Heritage Management, edited by R. C. G. M. Lauwerier and I. Plug, pp. 64-78. Oxbow Books. 2005 Bone Weathering and Food Procurement Strategies: assessing the relaibility of our behavioural inferences. In Biosphere to Lithosphere: New studies in vertebrate taphonomy, edited by T. O'Connor. Oxbow Books, Oxford. 2009 Specialization & diversification: a tale of two subsistence strategies from Late Glacial Italy. Before Farming 3(2):1-29. Pianka, E. R. 2000 Evolutionary Ecology. 6th ed. Addison Wesley Education Publishers, Inc. (Benjamin/Cummings), San Francisco. Pigati, J., J. Quade, J. B. Wilson, A. J. T. Jull and N. A. Lifton 2007 Development of low-background vacuum extraction and graphitization system for 14C dating of old (40-60 ka) samples. Quaternary International 166:4-14. Pike-Tay, A., V. Cabrera Valdes and F. Bernaldo de Quiros 1999 Seasonal variations of the Middle-Upper Paleolithic transition at El Castillo, Cueva Morin and El Pendo (Cantabria, Spain). Journal of Human Evolution 36:283317. Pitsios, T. K. 1985 Paleoanthropological research at the site of Apidima. Archaeologia 15:26-33 (in Greek). 1995 Prologue. Proceedings of the Meeting on Paleoanthropological Research in Apidima, Athens 18.4.1989. Acta Anthropologica (Athens) 1:9-18. 1999 Paleoanthropological research at the cave site of Apidima and the surrounding region (South Peloponnese, Greece). Anthropologischer Anzeiger 57(1):1-11. Potter, J. M. 1995 The effects of sedentism on the processing of hunting carcasses in the southwest: a comparison of two Pueblo IV sites in central New Mexico. Kiva 60:411428. Powell, J. 412 2003 The fish bone assemblage from the Cave of Cyclope, Youra: evidence for continuity and change. In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 173-179. British School at Athens Studies 10, London. Pyke, G. 1984 Optimal foraging theory: A critical review. Annual Review of Ecology and Systematic 15:59-74. Reisch, L. 1976 Beobachtungen an Vogelknochen aus dem Spätpleistozön der Höhle von Kephalari (Argolis, Grieschenlands). Archäologisches Korrespondenzblatt 6:261-265. Reitz, E. J. and D. Cordier 1983 Use of Allometry in Zooarchaeological Analysis. In Animals and Archaeology: 2. Shell Middens, Fishes and Birds, edited by C. Grigson and J. Clutton-Brock, pp. 237-252. British Archaeological Reports International Series 183, Oxford. Reitz, E. J., I. R. Quitmyer, H. S. Hale, S. J. Scudder and E. S. Wing 1987 Application of Allometry to Zooarchaeology. American Antiquity 52:304-317. Reitz, E. J. and E. S. Wing 2008 Zooarchaeology. 2nd ed. Cambridge Manuals in Archaeology. Cambridge University Press, Cambridge. Rensberger, J. M. and H. B. Krentz 1988 Microscopic effects of predator digestion on the surfaces of bones and teeth. Scanning Microscopy 2:1541-1551. Reynaud Savioz, N. and P. Morel 2005 La faune de Nadaouiyeh Aïn Askar (Syrie centrale, Pléistocène moyen): aperçu et perspectives. Revue de Paléobiologie, Genève 10:31-35. Richards, M. P., K. Harvati, V. Grimes, C. Smith, T. Smith, J.-J. Hublin, P. Karkanas and E. Panagopoulou 2008 Strontium isotope evidence of Neanderthal mobility at the site of Lakonis using laer-ablation PIMMS. Journal of Archaeological Science 35(5):1251-1256. Riel-Salvatore, J. 2009 What is a 'Transitional' Industry? THe Uluzzian of Southern Italy as a Case Study. In Sourcebook of Paleolithic Transitions, edited by M. Camps and P. Chauhan, pp. 377-396. Springer Science+Business Media, LLC. Roberts, T. J. 413 1977 The Mammals of Pakistan. Ernest Benn Limited, London. Roger, T. and A. Darlas 2008a Microvertébrés, Paléo-Environmental et Paléoclimat de la Grotte de Kalamakia (Péloponnèse, Grèce). In The Palaeolithic of the Balkans, edited by A. Darlas and D. Milhailovic, pp. 77-84. vol. 1819. BAR International Series, Oxford. 2008b Upper-Pleistocene Bird Remains from Kalamakia Cave (Greece). In The Palaeolithic of the Balkans, edited by A. Darlas and D. Milhailovic, pp. 69-76. vol. 1819. BAR International Series, Oxford. Rogers, A. R. 2000 On the Value of Soft Bones in Faunal Analysis. Journal of Archaeological Science 27(7):635-639. Rose, M. 1995 Fishing at Franchthi Cave, Greece: Changing environments and patterns of exploitation. Old World Archaeology Newsletter 18:21-26. Roubet, C. 1997 The Lithic Domain at Klithi: Technology of Production and the Chaîne Opératoire. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 1: Excavation and intra-site analysis at Klithi, edited by G. N. Bailey, pp. 125-154. McDonald Institute for Archaeological Research, Cambridge. 1999 Expressions of an Upper Palaeolithic management of lithic resources at Klithi (Greece). In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, September 1994, edited by G. N. Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 170-187. British School at Athens Studies 3, London. Roubet, C. and M. Lenoir 1997 The Backed Pieces at Klithi. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol 1: Excavation and intra-site analysis at Klithi, edited by G. N. Bailey, pp. 155-180. McDonald Institute for Archaeological Research, Cambridge. Runnels, C. 1995a The Palaeolithic and Mesolithic Remains. In The Berbati-Limnes Archaeological Survey 1998-1990, edited by B. Wells and C. Runnels, pp. 23-35. Svenska Institutet Athen, Stockholm. 1995b Review of Aegean Prehistory IV: the Stone Age of Greece from the Palaeolithic to the Advent of the Neolithic. American Journal of Archaeology 99(4):699-728. 414 2001 Review of Aegean Prehistory IV: The Stone Age of Greece from the Palaeolithic to the Advent of the Neolithic, with Addendum: 1995-1999. In Aegean prehistory: a review, edited by T. Cullen, pp. 225-258. Archaeological Institute of America, Boston. Runnels, C., E. Karimali and B. Cullen 2003 Early Upper Palaeolithic Spilaion: an Artifact-Rich Surface Site. Hesperia Supplements, Landscape Archaeology in Southern Epirus, Greece 1 32:135-156. Runnels, C. and T. H. van Andel 1993 The Lower and Middle Palaeolithic of Thessaly, Greece. Journal of Field Archaeology 20:299-317. Saint-Germain, C. 1997 The production of bone broth: a study in nutritional exploitation. Anthropozoologica 25-26:153-156. Sampson, A. 1998 The Neolithic and Mesolithic Occupation of the Cave of Cyclope, Youra, Alonnessos, Greece. The Annual of the British School at Athens 93:1-22. Sampson, A., J. K. Kozlowski and M. Kaczanowska 2003 Mesolithic chipped stone industries from the Cave of Cyclope on the island of Youra (northern Sporades). In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 123-130. British School at Athens Studies 10, London. Sampson, A., J. K. Kozlowski, M. Kaczanowska, A. Budek, A. Nadachowski, T. Tomek and B. Miekina 2009 Sarakenos Cave in Boeotia, from Palaeolithic to the Early Bronze Age. Eurasian Prehistory 6(1-2):199-231. Sampson, C. G. 2000 Taphonomy of Tortoises Deposited by Birds and Bushmen. Journal of Archaeological Science 27:779-788. Sánchez Goñi, M. F., I. Cacho, J. L. Turon, J. Guiot, F. J. Sierro, J. P. Peypouquet, J. O. Grimalt and N. J. Shackleton 2002 Synchroneity between marine and terrestrial responses to millennial scale climatic variability during the last glacial period in the Mediterranean region. Climate Dynamics 19:95-105. Sanchis Sera, A. and J. Fernández Peris 415 2008 Procesado y consumo antrópico de conejo en la Cova del Bolomor (Tavernes de la Valldigna, Valencia). Complutum 18(1):25-46. Schaller, G. B. 1977 Mountain Monarchs, Wild Sheep and Goats of the Himalaya. University of Chicago Press, Chicago. Schiegl, S., P. Goldberg, O. Bar-Yosef and S. Weiner 1996 Ash deposits in Hayonim and Kebara caves, Israel: macroscopic, microscopic and mineralogical observations, and their archaeological implications. Journal of Archaeological Science 23:763-781. Schmid, E. 1972 Atlas of Animal Bones for Prehistorians, Archaeologists, and Quaternary Geologists. Elsevier Science Publishers, Amsterdam. Schmitt, D. M., D. B. Madsen and K. Lupo 2004 The worst of times, the best of times: Jackrabbit hunting by Middle Holocene human foragers in the Bonneville Basin of western North America. In Colonisation, Migration and Marginal Areas: A Zooarchaeological Approach, edited by M. Mondini, S. Muñoz and S. Wickler, pp. 86-95. Oxbow Books, Oxford. Schmitt, D. N. and K. Lupo 1995 On Mammalian Taphonomy, Taxonomic Diversity, and Measuring Subsistence Data in Zooarchaeology. American Antiquity 60(3):496-514. Schmitz, R. W., D. Serre, G. Bonani, S. Feine, F. Hillgruber, H. Krainitzki, S. Paabo and F. H. Smith 2002 The Neandertal type site revisited: interdisciplinary investogations of skeletal remains from the Neander Valley, Germany. PNAS 99:133342-133347. Schoener, T. W. 1979 Generality of the Size-Distance Relation in Models of Optimal Foraging. The American Naturalist 114(6):902-914. Searle, K. R. and L. A. Shipley 2008 The Comparative Feeding Behaviour of Large Browsing and Grazing Herbivores. In The Ecology of Browsing and Grazing, edited by I. J. Gordon and H. H. T. Prins, pp. 117-148. Ecological Studies 195. Springer. Serre, D., A. Langaney, M. Chech, M. Teschler-Nicola, M. Paunoviae, P. Mennecier, M. Hofreiter, G. Possnert and S. Paabo 2004 No evidence of Neandertal mtDNA contribution to early modern humans. PLoS Biology 2:313-317. 416 Severinghaus, C. W. 1949 Tooth Development and Wear as Criteria of Age in White-Tailed Deer. Journal of Wildlife Management 13:195-216. Shackleton, J. C. 1990 Marine Molluscan Remains from Franchthi Cave. Excavations at Franchthi Cave, Greece, Fascicle 4. Indiana University Press, Bloomington and Indianapolis. Sharp, N. D. 1990 Fremont and Anasazi Resource Selection: An Examination of Faunal Assemblage Variation in the Northern Southwest. Kiva 56:45-65. Shipman, P., W. Bosler and K. L. Davis 1981 Butchering of giant geladas at an Acheulian site. Current Anthropology 22:257-268. Silva, M. and J. A. Downing 1995 CRC Handbook of Mammalian Body Masses. CRC Press, Boca Raton, Florida. Silver, I. A. 1969 The ageing of domestic animals. In Science in archaeology: a survey of progress and research, edited by D. Brothwell and E. S. Higgs, pp. 283-302. Praeger, New York. Simms, S. R. 1987 Behavioral Ecology and Hunter-Gatherer Foraging: An Example from the Great Basin. BAR International Reports 381, Oxford. Simpson, E. H. 1949 Measurement of Diversity. Nature 163:688. Sitlivy, V., K. Sobczyk, P. Karkanas and M. Koumouzelis 2007 Middle Paleolithic Lithic Assemblages of the Klissoura Cave, Peloponnesus, Greece: a Comparative Analysis. Archaeology, Ethnology and Anthropology of Eurasia 3:2-15. Smith, B. and J. B. Wilson 1996 A Consumer's Guide to Evenness Indices. Oikos 76:70-82. Smith, E. A. 1991 Inujjuamiut Foraging Strategies: Evolutionary Ecology of an Arctic Hunting Economy. Aldine de Gruyter, New York. 417 2000 Three styles in the evolutionary analysis of human behavior. In Adaptations and Human Behavior: An Anthropological Perspective, edited by L. Cronk, N. Chagnon and W. Irons, pp. 27-46. Aldine de Gruyter, New York. 2004 Why do good hunters have higher reproductive success? Human Nature 15:343-364. Smith, E. A. and B. Winterhalder 1985 On the logica and application of optimal foraging theory: A brief reply to Martin. American Anthropologist 87(3):645-648. Sordinas, A. 1969 Investigations in the prehistory of Corfu during 1964-1966. Balkan Studies 10(2):393-424. 2003 The 'Sidarian': maritime Mesolithic non-geometric microliths in western Greece. In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 89-97. British School at Athens, Studies 10, London. Sosis, R. 2002 Patch Choice Decisions among Ifaluk Fishers. American Anthropologist 104(2):583-598. Speth, J. D. 1983 Bison Kills and Bone Counts. University of Chicago Press, Chicao. 1991 Some unexplored aspects of mutualistic Plains-Pueblo food exchange. In Farmers, Hunters, and Colonists: Interaction Between the Southwest and Southern Plains, edited by K. Spielmann, pp. 18-35. University of Arizona Press, Tucson. 2004 Hunting Pressure, Subsistence Intensification, and Demographic Change in the Levantine Late Middle Paleolithic. In Human Paleoecology in the Levantine Corridor, edited by N. Goren-Inbar and J. D. Speth. Oxbow Books, Oxford. Speth, J. D. and J. L. Clark 2006 Hunting and overhunting in the Levantine Late Middle Paleolithic. Before Farming 3(1):1-42. Speth, J. D. and S. L. Scott 1989 Horticulture and large-mammal hunting: The role of resource depletion and the constraints of time and labor. In Farmers as Hunters: The Implications of Sedentism, edited by S. Kent, pp. 71-77. Cambridge University Press, Cambridge. Speth, J. D. and E. Tchernov 418 1998 The Role of Hunting and Scavenging in Neandertal Procurement Strategies: New Evidence from Kebara Cave (Israel). In Neandertals and Modern Humans in Western Asia, edited by T. Akazawa, K. Aoki and O. Bar-Yosef, pp. 223-239. Plenum Press, New York and London. 2001 Neandertal hunting and meat-processing in the Near East: evidence from Kebara Cave (Israel). In Meat-Eating and Human Evolution, edited by C. B. Stanford and H. T. Bunn, pp. 52-72. Oxford University Press, Oxford. 2002 Middle Paleolithic tortoise use at Kebara Cave (Israel). Journal of Archaeological Science 29(5):471-483. Spiess, A. E. 1979 Reindeer and Caribou Hunters, an Archaeological Study. Academic Press, New York. Starkovich, B. M. 2009 Dietary changes during the Upper Palaeolithic at Klissoura Cave 1 (Prosymni), Peloponnese, Greece. Before Farming [online version] 3(4):1-14. Starkovich, B. M. and M. C. Stiner 2010 Upper Palaeolithic Animal Exploitation at Klissoura Cave 1 in Southern Greece: Dietary Trends and Mammal Taphonomy. Eurasian Prehistory 7(2):107-132. Stephens, D. W. and J. R. Krebs 1986 Foraging Theory. Monographs in Behavior and Ecology. Princeton University Press, Princeton. Stiner, M. 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 Non-human Predators. Journal of Archaeological Science 18:455-482. 1992 Overlapping Species "Choice" by Italian Upper Pleistocene Predators. Current Anthropology 33(4):433-451. 1994 Honor Among Thieves: a Zooarchaeological Study of Neandertal Ecology. Princeton University Press, Princeton, New Jersey. 2001 Thirty Years on the "Broad Spectrum Revolution" and Paleolithic Demography. PNAS 98(13):6993-6996. 419 2002 On In Situ Attrition and Vertebrate Body Part Profiles. Journal of Archaeological Science 32:103-117. 2003a "Standardization" in Upper Paleolithic Ornaments at the Coastal Sites of Riparo Mochi and Üçagizli Cave. In The Chronology of the Aurignacian and of the Transitional Technocomplexes: Dating, Stratigraphies, Cultural Implications, edited by J. Zilhão and F. d'Errico, pp. 49-59. Instituto Portugues de Arqueologia, Lisbon. 2003b Zooarchaeological evidence for resource intensification in Algarve, Southern Portugal. Promontoria 1(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(3):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. 2009 Prey choice, site occupation intensity & economic diversity in the Middle – early Upper Palaeolithic at the Üçağızlı Caves, Turkey. Before Farming 3(3):1-20. 2010 Shell Ornaments from the Upper Paleolithic through Mesolithic Layers of Klissoura Cave 1 by Prosymna (Peloponese, Greece). Eurasian Prehistory 7(2):287308. Stiner, M. C., O. Bar-Yosef, S. L. Kuhn and S. Weiner 2005 Experiments in Fragmentation and Diagenesis of Bone and Shell. In The Faunas of Hayonim Cave, Israel: a 200,000-Year Record of Paleolithic Diet, Demography, and Society, edited by M. C. Stiner, pp. 39-58. Peabody Museum of Archaeology and Ethnology, Harvard University, Cambridge, Massachusetts. Stiner, M. C., R. Barkai and A. Gopher 2009 Cooperative hunting and meat sharing 400-200 kya at Qesem Cave, Israel. PNAS 106(32):13207-13212. Stiner, M. C., J. K. Kozlowski, S. L. Kuhn, P. Karkanas and M. Koumouzelis 2010 Klissoura Cave 1 and the Upper Paleolithic of Southern Greece in Cultural and Ecological Context. Eurasian Prehistory 7(2):309-321. Stiner, M. C., S. L. Kuhn, T. A. Surovell, P. Goldberg, L. Meignen, S. Weiner and O. BarYosef 2001 Bone Preservation in Hayonim Cave (Israel): a Macroscopic and Mineralogical Study. Journal of Archaeological Science 28:643-659. 420 Stiner, M. C. and N. Munro 2002 Approaches to Prehistoric Diet Breadth, Demography, and Prey Ranking Systems in Time and Space. Journal of Archaeological Method and Theory 9(2):181214. 2011 On the Evolution of Diet and Landscape during the Upper Paleolithic through Mesolithic at Franchthi Cave (Peloponnese Greece). Journal of Human Evolution 60:618-636. Stiner, M. C., N. Munro, T. A. Surovell, E. Tchernov and O. Bar-Yosef 1999 Paleolithic population growth pulses evidenced by small animal exploitation. Science 283:190-194. Stiner, M. C., N. D. Munro and T. A. Surovell 2000 The Tortoise and the Hare: Small-Game Use, the Broad-Spectrum Revolution, and Paleolithic Demography. Current Anthropology 41(1):39-73. Stiner, M. C., C. Pehlevan, M. Sagir and I. Ozer 2002 Zooarchaeological Studies at Üçağızlı Cave: Preliminary Results on Paleolithic Subsistence and Shell Ornaments. Arasterma Sonuclari Toplantisi, Ankara 17:29-36. Stiner, M. C., S. Weiner, O. Bar-Yosef and S. L. Kuhn 1995 Differential burning, fragmentation and preservation of archaeological bone. Journal of Archaeological Science 22:223-237. Strasser, T. F., E. Panagopoulou, C. Runnels, P. M. Murray, N. Thompson, P. Karkanas, F. W. McCoy and K. W. Wegmann 2010 Stone Age Seafaring in the Mediterranean: Evidence from the Plakias Region for Lower Palaeolithic and Mesolithic Habitation of Crete. Hesperia 79:145-190. Straus, L. G. 1977 Of Deerslayers and Mountain Men: Paleolithic Faunal Exploitation in Cantabrian Spain. In For Theory Building in Archaeology, edited by L. R. Binford, pp. 41-76. Academic Press, New York. 1987 Upper Palaeolithic Ibex Hunting in Southwest Europe. Journal of Archaeological Science 14:163-178. Stravopodi, E., S. Manolis and N. Kyparissi-Apostolika 1999 Palaeoanthropological findings from Theopetra cave in Thessaly: a preliminary report. In The Palaeolithic Archaeology of Greece and Adjacent Areas: Proceedings of the ICOPAG Conference, Ioannina, September 1994, edited by G. N. 421 Bailey, E. Adam, E. Panagopoulou, C. Perlès and K. Zachos, pp. 271-281. British School at Athens Studies 3, London. Stringer, C., J. C. Finlayson, R. N. E. Barton, Y. Fernandez-Jalvo, I. Caceres, R. C. Sabin, E. J. Rohodes, A. P. Currant, J. Rodriguez-Vidal, F. Giles-Pacheco and J. A. Riquelme-Cantal 2008 Neanderthal exploitation of marine mammals in Gibraltar. PNAS 105(38):14319-14324. Stringer, C., J.-J. Hublin and B. Vandermeersch 1984 The Origin of Anatomically Modern Humans in Western Europe. In The Origins of Modern Humans: A World Survey of the Fossil Evidence, edited by F. Smith and F. Spencer, pp. 51-135. Alan R. Liss, New York. Stroulia, A. 2010 Flexible Stones: Ground Stone Tools from Franchthi Cave. Excavations at Franchthi Cave, Greece, Fascicle 14. Indiana University Press, Bloomington and Indianapolis. Stutz, A. J., N. Munro and G. Bar-Oz 2009 Increasing the resolution of the Broad Spectrum Revolution in the Southern Levantine Epipaleolithic (19-12 ka). Journal of Human Evolution 56:294-306. Surmely, F., P. Alix, S. Costamagno, P. Daniel, M. Hays, R. Murat, R. Renard, J. Virmont and J. P. Texier 2003 Découverte d'un gisement du Gravettien ancien au Iieu-dit Ie Sire (Mirefleurs, Puy-de-Dôme). Bulletin de la Société préhistorique française 100:29-39. Svoboda, I., H. Van der Plicht and V. Kuzelka 2002 Upper Palaeolithic and Mesolithic human fossils from Moravia and Bohemia (Czech Republic). Antiquity 76:957-962. Swingland, I. R. and I. R. Stubbs 1985 The ecology of a Mediterranean tortoise (Testudo hermanni): Reproduction. Journal of Zoology, London 205:595-610. Symmons, R. 2005 New density data for unfused and fused sheep bones, and a preliminary discussion on the modellingn of taphonomic bias in archaeofaunal age profiles. Journal of Archaeological Science 32:1691-1698. Théry-Parisot, I. 2002 Fuel Management (Bone and Wood) During the Lower Aurignacian in the Pataud Rock Shelter (Lower Paleolithic, Les Eyzies de Tayac, Dordogne, France). Contribution of Experimentation. Journal of Archaeological Science 29:1415-1421. 422 Thomas, D. H. and D. Mayer 1983 Behavioral faunal analysis of selected horizons. In The Archaeology of Monitor Valley 2: Gatecliff Shelter, edited by D. H. Thomas, pp. 353-391. American Museum of Natural History Anthropological Papers 59 (1). Todd, L. C. and D. Rapson 1988 Long bone fragmentation and interpretation of faunal assemblages: approaches to comparative analysis. Journal of Archaeological Science 15(307-325). Tomek, T. and Z. M. Bocheński 2002 Bird Scraps from a Greek Table: The Case of Klisoura Cave. Acta Zoologica Cracoviensia 45:133-138. Tortosa, J. E. A., V. V. Bonilla, M. P. Ripoll, R. M. Valle and P. G. Calatayud 2002 Big Game and Small Prey: Paleolithic and Epipaleolithic Economy from Valencia (Spain). Journal of Archaeological Method and Theory 9(3):215-268. Trantalidou, K. 2003 Faunal remains from the earliest strata of the Cave of Cyclope, Youra. In The Greek Mesolithic: Problems and Perspectives, edited by N. Galanidou and C. Perlès, pp. 143-172. British School at Athens Studies 10, London. Trinkhaus, E. 2007 European Early Modern Humans and the Fate of the Neandertals. PNAS 104(18):7367-7372. Trinkhaus, E., S. Milota, R. Rodrigo, M. Gherase and O. Moldovan 2003 Early modern human cranial remains from the Pestera cu Oase, Romania. Journal of Human Evolution 45:245-253. Trinkhaus, E., O. Moldovan, S. Milota, A. Bilgar, L. Sarcina, S. Athreya, S. E. Bailey, R. Rodrigo, M. Gherase, T. Higham, C. Bronk Ramsey and H. van der Plicht 2003 An early modern human from the Pestera cu Oase, Romania. PNAS 100:11231-11236. Tsoukala, E. 1992 Quaternary Faunas of Greece. In Mammalian Migration and Disperal Events in the European Quaternary, edited by W. v. Koenigswald and L. Werdelin, pp. 7992. vol. 153. Courier Forsch. Inst. Senckenberg, Frankfurt. 1999 Quarternary (sic) large mammals from the Apidima Caves (Lakonia, S Peloponnese, Greece). Beitrage zur Palaeontologie 24:207-229. Turland, N. J., L. Chilton and J. R. Press 423 1993 Flora of the Cretan Area: Annotated Checklist and Atlas. The Natural History Museum, London. Turner, C. and M. F. Sánchez Goñi 1997 Late Glacial Landscape and Vegetation in Epirus. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece, edited by G. N. Bailey, pp. 559-585. vol. 2. Oxbow Books, Oxford. Tutin, T. G., V. H. VHeywood, N. A. Burger, D. M. Moore, D. H. Vallentine, S. M. Walters and D. A. Webb 1980 Flora Europaea Volume 5. Cambridge University Press, Cambridge. Tzedakis, P. C. 1993 Long-term tree populations in northwest Greece through multiple Quaternary climatic cycles. Nature 364:437-440. 1994 Vegetation Change through Glacial-Interglacial Cycles: A Long Pollen Sequence Perspective. Philosophical Transactions: Biological Sciences 345(1314):403-432. 1999 The last climatic cycle at Kopais, central Greece. Journal of the Geological Society, London 156:425-434. Tzedakis, P. C., M. R. Frogley, I. T. Lawson, R. C. Preece, I. Cacho and L. de Abreu 2004 Ecological thresholds and patterns of millennial-scale climate variability: the response of vegetation in Greece during the last glacial period. Geology 32:109-112. Tzedakis, P. C., K. A. Hughen, I. Cacho and K. Harvati 2007 Placing late Neanderthals in a climatic context. Letters to Nature 449:206208. Tzedakis, P. C., I. T. Lawson, M. R. Frogley, G. M. Hewitt and R. C. Preece 2002 Buffered Tree Population Changes in a Quaternary Refugium: Evolutionary Implications. Science 297:2044-2047. Ugan, A. 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(1):75-89. Valensi, P. 424 2000 The Archaeozoology of Lazaret Cave (Nice, France). International Journal of Osteoarchaeology 10:357-367. Valladas, H., N. Mercier, L. Froget, J.-L. Joron, J.-L. Reyss, P. Karkanas, E. Panagopoulou and N. Kyparissi-Apostolika 2007 TL age-estimates for the Middle Palaeolithic layers at Theopetra cave (Greece). Quaternary Geochronology 2:303-308. Van Andel, T. H. and J. C. Shackleton 1982 Late Paleolithic and Mesolithic Coastlines of Greece and the Aegean. Journal of Field Archaeology 9:445-454. Van Andel, T. H. and S. B. Sutton 1988 Landscape and People of the Franchthi Region. Excavations of Franchthi Cave, Greece, Fascicle 2. Indiana University Press, Bloomington and Indianapolis. Van Andel, T. H. and P. C. Tzedakis 1996 Palaeolithic Landscapes of Europe and Environs, 150,000-25,000 Years Ago: An Overview. Quaternary Science Reviews 15:481-500. van Vuure, C. 2005 Retracing the Aurochs: history, morphology and ecology of an extinct wild ox. Pensoft Publishers, Sofia. Vandiver, P. B., O. Soffer, B. Klima and I. Svoboda 1989 The origin of ceramic technology at Dolni Vestonice, Chechoslovakia. Science 246:1002-1008. Vanhaeren, M., F. d'Errico, C. Stringer, S. L. James, J. A. Todd and H. K. Mienis 2006 Middle Paleolithic shell beads in Israel and Algeria. Science 312:1785-1788. Vavalekas, K., C. Thomaides, E. Papaevangellou and N. Papageorgiou 1993 Nesting biology of the Rock Partridge Alectoris graeca graeca in northern Greece. Acta Ornithologica 28(2):97-101. Vehik, S. C. 1977 Bone fragments and bone grease manufacturing: a review of their archaeological use and potential. Plains Anthropologist 22(77):169-182. Villa, P. and F. d'Errico 2001 Bone and ivory points in the Lower and Middle Paleolithic of Europe. Journal of Human Evolution 41:69-112. Villaverde, V., R. M. Valle, P. G. Calatayud and M. P. Fumanal 425 1996 Mobility and the role of small game in the middle Paleolithic of the central region of the Spanish mediterranean: A comparison of the Cova Negra with other Paleolithic deposits. In The Last Neandertals, the First Anatomically Modern Humans, edited by E. Carbonell and M. Vaquero, pp. 267-288. Universitat Rovira i Virgili, Tarragona, Spain. Waguespack, N. M. 2001 Caribou sharing and storage: Refitting the Palangana site. Journal of Anthropological Archaeology 21(3):396-417. Walker, M. J., J. Gilbert, M. López, A. V. Lompardi, A. Pérez-Pérez, J. Zapata, J. Ortega, T. Higham, A. Pike, J.-L. Schwenninger, J. Zilhão and E. Trinkhaus 2008 Late Neandertals in Southeastern Iberia: Sima de las Palomas del Cabezo Gordo, Murcia, Spain. Proceedings of the Prehistoric Society 105(52):20631-20636. Watts, W. A., J. R. M. Allen and B. Huntley 1996 Vegetation History and Palaeoclimate of the Last Glacial Period at Lago Grande di Monticchio, Southern Italy. Quaternary Science Reviews 15:133-153. Weaver, T. D., R. H. Boyko and T. E. Steele 2011 Cross-platform program for likelihood-based statistical comparisons of mortality profiles on a triangular graph. Journal of Archaeological Science 38:24202423. Weiner, S., P. Goldberg and O. Bar-Yosef 1993 Bone preservation in Kebara Cave, Israel using on-site Fourier-transform infrared spectrometry. Journal of Archaeological Science 20(613-627). Wenban-Smith, F. F. 1997 Refitting of Lithic Artefacts. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece. Vol. 1: Excavation and intra-site analysis at Klithi, edited by G. N. Bailey, pp. 95-104. McDonald Institute for Archaeological Research, Cambridge. Weniger, G. C. 1987 Magdalenian settlement patterns and subsistence in central Europe: The southwestern and central German cases. In The Pleistocene Old World: Regional Perspectives, edited by O. Soffer, pp. 201-215. Plenum, New York. West, D. 1996 Horse hunting, processing and transport in the Middle Danube. In Paleolithic in the Middle Danube Region, edited by J. Svoboda, pp. 207-245. Archeologicky ustav AV CR, Brno. 426 1997 Hunting Strategies in Central Europe During the Last Glacial Maximum 672. BAR International Series, Oxford. Wijmstra, T. A. 1969 Palynology of the First 30 Metres of a 120 m Deep Section in Northern Greece. Acta Botanica Neerlandica 18:511-527. Wild, E. M. 2005 Direct dating of Early Upper Palaeolithic human remains from Mladec. Nature 435:332-335. Wilkinson, T. J. and S. T. Duhon 1991 Franchthi Paralia: The Sediments, Stratigraphy, and Offshore Investigations. Excavations at Franchthi Cave, Greece, Fascicle 6. Indiana University Press, Bloomington and Indianapolis. Willis, K. 1997 Vegetational History of the Klithi Environment: a Palaeoecological Viewpoint. In Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece, edited by G. N. Bailey, pp. 395-413. vol. 2. Oxbow Books, Oxford. Winterhalder, B. 1981 Foraging strategies in the Boreal Forest: An analysis of Cree hunting and gathering. In Hunter-Gatherer Foraging Strategies: Ethnographic and Archeological Analyses, edited by B. Winterhalder and E. A. Smith, pp. 66-98. University of Chicago Press, Chicago. 2002 Behavioral and Other Human Ecologies: Critique, Response and Progress through Criticism. Journal of Ecological Anthropology 6:4-22. Wolverton, S. 2002 NISP:MNE and %whole in analysis of prehistoric carcass exploitation. North American Archaeologist 23(2):85-100. Yellen, J. E. 1977 Cultural patterning in faunal remains: evidence from the !Kung Bushmen. In Experimental Archeology, edited by D. Ingersoll, J. E. Yellen and W. Macdonald, pp. 271-331. Columbia University Press, New York. 1991 Small mammals: !Kung San utilization and the production of faunal assemblages. Journal of Anthropological Archaeology 10:1-26. Zeanah, D. W. 1999 Transport costs, central place foraging, and hunter-gatherer alpine land use strategies. In Intermountain Archaeology: Selected Papers of the Rocky Mountain 427 Anthropological Conference, edited by D. B. Madsen and D. Metcalfe, pp. 1-14. University of Utah Anthropological Papers, Salt Lake City. 2004 Sexual division of labor and central place foraging: a model for the Carson Desert of western Nevada. Journal of Anthropological Archaeology 23:1-32. Zilhão, J. 2006 Genes, Fossils, and Culture. An Overview of the Evidence for NeandertalModern Human Interaction and Admixture. Proceedings of the Prehistoric Society 72:1-20. Zilhão, J., D. E. Angelucci, E. Badal-Garcia, F. d'Errico, F. Daniel, L. Dayet, K. Douka, T. F. G. Higham, M. J. Martinez-Sanchez, R. Montes-Bernardez, V. Villaverde, R. Wood and J. Zapata 2010 Symbolic use of marine sheels and mineral pigments by Iberian Neandertals. PNAS 107(3):1023-1028. Zilhão, J. and F. d'Errico 1999 The Chronology and Taphonomy of the Earliest Aurignacian and Its Implications for the Understanding of Neandertal Extinction. Journal of World Prehistory 13(1):1-68. Zilhão, J. and P. B. Pettitt 2006 On the new dates for Gorhm's Cave and the late survival of Iberian Neanderthals. Before Farming 3:1-9.
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project