|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 LATE PLEISTOCENE TEC HNOLOGICAL CHANGE AN D HUNTER GATHERER BEHAVIOR AT MOCHE BO RAGO ROCKSHELTER, SO DO WOLAYTA, ETHIOPIA: FLAKED STONE ARTIFAC TS FROM THE EARLY OI S 3 (60 43 KA) DEPOSITS By ERICH C. FISHER A DISSERTATION PRE SENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Erich C. Fisher
3 To the Wolayta pe ople living at Moche Borago today Tosimo.
4 ACKNOWLEDGEMENTS I thank my committee chair, and friend, Dr. Steven Brandt first and foremost for his support, advice, and wisdom these past few years. I also thank my doctoral committee for their advice throughout the entire Ph.D. process. My research was graciously funded by the J. William Fulbright U.S. student program, and I would particularly like to thank my Fulbright program adviser, Jermaine Jones as well as Yohannes Birhanu at the U.S. Embassy i n Addis Ababa. I also acknowledge the support of Dr. Zinabu Gebremariam, Markos Tekle, and the faculty and staff at Hawassa University who warmly received me, gave me space to work, and accommodation. Similarly, I thank Wezeru Mamitu Yilma and Ato Menker Bitew at the National Museum of Ethiopia for providing space to analyze the collections, and the Sodo Bureau of Culture, SNNPR Bureau of Culture, and, most importantly, the Authority for Research and Conservation of Cultural Heritage for their continued s upport of SWEAP research. In particular, I thank Ato Jara Hailemariam, Director of ARCCH, and Dr. Yonas Beyene, Head of Archaeology. I would be remiss not to acknowledge my appreciation to M s Anna Fernyhough and Dr. Gebre Yntiso for accommodation and fr iendship while I was living in Ethiopia. I also thank Stephen Burns, Jason Cosford, Hai Cheng, Antje Voelker, Syee Weldeab, Seifu Kebede, Mohammed Umer, and Henry Lamb for providing ideas and/or data about paleolclimates related to my research. Lastly, I extend my most gracious thank you to my parents, my family, and my friends who have supported me tirelessly throughout these past few years.
5 TABLE OF CONTENTS P age ACKNOWLEDGEMENTS ................................ ................................ ................................ ............. 4 LIST OF TABLES ................................ ................................ ................................ ........................... 9 LIST OF FIGURES ................................ ................................ ................................ ....................... 12 CHAPTERS 1 INTRODUCTION ................................ ................................ ................................ .................. 16 2 THE PALEOCLIMATIC AND PALEOENVIRONMENTAL CONTEXT IN THE HORN OF AFRICA FROM OIS 4 TO OIS 3 ................................ ................................ ........ 22 Glacial and Interglacial Cycling ................................ ................................ ............................. 22 Major Data Sources ................................ ................................ ................................ ................ 24 Millennial Scale Events During OIS 4 and OIS 3 ................................ ................................ .. 25 Heinrich Events ................................ ................................ ................................ ............... 26 Dansgaard Oeschger Events ................................ ................................ ............................ 28 West African and SW Asian Monsoonal Systems ................................ .......................... 29 The Paleoclimatic and Pale oenvironmental Context of Europe, Asia, and Africa Between 65 and 43ka ................................ ................................ ................................ .......... 31 OIS 4: 73.5 to 65 ka ................................ ................................ ................................ ......... 31 OIS 4 to OIS 3: 65 to 55 ka ................................ ................................ ............................. 32 Early OIS 3: 55 to 48 ka ................................ ................................ ................................ .. 33 Early Mid OIS 3: 48 to 43 ka ................................ ................................ .......................... 36 Conclusions ................................ ................................ ................................ ............................. 37 3 A BRIEF HISTORY OF THEORETICAL IDEAS ABOUT HUNTER GATHERER BEHAVIORAL VARIABILITYAND ARID CLIMATIC EVENTS DURING THE LATE PLEISTOCENE ................................ ................................ ................................ ........... 47 Modern Human Behaviors ................................ ................................ ................................ ...... 47 Late Pleistocene Behavioral Variability in the Archaeological Record ................................ 48 Arid Adaptation Ideas Since the Ea rly 20th century ................................ .............................. 49 1900 1940 ................................ ................................ ................................ ........................ 50 1940 Present Day ................................ ................................ ................................ ............ 51 Conclusions ................................ ................................ ................................ ............................. 53 4 ARCHAEOLOGICAL SITES IN EASTERN AFRICA DATING TO THE LATE P LEISTOCENE (OIS 4 AND OIS 3) ................................ ................................ ..................... 56 OIS 3 Archaeological Record: Limitations ................................ ................................ ............ 56 Chronological and Sampling Problems ................................ ................................ ........... 56 Problems Arising from Late Pleistocene Demography and Settlement Patterns ............ 58
6 OIS 3 Archaeological Record: Available Data ................................ ................................ ....... 59 Mode 4 and Mode 5 Stone Tools ................................ ................................ .................... 60 Conclusions ................................ ................................ ................................ ............................. 63 5 MOCHE BORAGO ROCKSHELTER: LOCATION, ORAL HISTORY, AND PRIOR ARCHAEOLOGICAL RESEARCH ................................ ................................ ...................... 65 Physiography of the Wolayta Region ................................ ................................ ..................... 65 Ethiopian Rift Valley System ................................ ................................ .......................... 66 Climate ................................ ................................ ................................ ............................ 66 Physiography and Vegetation ................................ ................................ .......................... 67 Language ................................ ................................ ................................ ......................... 69 Political History ................................ ................................ ................................ ............... 70 Local History of Moche Borago Rockshelter ................................ ................................ ......... 71 Prior Archaeological Research at Moche Borago ................................ ................................ .. 72 GEPCA Excavations: 1998 and January February 2000 ................................ ................ 72 GEPCA Excavation: November 2000 ................................ ................................ ............. 73 GEPCA Excavation: December 2001 ................................ ................................ ............. 74 SWEAP Excavations: 2006 2008 ................................ ................................ .................... 75 SWEAP Excavations: 2006 ................................ ................................ ............................. 78 SWEAP Excavation: 2007 ................................ ................................ .............................. 80 SWEAP Excavation: 2008 ................................ ................................ .............................. 81 Esay Rockshelter ................................ ................................ ................................ .................... 83 Conclusions ................................ ................................ ................................ ............................. 83 6 THE LITHO STRATIGRAPHIC SEQUENCE AT MOCHE BORAGO ROCKSHELTER DURING EARLY OIS 3 ................................ ................................ .......... 87 Litho Stratigraphic Units Versus Culture Stratigraphic Units ................................ ............... 87 Methods ................................ ................................ ................................ ................................ .. 88 Stratigraphic Descriptions and Profiles ................................ ................................ ........... 88 Multidimensional GIS (mDGIS) Modeling ................................ ................................ ..... 89 Bulk Sample Analysis ................................ ................................ ................................ ..... 90 Inductively Coupled Plasma Mass Spectrometry ................................ ........................... 90 X Ray Florescence ................................ ................................ ................................ .......... 91 Magnetic Susceptibility ................................ ................................ ................................ ... 93 Radiocarbon ................................ ................................ ................................ ..................... 93 Litho Stratigraphic Groups of the Block Excavation Area ................................ .................... 95 PKT ................................ ................................ ................................ ................................ .. 95 T Group Deposits ................................ ................................ ................................ ............ 96 Layer d escriptions ................................ ................................ ................................ .... 97 Occupational Hiatus #1 ................................ ................................ ................................ ... 98 S Group Deposits ................................ ................................ ................................ ............ 99 Layer d escriptions ................................ ................................ ................................ .. 100 Occupational Hiatus #2 ................................ ................................ ................................ 102 R Group ................................ ................................ ................................ ......................... 103 Late and Terminal Pleistocene Unconformity ................................ ............................... 104
7 Occupational Hiatus #3/BXA H Group ................................ ................................ ........ 106 Layer Descriptions ................................ ................................ ................................ 106 Stratigraphic Summary of TU2 ................................ ................................ ............................ 107 L Group ................................ ................................ ................................ ......................... 107 S Group ................................ ................................ ................................ ......................... 108 Occupational Hiatus #2 ................................ ................................ ................................ 108 R Group ................................ ................................ ................................ ......................... 108 Occupational Hiatus #3/TU2 H Group ................................ ................................ ......... 109 Stratigraphic Summary of N42E38 ................................ ................................ ...................... 110 Current Basal Sediments ................................ ................................ ............................... 110 L Group ................................ ................................ ................................ ......................... 111 S Group ................................ ................................ ................................ ......................... 112 Occupational Hiatus #3/N42E38 H* Group ................................ ................................ .. 113 A Working Model of the OIS 3 Depositional History at Moche Borago Rockshelter ......... 113 ~60 ka to 45 ka ................................ ................................ ................................ .............. 114 45 ka to 43 ka ................................ ................................ ................................ ................ 115 Conclusions ................................ ................................ ................................ ........................... 118 7 DESCRIPTIVE AND STATISTICAL ANALYSES OF THE T GROUP AND S GROUP ASSEMBLAGES AT MOCHE BORAGO: RAW MATERIALS, CORES, AND LITHIC DEBITAGE ................................ ................................ ................................ ... 121 General Description of the Assemblages ................................ ................................ .............. 122 Minimum Number of Lithics ................................ ................................ ................................ 12 2 Hammerstones and Raw Material Nodules ................................ ................................ .......... 123 Raw Materi als ................................ ................................ ................................ ....................... 124 Cores ................................ ................................ ................................ ................................ ..... 128 Core Size ................................ ................................ ................................ ....................... 129 Diversity of Single and Multiple Platform Cores ................................ .......................... 130 Single p latform c ores ................................ ................................ ............................. 131 Pyramidal and p rismatic c ores ................................ ................................ ............... 132 Co re f lake t ype ................................ ................................ ................................ ....... 132 Debitage ................................ ................................ ................................ ................................ 133 Conclusions ................................ ................................ ................................ ........................... 139 8 DESCRIPTIVE AND STATISTI CAL ANALYSIS OF THE T GROUP AND S GROUP ASSEMBLAGES AT MOCHE BORAGO: UNSHAPED AND SHAPED TOOLS ................................ ................................ ................................ ................................ 160 Overview of the Unshaped and Shaped Tools ................................ ................................ ...... 160 Unshaped Tools ................................ ................................ ................................ .................... 161 Miscellaneous Shaped Tools ................................ ................................ ................................ 162 Microliths ................................ ................................ ................................ .............................. 163 Scrapers ................................ ................................ ................................ ................................ 164 Burins ................................ ................................ ................................ ................................ .... 166 Points ................................ ................................ ................................ ................................ .... 167 The Point Assemblage ................................ ................................ ................................ ... 169
8 Point Descriptions ................................ ................................ ................................ ................. 170 Technological Summary of the Moche Borago Points ................................ .................. 170 Typolog ical Summary of the Moche Borago Points ................................ ..................... 173 Diachronic Technological and Typological Point Patterns ................................ .................. 175 Conclusions ................................ ................................ ................................ ........................... 176 9 SYNTHESIS ................................ ................................ ................................ ......................... 212 What is the E vidence at G lobal, C ontinental, and L ocal Scales for C limatic F luctuations d uring E arly and M iddle OIS 3? How M ight t h ese Changes Have Affected L ocal E cology A round Moche Borago? ................................ ................................ ..................... 213 If Evidence Exists at V arious S cales for C limatic and E nvironmental F luctuations d uring OIS 3, How Did The M obility P atterns, S ubs istence Strategies, and Social O rganization of H unter G atherers Vary in R esponse to P aleoenvironmental I nstability? ................................ ................................ ................................ ......................... 216 If E vidence E xists at Various Scales for C limatic and E nvironmental F luctua tions d uring OIS 3, How D id Hunter G atherer Stone T ool T echnology Vary in R esponse to Paleoenvironmental F luctuations? ................................ ................................ .................... 220 Moche Borago in Broader Perspective ................................ ................................ ................. 224 Conclusions ................................ ................................ ................................ ........................... 227 APPENDIX A LIST OF ABBREVIATIONS AND DEFINITIONS USED IN THE TEXT ...................... 230 B LIST OF ARCHAELOG ICAL SITES FROM THE HORN OF AFRICA, EAST AFRICA, AND NORTH AFRICA, DATING TO THE LAST GLACIAL PERIOD WHICH HAVE DEPOSITS CONTAINING MODE 3 AND MODE 4/5 LITHICS .......... 232 C DETAILED DESCRIPTIONS OF THE MICROLITHS AT MOCHE BORAGO ............. 247 Level 12 ................................ ................................ ................................ ................................ 247 Level 13 ................................ ................................ ................................ ................................ 247 Level 14 ................................ ................................ ................................ ................................ 247 Level 16 ................................ ................................ ................................ ................................ 247 Level 18: ................................ ................................ ................................ ............................... 248 Level 23: ................................ ................................ ................................ ............................... 248 D DETAILED DESCRIPTIONS OF THE POINTS AT MOCHE BORAGO ........................ 249 REFERENCE LIST ................................ ................................ ................................ ..................... 254 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 270
9 LIST OF TABLES Table P age 3 1: Description of Paleoclimatic Sites in Figure 3 1 ................................ ............................... 41 3 2: List of Dansgaard Oeschger Events from the Last Glacial to the Present ......................... 44 6 1: The AMS Radiocarbon Assay from Moche Borago ................................ ........................ 120 7 1: Summary of the major lithic groups by excavation level and stratigraphic group. ......... 140 7 2: Summary of the major lithic groups by excavation level and stratigraphic group without medial or lateral flake d ebitage (MNL) ................................ .............................. 141 7 3 : Metric information for nodules and hammerstones ................................ ......................... 143 7 4: MNL Frequencies and percentages of the various r aw material groups in the T Groups and S Group ................................ ................................ ................................ ........ 145 7 5: MNL Frequencies of raw material groups within the T Group and S Group deposits. .. 146 7 6: MNL Frequencies of individual raw material types organized by stratigraphic aggregate ................................ ................................ ................................ .......................... 148 7 7: Frequency and volumetric measurements for core types and core groups from each of t he three litho stratigraphic units discussed in the text (non MNL) ............................ 149 7 8: Frequencies of core types in the S Group and T Group Upper and Lower (non MNL) 151 7 9: Relative percents of core flake removal types per excavation levels (non MNL) .......... 152 7 10: Percentages of the major lithic groups by stratigraphic aggregates for the entire assemblage (upper) and MNL only (lower) ................................ ................................ ..... 153 7 11: Average volume (mm 3 ) of cores, shaped and unshaped tools per excavation level and litho stratigraphic group (non MNL) ................................ ................................ ............... 154 7 12: Total frequencies and percentages of various shaped and unshaped tool categories (non MNL) ................................ ................................ ................................ ....................... 159 7 13: MNL proportions of the blank type s used to make shaped and unshaped tools within each litho stratigraphic unit ................................ ................................ ............................. 159 8 1: Frequency (upper) and percentages (lower) of the unshaped and shaped tools per total number of artifacts in this class ................................ ................................ ............... 179 8 2: Percentage of unshaped and shaped tools per stratigraphic aggregate ............................ 179
10 8 3: MNL frequency (upper) and percenta ges (lower) of the unshaped and shaped tools in the S Group and T Group ................................ ................................ ................................ 180 8 4: Frequency and percentage per stratigraphic group of unshaped tools based on non MNL counts (above) and MNL count s (below) ................................ .............................. 181 8 5: Percentages of edge damage on unshaped tools (MNL) based on the location of edge damage ................................ ................................ ................................ ............................. 182 8 6: Edge damage typ e on unshaped tools (MNL) ................................ ................................ .. 183 8 7: Relative percentage of blank types used for modified and utilized unshaped tools (non MNL) ................................ ................................ ................................ ....................... 183 8 8: Mean aspect ratio and length values for whole modified and utilized unshaped tools (MNL) ................................ ................................ ................................ .............................. 184 8 9: Basic metrical data and descriptions for miscellaneous shaped tools in the S Group and T Group (MNL) ................................ ................................ ................................ ........ 185 8 11: Frequency of scrapers and scraper fragments by type within each level and stratigraphic aggregate ................................ ................................ ................................ ..... 188 8 12: Relative percentage of scraper types and scraper fragments per level ............................ 189 8 13: MNL frequency of scrapers and scraper fragments by type within each level and stratigraphic aggregate ................................ ................................ ................................ ..... 190 8 14: Relative percentage of scraper types and scraper fragments per level based on MNL counts ................................ ................................ ................................ ............................... 191 8 15: Basic metrical data and des criptions of the scrapers (non MNL) found in the S Group and T Group ................................ ................................ ................................ .......... 192 8 15: Continued ................................ ................................ ................................ ......................... 193 8 15: Continued ................................ ................................ ................................ ......................... 194 8 15: Continued ................................ ................................ ................................ ......................... 195 8 16: Mean aspect ratios of all scrapers types (MNL) subdivided by litho stratigraphic group ................................ ................................ ................................ ................................ 196 8 17: Mean aspect ratios of MNL end scrapers only subdivided by litho stratigraphic group. ................................ ................................ ................................ ............................... 196 8 18: Distribution of retouch location per scraper type (MNL ) ................................ ................ 197 8 19: Location of retouch per scraper type ................................ ................................ ............... 198
11 8 20: Angle of the working edges for end and side scrapers ................................ .................... 199 8 21: Frequency of working edge form for end scrapers (above) and side scrapers (below) ... 199 8 22: Basic metric information on notched flakes from the S Group and T Group ................. 200 8 24: Frequencies of point retouch typ e ................................ ................................ .................... 203 8 25: Percentages of point retouch type ................................ ................................ .................... 203 8 26: Completeness of the points (non MNL) within the T Group and S Group ..................... 204 8 27: Frequency of bulbar thinning on points ................................ ................................ ........... 20 4 8 28: Percentage of bulbar thinning found on points ................................ ................................ 205 8 29: Basic metric information for whole points ................................ ................................ ...... 206 8 29: Continued ................................ ................................ ................................ ......................... 207
12 LIST OF FIGURES Figure P age 3 1: Paleoclimatic proxy sites across the circum Indian Ocean ar ea that are described in the text ................................ ................................ ................................ ................................ 40 3 2: Timeline of the Last Glacial showing the boundaries of Oxygen Isotope Periods and Dansgaard Oeschger events ................................ ................................ ............................... 43 3 3: Generalized wind patterns of the summer monsoon systems in Africa and SW Asia ....... 45 3 4: Contemporary mean annual precipitation rate across the Horn of Africa. Data cre dit: FAO / UNEP Desertification and Mapping Project ................................ ........................... 46 4 1: Map of the Horn of Africa (red) and East Africa (yellow) showing the archaeological sites discussed in the text ................................ ................................ ................................ ... 64 5 1: GEPCA excavation areas at Moche Borago from 1998 to 2001 ................................ ....... 85 6 1: This composite image shows the stratigraphic profiles from the BXA (left), TU2 (c enter) and N42 (right) excavation areas ................................ ................................ ....... 119 7 1 : BN 3115.1, Level 23. Stratig raphic Unit DCC (T Group Upper) ................................ .. 142 7 2: XRF spec tra of Common Black Obsidian (CBO) ................................ ............................ 144 7 3: MNL proportions o f CBO and non CBO raw materials in the S Group. The data are further subdivided by lithic group ................................ ................................ .................... 147 7 4: MNL proportions of CBO and non CBO raw materials in the T Group Upper. The data are f urther subdivided by lithic group ................................ ................................ ...... 147 7 5: MNL proportions of CBO and non CBO raw materials in the T Group Lower. The data are further subdivided by lithic group ................................ ................................ ...... 148 7 6 : Selected single and multi platform cores from the G10 assemblage. Arrows denote step or hinge fractures ................................ ................................ ................................ ...... 150 7 7: Bar chart showing the mean length of whole flak es per excavation level (MNL) .......... 155 7 8: Bar chart showing the mean aspect ratio (length/width) of whole flake per excavation level (MNL) ................................ ................................ ................................ ..................... 156 7 9: MNL frequency of raw material groups f rom excavation level 11 to 32 ........................ 157 7 10: Relative proportions of lithic groups in the S Group and T Group stratigraphic aggregates ................................ ................................ ................................ ........................ 158
13 8 1: Average aspect ratio of unshaped (modified and utilized) whole flake blanks between the three litho stratigraphic groups ................................ ................................ .... 184 8 2: Microliths from the BXA T Group and S Group deposits ................................ .............. 187 8 3: Angle burin from Level 13 with possible backing along the left lateral edge ................. 202 8 4: Dihedral burin recovered from Level 30 ................................ ................................ ......... 202 8 5: Histogram of the mean lengths of whole points per stratigraphic aggregate ................... 208 8 7: Points from the Moche Borago OIS 3 deposits. These points come from the BXA and TU2 areas, as well as the T Group, S Group, and the R Group ............................... 209 8 8: Point technological tra ditions represented at Moche Borago ................................ .......... 210 8 9: Point typology for Moche Borago. The points represented are ordered in a roughly chronological sequence ................................ ................................ ................................ .... 211
14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy LATE PLEISTOCENE TECHNOLOGICAL CHANGE AND HUNTER GATHERER BEHAVIOR AT MOCHE BORAGO ROC KSHELTER, SODO WOLAYTA, ETHIOPIA: FLAKED STONE ARTIFACTS FROM THE EARLY OIS 3 (60 43 KA) DEPOSITS By ERICH C. FISHER August 2010 Chair: Steven A. Brandt Major: Anthropology The current archaeological evidence suggests that hunter gatherers living in A frica during the Late Pleistocene rapidly expanded into Europe and Asia by 50 40 ka. It is widely believed that these populations of humans practiced a modern behavioral repertoire that included, among other things, complex social and technological skil ls. However, there are only a handful of sites across the Horn of Africa that date to this critical time period, known as Oxygen Isotope Stage 3. Consequently, very few studies have been able to comment on any of the behavioral or environmental precondit ions that may have influenced or affected these populations prior to, and during, the hypothesized migrations via the Horn and out of Africa. This dissertation presents new data from the Moche Borago rockshelter, located on the western flank of the volcano Mt. Damota near the current town of Sodo in Wolayta, southwest Ethiopia. Moche Borago has been studied since 2006 by the Southwest Ethiopia Archaeological Project and the sequence provides a n intact series of archaeological deposits dating to early OIS 3 ~54 40 ka. Th is dissertation focuses on the analysis of stone tools from the lower deposits at Moche Borago dating from ~54 43 ka to infer technological behavioral change in southwest Ethiopia during this time period.
15 These findings are compared against broader paleoclimatic proxy records which suggest that there may have been rapid and repeated fluctuations in the West African and S outhwest Asia n monsoon systems at this time. Monsoonal flux may have affected local ecology and resources, and it is hypothesized here that changes in the stone tools reflect local behavioral adaptations to monsoon driven local environmental change around Moche Borago. The rapid and repeated nature of the climatic and environmental flux during this time period may ha ve forced resident human populations in the Horn of Africa to develop a flexible behavioral strategy that allowed these people to quickly adopt their existing social and technical behaviors to suite changing contexts. Behavioral adaptations to monsoon dri ven environmental changes may partially explain the unique stone tool record in the Horn of Africa during the Late Pleistocene.
16 CHAPTER 1 INTRODUCTION During the Last Glacial period (73.5 14.7 ka) anatomically and behaviorally modern hunter gatherers moved out of Africa. Archaeological evidence shows that the se colonizers spread into Australia by 6 0 to 4 0 ka (Roberts et al., 1999; Thorne et al., 1999; Bowler et al., 2003) and into Europe by 4 5 to 4 0 ka (Zilho, 2006; Zilho et al., 2007) Eventually, the descendants of these hunter gatherers arrived in North America by ~20 ka (Forster et al., 1996; Oppenheimer, 2003) The many environments these people moved th r ough undersc ore their innate ability to adapt to and exploit diverse ecological conditions successfully The main reason attributed to Pleistocene and into the Late Pleistocen e (~300 30 ka). These behaviors are broadly seen as a suite of new social, technological, and symbolic tools that gave African hunter gatherers the ability to contend with a wider range of natural and social environments and exploit resources more effecti vely and efficiently. A subject of current debate is the specific nature of these (McBrearty and Brooks, 2000; Wadley, 2001; Marean and Assefa, 2005) My interest in this debate focuses upon what role(s) Late Quaternary paleoenvironmental changes may have had in the establishment of these modern human behaviors. Numerous scholars hypothesize that these behavioral changes were direct or indirect response s by African hunter gatherers to either the long, cold, arid conditions of the Penultimate Glacial ( 195 128 ka ) and/or the Last Glacial periods ( 73.5 14.7 ka ) or to the long stretches of warm and humid climates that characterized the Last Interglacial pe riod ( 128 73.5 ka ) (Clark, 1960; Clark, 1988; McBrearty and Brooks, 2000; Ambrose et al., 2002; Barton et al., 2009) Many of the models
17 generated by these scholars argue for direct causal relationships betwee n paleoenvironmental change and behavioral change. However, these models may be too simplistic because they fail to recognize the variability and instability that characterized Glacial and Interglacial climates of the Late Quaternary. These climates have been revealed by newer and more detailed climatic records (Wolff et al., 2010; Capron et al., 2010) Multiple marine and terrestrial proxies from North America, Europe, and Asia reveal Oxygen Isotope Stage 3 (OIS 3) (~59 28 ka), as one of the most climatically unstable oxygen isotope stages of the last 200,000 years if not the most climatically unstable oxygen isotope stage. Although relatively warmer and more humid than the cold and arid Oxygen Isotope Stage 4 (~72 59 ka) that preceded it, OIS 3 was punctuated by numerous short periods of even warmer and wetter climates called Dansgaard Oeschger (D O) events, as well as significantly colder and more arid events called Heinrich events. The climatic fluctuation s during OIS 3 are known to have changed European and African vegetation (Sanchez Gone et al., 2008; Hessler et al., 2010) Across eastern Africa 1 these climatic variations are widely believed to have influenc ed the intensity of monsoonal moisture originating from the Atlantic and Indian oceans (Wang et al., 2001; Burns et al., 2003; Weldeab et al., 2007; Revel et al., 2010) These fluctuations in turn affected veg etation patterns across the region (Kiage and Liu, 2006) How African hunter gatherer subsistence, settlement, and social strategies may have responded to OIS 3 fluctuations, as well as what role(s) early OIS 3 (60 45 ka) climatic instability may have played in the development of modern human behavior at the threshold of 1 Burundi, Eritrea, Ethiopia, Djibouti, Kenya, Rwanda, Somalia, Tan zania, and Uganda.
18 hunter gatherer migrations through and out of A frica, have never been seriously considered. This lack of study is due in part to: scientists questioning whether or not any secure terrestrial paleoclimatic data from Africa can be correlated definitely to Heinrich and D O events the complete absence of African archaeological sites firmly dated to the end of OIS 4 and early OIS 3 However, recent excavations by the Southwest Ethiopian Archaeological Project (SWEAP) at Moche Borago (MB) r ockshe l ter, located in the highlands of Southwestern (SW) Ethiopia, have revealed a radiocarbon dated archaeological sequence that spans the critical time period from 60 to 4 0 ka (Brandt et al., 2006; Hildebrand et al., 2008; Brandt et al., 2010) My dissertation draws upon t his new database by posing four major research questions: What is the evidence at global, continental, and local scales for climatic fluctuations during early and middle OIS 3? How might these climatic changes have affected the local ecology around Moche Borago? If evidence exists at various scales for climatic and environmental fluctuations during OIS 3, how did the mobility patterns, subsistence strategies, and social organization of hunter gatherers vary in response to paleoenvironmental instability? If evidence exists at various scales for climatic and environmental fluctuations during OIS 3, how did hunter gatherer stone tool technology vary in response to paleoenvironmental fluctuations? By addressing these major questions, the goal of this disser tation is to provide data at the local scale that can contribute to a better understanding of human behavioral adaptations during OIS 3 on the regional, continental, and global scales. In the future, the data from Moche Borago can be
19 used to develop more informed hypotheses relating to the establishment of modern human behavior in the Late Pleistocene. This dissertation is organized into nine chapters. Following this introductory chapter, Chapter 2 provides an overview of current research on Late Pleisto cene paleo climates and environments of Europe, Asia, and Africa in order to answer the first and second of the local scales for major and rapid climatic fluctua tions in eastern Africa during early OIS 3? I use a multi scalar approach to tackle this question. At the broadest scale are high latitude ice cores from Greenland that can be linked to continental records in Europe and Africa showing mostly concurrent c limatic and environmental changes (Grootes et al., 1993; Stuiver and Grootes, 2000; Sanchez Gone et al., 2008; Hessler et al., 2010; Fletcher et al., 2010) Records from the northwest (NW) African Atlantic mar gin, Mediterranean coast, the Nile Basin, the Levant, and the circum Indian Ocean area show similar paleoclimatic fluctuations during OIS 3, which affected the intensity of monsoon systems (Bar Matthews et al., 2000; Burns et al., 2003; Weldeab et al., 2007; Revel et al., 2010) Lake cores (Livingstone, 1975; Gasse, 2000; Kebede et al., 2005) and pollen assemblages (Kiage and Liu, 2006; Hessler et al., 2010) provide data on a regional level. At the most proximal scale, the geomorphological history of the deposits at Moche Borago reveals periods of much wetter conditions in the area. These wetter periods are hypothesized to be due to increased monsoonal moisture. Chapter 3 summarizes the intellectual history of models linking arid climatic events and human behavior in African archaeology, and also reviews ethnographic observations about how African hunter gatherers adap t to arid environments. This intellectual history is applied to
20 Questions 3 and 4 which concern how the archaeological data from Moche Borago can be used to infer hunter gatherer responses to Late Pleistocene paleoenvironmental change. Chapter 4 gives summary background information about archaeology in the Horn of Africa dating from OIS 4 to OIS 3. A relatively few archaeological sites date to OIS 3, and many of the currently available sites are poorly dated or have gaps in the depositional sequence. The lack of archaeological data makes comparisons between sites difficult, and knowing where and how hunter gatherers were living in the Horn is still largely unresolved. Chapter 5 provides descriptions of the region around Moche Borago the current inha bitants and excavations at the site The site was first excavated in 1996 by a research team from the Centre National de la Recherche Scientifique whose main interest was Holocene archaeology. Their excavations revealed Pleistocene deposits below the Ho locene layers, which have been the focus for the Southwest Ethiopia Archaeology Project (SWEAP) since 2006. My interpretations of the site stratigraphy and context are based on SWEAP research at the site during four field seasons. The dataset of stone ar tifacts used in this study comes from the main rockshelter is also introduced in this chapter. Esay is notable for a shallow stream of water that flows across the floor of the site, providing a modern day analog to past fluvial features, as evident in the geomorphic history of Moche Borago. Chapter 6 details the litho stratigraphic sequence at Moche Borago and site specific evidence for climatic change during early OIS 3. The depositional history of the Pleistocene layers at the site is complex, showing contributions by fluvial, aeolian, and volcanic process. The Pleistocene layers subdivided naturally into litho stratigraphic units by thick, archaeologically sterile v olcanic ashes. Evidence for fluvial features at the site are found in one
21 of the litho wetter at this time. Chapter 7 provides the analysis of lithic raw materials, core s, and debitage The stone artifact assemblage is made mainly of locally available obsidian. The oldest and deepest litho stratigraphic unit is characterized by abundant debitage and relatively few cores, which is unlike the S Group that has little debita ge and many more cores. Chapter 8 discusses the unshaped and shaped tools focusing specifically upon points Relatively few unshaped and shaped tools in the lowest litho stratigraphic unit are found at Moche Borago compared to the overlying S Group tha t has more types of tools. The points show technological and typological patterns put into a regional context, suggesting that hunter gatherers employed regional cultural traditions during OIS 3. Chapter 9 returns to the original four research questions a nd hypotheses to explain what the data from Moche Borago may reveal about hunter gatherer responses to paleoenvironmental changes during OIS 3. Each of these four research questions is posed in turn, and relevant data from each chapter are synthesized int o a succinct answer describing how hunter gatherers may have been living and behaving at Moche Borago during OIS 3.
22 CHAPTER 2 THE PALEOCLIMATIC AN D PALEOENVIRONMENTAL CONTEXT IN THE HORN OF AFRICA FROM OIS 4 TO OIS 3 This chapter presents a summar y of the available paleoclimatic and paleoenvironmental data from across the Northern Hemisphere that date from OIS 4 to OIS 3. The purpose of this chapter is to contextualize the findings of my flaked stone artifact analysis from Moche Borago and interpr et ations of these data about behavioral variability during OIS 3. This chapter will focus on changes that may have affected the ecological conditions across the Horn of Africa and southwestern Ethiopia during OIS 3. However, few direct data sources are available from the Horn so I have adopted a multi scalar approach that draws together data from across the Northern Hemisphere, the region, and more local surroundings to Moche Borago. This approach relies on a top down chain of evidence starting from the high latitude ice core records to regional and local data sources that enable me to speculate about what the climatic and environmental conditions around Moche Borago may have been like at various time periods during OIS 3. Following the findings of Burn s (2003; 2004) Cai (2006) and Revel (2010) I believe that hemispheric and regional climatic instability during OIS 3 induced monsoonal variability across the circum Indian Ocean region and the Horn of Africa. I speculate that the changes in monsoonal precipitation across the southwestern Ethiopian highlands altered the local ecology and that local bands of hunter gatherers adapted their lithic behaviors to suit. Glacial and Interglacial Cycling The Last Glacial period occurred from 73.5 to 14 ka It is part of the broader cycling of the which are thought to be due to variations in the e tilt of the e arth, and the wobble in the e (Milankovitch, 1941) These variations of movement affect the radiation that reaches the e arth insolation (Bradley, 1999) By s tudying the distribution of insolation across the e arth,
23 Milankovitch (1941) found periodic anomalies above 65 north latitude that were in phase with continental ice sheet growth. He hypothesized that when the position and movem ent of the earth around the sun minimize insolation above 65 north latitude, then the northern latitudes become cooler and the continental ice sheets expand. The timing of insolation anomalies is calculated to occur approximately every 100,000 years whi ch is coincident with the historical phases of glacial/interglacial cycles seen in the paleoclimatic proxy records dating to the Pleistocene Today, stable oxygen isotopes are a commonly employed method to classify the natural glacial/interglacial cycling studied from numerous sources containing fossil water, but these records are still most often reconstructed from marine biogenic materials such as corals or shell and high latitude ice cor es. Oxygen isotope records rely on variations in the ratio ( 18 O ) of naturally occurring 18 O and 16 O to classify glacial and interglacial climates. The variations in the atomic weights of the different oxygen isotopes ( 18 O is heavier than 16 O ) mean that when temperatures drop d uring glacial periods, 1 8 O is natural ly distilled from precipitation via the process of fractionation (Dansgaard, 1964) The process of fractionation creates lower oxygen isotope ratios during glacial periods, which makes the classification possible (Bradley, 1999) According to the stable oxygen isotope record, the Last Glacial is composed of three distinct stages. The ages I use to define each stage are drawn from the most recent calculations, which are described in Sanchez Goni et al. (2010) The earliest stage, known as Oxygen Isotope Stage 4 (OIS 4) lasted from 73.5 to 59.4 ka and this stage was associated with generally cold and arid conditions. The next stage was OIS 3 (59.4 27.8 ka) Climatic condition s were slightly warmer and wetter during OIS 3. Multiple records suggest an intensified seasonality and general climatic instability across Africa during this time (Flores et al., 2000; Kiage and Liu, 2006) The
24 latest period was OIS 2 last ing from 27.8 to 14 ka. OIS 2 coincides with the end of the Pleistocene epoch and the beginning of the Holocene epoch. OIS 2 is generally known for cold and arid conditions that culminated in the Last Glacial Maximum betwe en 27.2 and 23.5 ka. Major Data Sources Voelker (2002) provide d a comprehensive summary of the current available climatic and environmental proxy data for OIS 3. are represented in Figure 3 1 and Table 3 1 In total, ~250 sites have yielded various types and qualities of data pertaining to climatic fluctuations during OIS 3 T he distribution of sites shows a bias for data derived from high latitude locations suc h as ice sheets and for marine cores leaving large absences of data interior to most continents. The emphasis on ice cores is because these data types provide a wealth of isotopic and other climatic proxy data spanning long time frames within well underst ood formation processes and contexts Marine records also do not suffer many of the complicating effects of diagenesis that terrestrial records may undergo (Bradley, 1999) Three primary issues regarding most of these data sources include dating, resolution, and local influences. Ice and marine cores typically rel y on accelerator mass spectrometry 14 C dating to ~40 ka, but thereafter these records are tuned orbitally using layer counting and age models (Bradley, 1999; Svensson et al., 2006; Svensson et al., 2008) Thi s relative dating method is prone to error and the chronologies are often revisited and adjusted. The most recent readjustment is the Greenland Ice Core Chronology 2005 (GICC05) (Andersen et al., 2006; Svensson et al., 2006; Svensson et al., 2008; Wolff et al., 2010) Speleothem records are now being dated using Uranium series, which provides an independent and direct dating method. However, these records are sensitive to the context of formation (hydrologic al, geological, chemical, and climatic) and they require a closed system to
25 minimize re crystallization of calcite thereby limiting porosity and detrital contamination. Low uranium content in the parent material can also limit the resolution of speleoth em records (Lundblad and Holmgren, 2005) Other ter restrial proxy records provided by lake cores palynology, and phytoliths can provide regional and local climatic records given the right circumstances. Terrestrial freshwater lake core records are largely dependent upon the existence and fluctuations of the lake sediments through time and are generally only dated using 14 C up to ~40 ka (e.g. Gasse, 1977; Kiage and Liu, 2006) E xperimental methods such as 230 Th/ 232 Th dating are being applied at sites such as Sacr ed Lake, Kenya (Olago et al., 2001) Local influences may mask or enhance certain paleoclimatic signals in these data sources 18 O ice core records for example, all show similar patterns of air temperature flux that occurred above Greenland (Dansgaard et al., 1993) Similar climatic patterns are seen in the 18 O records from Hulu cave ( China ) and Moomi cave ( Socotra Island ) speleothem s, but these re cords are also believed to show the regional fluctuations in monsoon systems (Burns et al., 2003; Cai et al., 2006) Millennial Scale Events D uring OIS 4 and OIS 3 Beginning during OIS 4, ~75 ka, Northern He misphere climates started to become increasingly unstable (Wolff et al., 2010) The c limatic fluctuations take two forms. On the one hand periods of intense c old and arid conditions called Heinrich events (H) occurred. These events were likely caused by massive discharges of freshwater ice from the Hudson Strait (Canada) into the North Atlantic Ocean The freshwater influx may have slowed or shut down the Nor th Atlantic thermohaline circulation, which regulates oceanic temperatures and, by proxy, terrestrial temperatures and precipitation (Hemming, 2004) On the other hand, rapid increases
26 in N orthern H emisphere temperatures and humidification called Dansgaard Oeschger eve nts (D O) also occurred Heinrich and Dansgaard Oeschger events began during OIS 4 and increased in intensity during OIS 3, affecting both the Northern and S outhern h emispheres (Hessler et al., 2010; Sanchez G one and Harrison, 2010) Heinrich and Dansgaard Oeschger events are linked phenomenon and t he combined effects of ice sheet and ocean atmosphere interactions by both maintained a frequently fluctuating climatic regime during the Late Pleistocene (especia lly OIS 3). These fluctuations occurred up to the decadal level and the general patterning forms the foundation for the saw toothed, stepwise pattern of climatic changes seen during this period (Leuschner and Sirocko, 2000; Sanchez Gone and Harrison, 2010; Hessler et al., 2010; Wolff et al., 2010) Heinrich Events Heinrich events are quasi periodic excursions (~7,200 years) of massive amounts of Laurentide and Scandinavian sheet ice which discharged from th e Hudson Strait (Canada) into the Atlantic Ocean. The provenance of the ice source is established from high detritral carbonate concentrations which were carried within the ice and deposited along a belt in the Atlantic Ocean approximately 40 to 55 nort h latitude (Ruddiman, 1977; Broecker et al., 1992; Grousset et al., 1993; Leuschner and Sirocko, 2000; Hemming, 2004) The ca u sal factors for Heinrich events are debated. Broecker (1994) suggest ed that Heinrich events were caused by exogenous orbital fluctuations which facilitated ice sheet growth and eventual expansion. MacAyeal (1993) however, proposed an endogenous explanation He argued that following a period of increased ice growth (i.e. binge) the ice sheet became unstable and was purged.
27 Building upon these ide as Maslin (2001) put forth a stepwise model based on freshwater influx from minor ice sheet discharges in the North Atlantic Ocean. According to his model, minor ice sheet discharges disrupted the North Atlantic Deep Water Circulation ( NADW ) which regulates oceanic temp erature The d isruption of the NADW caused S outhern H emisphere heat piracy that in turn, melted Antarctic ice and disrupted the Antarctic Bottom Water Current (AABW). Thus, heating in the Southern Hemisphere is inversely related to cooling in the Northe rn Hemisphere. After several bi hemispheric cycles the progressive ice melting phases in the Northern Hemisphere undercut the Laurentide ice sheet forcing a full fledged Heinrich excursion. Regardless of the exact reason behind the formation of Heinri ch events, two points appear certain. First, Heinrich events occur on a roughly 7,200 year periodicity. The duration of Heinrich events ranges from centennial to millennial scales but lasts no longer than ~2,200 years (Hemming, 2004) Second, climatic changes attribu ted to Heinrich cooling events are abundant in terrestrial and marine records. For example, during Heinrich events depletion of the 18 O signature in foraminifera indicate s massive influx es of freshwater ice during these events, which lowered Atlantic oce to 2C (Bond et al., 1993; Leuschner and Sirocko, 2000; Hemming, 2004) These data are supported by modeling results which also suggest that massive i nflux of freshwater into the Atlantic decrease s NADW circulation, slowing or shut ting down entirely the thermohaline circulation system in the Atlantic (Van Campo et al., 1990; Vidal and Arz, 2004; Weldeab et al ., 2007) Sea surface temperatures (SST) also decrease, thereby changing the location of the Inter tropical Convergence Zone ( ITCZ ) which controls the intensity and patterns of monsoonal winds across Africa and SW Asia (Vidal and Arz, 2004; Tierney et al., 2008)
28 Leuschner and Sirocko (2000) and Voelker (2002) each demonstrate d that cooling episodes associated with Heinrich events also are found in both h emispheres and on every continent, though records pertaining dire ctly to Africa are especially scarce. Dansgaard Oeschger Events Dansgaard Oeschger ( D O ) events are cyclic warming events that occurred throughout the Last Glacial (Dansgaard et al., 1993; Grootes et al., 1993 ) Table 3 2 lists D O cycles occurring between 53 and 46 ka and these events are represented graphically in Figure 3 2 alongside the oxygen isotope stages. The beginning s of D O events are often punctuated H igh resolution records from Moomi Cave ( So cotra Island ) in the Indian Ocean indicate for example, that the transition from the arid H5 event ( ~ 48 ka) to D O 12 ( ~55 44 ka) occurred only within the span of 25 years (Burns et al., 2003; Burns et al., 2004) In northern high latitudes, at least, evidence also suggests dramatic temperature fluctuations at the onset of D O events. Thes e fluctuations are estimated between 8 C (DO13) and 16 C (DO19), but they seem to fall most commonly between 10 and 12 C (Wolff et al., 2010) Individual D O events occur on a roughly 1,470 year periodicity (Bond et al. 1997) H owever a larger pattern may be involved. Heinrich events often precede initial D O events followed by a triplet of subsequent D O events Each of the subsequent D O events generally appears smaller than the prior event indicating a gradual cooling over the subsequent 1 000 to 3 000 years leading up to the next Heinrich event (Burns et al., 2003; Huber et al., 2006; Cosford et al., 2008) This triplet pattern of D O events and subsequent Heinrich events support s the hypothesis for a stepwise recursive relationship between North ern Hemisphere and South ern Hemisphere oceanic circulation and phases of punctuated Laurentide ice sheet discharg e (Bond et al., 1997; Maslin et al., 2001)
29 West African and SW Asian Monsoonal System s Today, the highlands of the Horn of Africa receive the majority of annual precipitation from the African (Atlantic or West African ) boreal summer monsoon systems. The northerly movement of the ITCZ between July and August draws moisture rich air masses via the mid latitude westerlies from the Atlantic Ocean off the coast of NW Africa and across the continent (Figure 3 3 ). As these air masses hit the highlands of the Horn, the moisture is released as precipitation (Mohammed et al., 2004; Revel et al., 2010) In SW Ethiopia the orographic influence of the highlands on the mon soon air masses results in >2,000 mm annual average precipitation (Livingstone, 1975; Gasse, 2000; Mohammed et al., 2004) (Figure 3 4 ) The SW Asian ( Indian ) monsoons influence the highlands of the Horn in ot her ways. The SW Asian monsoons are part of the larger Asian monsoon system. T he East Asian monsoons derive their precipitation mainly from tropical western Pacific waters whereas the SW Asian water source is the Indian Ocean (Cosford et al., 2008) The Qinghai Tibet Plateau isolates the precipitation from the two monsoonal systems but these systems are both driven primarily by a land sea thermal gra dient during the boreal summer and the northward movement of the ITCZ (Burns et al., 2003) Recent isotopic research on meteoric water shows a latitu dinal gradient across the Horn that may relate to differences in precipitation source waters, hence monsoon systems. Freshwater lakes and groundwater from s outhern Ethiopia and Kenya appear to be more strongly influenced from an Indian Ocean water source than more northerly locations of the Horn (Kebede et al., 2005; Kebede et al., 2008; Kebede et al., 2009) These locations may therefore be more influenced by the SW Asian monsoons than the more northerly loca tions. Kebede (per comm.) hypothesized that past deviations in the annual intensity and pattern of the West African monsoonal system over the southern Ethiopian highlands may have been buffered by the SW
30 Asian monsoons. When discussing the SW Ethiopian h ighlands then, it is prudent to consider the fluctuation of both the African and SW Asian monsoons. The 18 O cave speleothem records from Xiabailong, Hulu, and Xiangshui caves in China (Wang et al., 2001; Cai et al., 2006; Cosford et al., 2008) Moomi Cave ( Socotra Island ) in the Indian Ocean ( Burns et al., 2003; Burns et al., 2004) and also Arabian Sea cor e s (Schulz et al., 1998) indicate that the SW Asian and East Asian monsoon systems were affected by climatic variability during the Late Pleistocene. Nile paleohydrology, West African Atlantic marine cores, and continental records provide a contemporaneous and synchronous account of African monsoonal variability during this time period (Lezine and Casanova, 1991; Flores et al., 2000; Weldeab et al., 2007; Revel et al., 2010) As numerous studies have noted, the timing and dur ation of the climatic variability evident from these records are in step with high latitude millennial scale Dansgaard Oeschger and Heinrich events found in the Greenland ice cores (Wang et al., 2001; Burns et a l., 2003; Burns et al., 2004; Cai et al., 2006; Cosford et al., 2008; Revel et al., 2010) The link between these records is that insolation increased during D O events changing the land sea thermal gradient and intensifying the monsoons from West Afric a to Western Asia. The Horn of Africa and the highlands of the Horn being uniquely positioned between the confluence of the African and Asian monsoons very likely experienced increased monsoonal precipitations during D O events that are in phase with the broader regional and hemispheric patterns. In summary, changing orbital conditions during the Pleistocene forced a succession of glacial and interglacial periods with a roughly 100,000 year periodicity. Among the most common climatic changes during the Late Pleistocene were the dramatic and rapidly occurring millennial scale events such as Heinrich stadials and Dansgaard Oeschger interstadials. These
31 events are found synchronously in high latitude ice core records and numerous multi proxy low latitude records. Dansgaard Oeschger events, in particular, may have affected regional monsoons systems across Africa and Asia. The influence of the African and Asian monsoons in the Horn of Africa today, and especially the SW Ethiopia highlands, indicates that this area likely experienced monsoonal variability during the Late Pleistocene The P aleoclimatic and P aleoenvironmental C ontext of Europe, Asia, and Africa B etween 65 and 43 ka This section synthesizes paleoclimates (PC) and paleo environments (PE) in Europe Asia, and Africa ~ 65 to 43 ka which covers the end of OIS 4 and the early and middle stages of OIS 3 Th is synthesis is subdivided into three subsections 65 to 55 ka 55 to 48 ka and 48 to 43 ka following the natural subdivision s proposed by Huber ( 2006:508) using D O events Accordingly, period 65 to 55 ka corresponds with D O1 8 15 and H6; period 55 to 48 ka corresponds with D O14 13 and H5; and period 48 to 4 0 ka corresponds with D O12 9 (cf. Wolff et al., 2010) I have structured each section so that the descriptions start from the broadest regional scale and move toward the local conditions in the Horn of Africa and, as best as pos sible, the SW Ethiopia highlands. OIS 4: 73.5 to 65 ka OIS 4 was a period of cold and arid glacial conditions across the N orthern Hemisphere. The environmental conditions across Africa during OIS 4 are often described using the better document ed glacial O IS 2 (27.8 14.7 ka) analog which was the last stage of the Last Glacial. Generally speaking annual temperatures across the continent were cooler, precipitation was lower, and vegetation zones contracted while deserts expanded. In the Horn of Africa, l ake cores show that mean annual temperature was up to 5 C cooler during OIS 2 than today (Gasse, 2000) Lower temperatures also led to decreased precipitation
32 rates The water levels at many of the rift valley lakes dropped significantly. Lake Tanganyika dropped up to 300 m, Lake Mobuto Sese Seko (Albert) was 46 m lower, while Lake Victoria and Lake Abhe (Ethiopia) may have been desiccated (Gasse, 2000) Across North Africa, the expansion of the Sahara desert during OIS 2 followed widespread desiccation of the Last Interglacial grasslands (Clos e et al., 1996) Decreased discharge and increased sediment load in the Atbara river and the Blue Nile and White Nile resulted in the blockage and occasional cessation of these riverine systems (Wendorf and Schild, 1989; Close et al., 1996) In Central Africa, the once widespread tropical rainforests fragmented and were replaced by patchy woody grasslands (Cornelissen, 2002) OIS 4 to OIS 3: 65 to 55 ka Northern Hemisphere climates were becoming increasingly unstable by the end of OIS 4. The GISP2, GRIP, and NGRIP 18 O ice core s show that across the Northern Hemisphere high latitude cold and arid condit ions during H6 ( ~60 ka) w ere followed in rapid succession by D O events 17 (~59 ka) and 16 (~58 ka) (Grootes et al., 1993; Sowers et al., 1993; Meese et al., 1994; Stuiver and Grootes, 2000; North Greenland Ice Core Project, 2004) North Atlantic sea surface temperatures ( SST s) from core CH73 139c corroborate these findings (Labeyrie et al., 1991) These climatic fluctuations appear to have had widespread influence on local paleoenvironments across Europe Atlantic marine core records (MD95 2042; SU81 18; MD99 2331; MD04 2845) show a rapid expansion of the Mediterranean forest regime in SW Iberia during D O 1 7 to 1 6 (Sanchez Gone et al., 2008) The expansion in Mediterranean forest cover in Europe during D O events indicates that the climatic pattern s were likely reminiscent of the current Mediterranean climate with ho t and dry summers and winter rainfall (Sanchez Gone et al., 2008) Similar influence s in winter rainfall patterns are found also in the Soreq cave ( Israe l) speleothem records These records
33 show p eriodic low 18 O and 13 C values that are correlated to sapropel events in the e astern Mediterranean Sea (Bar Matthews et al., 2000) S apropels are pluvial events that reflect changes in Mediterranean Sea SST or salinity due to i ncreased rainfall and freshwater riverine runoff. Bar Matthews et al. (2000) believed that the sapropels and fluctuations in Soreq cave 18 O and 13 C are influenced by monsoonal activity. Similarly, total organic carbon (TOC) records from Arabian Sea co res (111KL; 136KL) show fluctuations in TOC that are consistent with the pattern of sapropel, D O, and H climatic shifts during this period (Schulz et al., 1998) Schulz et al. (1998) identified the se TOC rich bands to show greater precipitation that was likely due to increased monsoonal activity. Early OIS 3: 55 to 48 ka The intensit y of D O events 17 to 15 declined sequentially after H6. By D O 15, the magnitude of annual temperature increases above Greenland may have been only 9 to 10 C above current conditions (Wolff et al., 2010) However, D O 14 (~54 ka) appears to be as strong as D O 17 with a 12.5 C mean annual temperature spike (Wolff et al ., 2010) Atlantic marine core records (MD95 2042; SU81 18; MD99 2331; MD04 2845) show that mid Atlantic forest s across NW Iberia and w estern France expanded significantly during D O 14 (Sanchez Gone et al., 2008) Mediterranean forest s however, do not show significant changes implying that the effects of millennial scale events across w estern Europe were also strongly influenced by regional influences including the varying effects of orbital orientation (Sanchez Gone et al., 2008) The e astern Mediterranean Sea and Arabian Sea were also experiencing wetter conditions during this period. At Soreq cave, low 18 O and 13 C speleothem values at 54 ka may correlate to the poorly studied s apropel 2 event (Bar Matthews et al., 2000) This time is coincident to D
34 O 14. A high lake level stand from Lake Lisan, the precursor to t he current Dead Sea, corroborates these findings (Neev and Emery, 1967; Stein et al., 1998; Bar Matthews et al., 2000) Total organic carbon (TOC) records from the Arabian Sea provide a particularly clear seq uence of D O events during this period. 2 TOC peaks are present at ~56 ka (D O 16), ~53 ka (D O 15), ~51 ka (D O 14), and ~47 ka (D O 13) (Schulz et al., 1998) The very sudden and large increase in TOC at the onset of D O 14 is analogous to the similarly identified D O 14 event above Greenland where a 12.5 C mean annual temperature sp ike occurred (Wolff et al., 2010) While the TOC records do not record paleothermometry, the similarities between D O 14 in the Arabian Sea and above Greenland lend credence to the hypothesis that similar temperature shifts were occurring between these two areas during D O events. What do these patterns reveal about monsoonal variability at this time? The Arabian TOC records and speleothems from Moomi cave and Xiaobailong cave indicate coupling between millennial scale interstadial (i.e. D O) events and the SW Asian monsoons. Schulz et al. (1998) believed the TOC records to be strongly correlated to monsoonal variability. O rganic rich bands (high TOC) are for med during periods of monsoon induced biological productivity. The 18 O speleothe m records from Moomi cave (Socotra Island) and Xiaobailong cave (SW China) both show isotopic changes relating to precipitative influences de rived from the SW Asian summer monsoon s (Burns et al., 2 003; Burns et al., 2004; Cai et al., 2006) Thus, during this period, SW Asian monsoonal variability was clearly linked to D O events in the circum Indian Ocean region. Proxy records pointing to West African monsoonal variability show similar patterns to SW Asian monsoonal variability, though the records are generally of lower resolution. Continental 2 Discrepancies occurr between D O ages in the TOC records and other records, such as Wolff (2010) because the TOC records us e an orbitally tuned age model based on layer counting.
35 North Africa records indicate a broad humid phase from 52 to 44 ka attributed to increased West African monsoonal activity (Lezine and Casanova, 1991) Sediment cores from the coast of West Africa suggest a period of high riverine runoff ~55 ka that is also interpreted to indicate increased West African monsoons at this time (Weldeab et al., 2007) This time period is approximate to D O 15. Recent and unpublished 18 O records from the Gulf of Guinea, however, may contradict the decoupling hypothesis of the monsoons proposed by Weldeab (2007) These newer records show that West African monsoonal variability is tightly linked to Greenland temperature instabilities during OIS 3 (Weldeab per comm.). The pattern of sapropel events correlated to the Soreq cave records shows the influence of millennial scale climatic variability on the West African monsoons in the eastern Mediterranean region (Bar Matthews et al., 2000) L ow 18 O and 13 C values at 54 ka are coincident to sapropel 2 and D O 14. Most importantly, the formation of Mediterranean sapropels is linked to increased precipitation in the Nile source region, the Ethiopian highlands ( Bar Matthews et al., 2000) Nile paleohydrology records provide another independent data source that is also directly correlated to the Ethiopian highlands. These records show a broad period of pluvial conditions from 60 to 50 ka that are interpreted as increased monsoonal activity in the Ethiopian highlands at this time (Revel et al., 2010) Taken together, since the beginning of OIS 3, and certainly from 55 to 48 ka, a series of D O events in the Northern Hemisphere occurred that had a clear impact on the intensity of both the African and SW Asian monsoons systems. This impact included large magnitude temperature fluctuations of D O events above Greenland to concurrent increases in regional monsoonal precipitation across North Africa, the eastern Mediterranean Sea, the circum Indian Ocean, and finally the Horn of Africa. No concurrent terrestrial records exist from the SW
36 Ethiopian highlands dating to this period. Howe ver, these data provide strong supporting evidence that regional monsoonal variability changed the precipitation patterns across the Horn of Africa and, by proxy, SW Ethiopia in sync with D O events. Most importantly, I believe these precipitation changes almost certainly impacted the local ecology in the Horn and SW Ethiopia. Early Mid OIS 3: 48 to 43 ka B y 48 ka the glacial aridity of H5 w as present across Europe, Arabia, Asia, and North and East Africa. Multiple Greenland ice cores each show severely d epleted 18 O between ~48 and 47 ka (Grootes et al., 1993; Greenland Ice core Project, 1993; Stuiver and Grootes, 2000; North Greenland Ice Core Project, 2004) Atlantic Ocean SSTs adjacent to w estern Europe w ere >10C cooler than prior periods, and p ollen records indicate contractions of the European mid Atlantic forests and the Mediterranean forests (Sanchez Gone et al., 2008) The Lago Grande di Monticchio se diment cores from s outhern Italy similarly show decreases in the relative abundance of woody taxa at this time (Allen et al., 1999) The TOC records from the Arabian Sea sediment cores suggest similarly cool conditions at this time. These records show a ~1.5 ka period of drastically lower TOC pterapod rich sediments implying colder conditions during H5 (Schulz et al., 1998) The SSTs der ived from marine cores near the Eastern Maldives (MD90963) and Somalia (TY93929/P; MD85668; MD85674), and eastern Mozambique (MD79257) are also lower (Bard et al., 1997; Rostek et al., 1997; Bard, 2003) Climatic conditions improved quickly at the onset of D O 12, 46.8 ka. In France, the Villars cave 18 O speleothem shows the highest growth rate s and greatest variation in 18 O at this time (Genty et al., 2003) The findings indicate a sudden shift from cooler to warmer and more humid cond itions on par with a 10C (50F) temperature increase during D O 12, which is similarly seen in other marine and terrestrial records at this time (Guiot et al., 1993; Cayre et al.,
37 1999; Genty et al., 2003) C oncurrently, Atlantic forests across western France and NW Iberia expand at this point (Sanchez Gone et al., 2008) while woody taxa increased in abundance across southern Italy (Allen et al., 1999) The D O 12 event had a similar impact across the circum Indian Ocean region. T he Moomi cave 18 O speleothem record indicates gradually improving climatic conditions after 50 ka until 47.2 ka when a rapid enrichment of 18 O occurs during D O 12. The Moomi M1 2 speleothem was sampled in 8 year increments acros s this period, and the resolution of this record shows that the shift from cold and arid H5 to the warmer and wetter D O 12 occurred within 25 years (Bu rns et al., 2003) Mean annual temperatures above Greenland increased 12 C during D O 12 (Wolff et al., 2010) Around 43 ka, the D O 11 events are evident i n both high latitude ice cores and regional records surrounding the Horn of Africa. Above Greenland, D O 11 is associated with the greatest increase in mean annual temperatures known during OIS 3, 1 5 C (Wolff et al., 2010) However, for reasons that are still unclear, the D O 11 event is poorly seen in the Arabian Sea TOC records (Schulz et al., 1998) and Xiaobailong speleothem records (Cai et al., 2006) Nile paleohydrology records indicate a broad per iod of pluvial conditions at this time with flood maxima from 38 to 30 ka (Revel et al., 2010) Continental North Africa records show a similar broad pluvial event from 52 t o 44 ka (Lezine and Casanova, 1991) Conclusion s What do these data imply about ecology changes and hunter gatherer adaptation in SW Ethiopia during OIS 3? First, a clear relationship is evident between high latitud e ice core records, mid latitude terrestrial and marine records, and low latitude tropical marine and terrestrial records from across Africa and the circum Indian Ocean region. These similarities may extend to parallels between the paleothermometry of the se regions during D O events,
38 which would also help to explain the increase in monsoonal activities during these times. If the teleconnection that I have laid out between regions and records holds true, then D O 12 (46.8 ka) and D O 11 (43.3 ka) almost ce rtainly had a major impact on the Horn of Africa. Mean annual temperatures are estimated to have increased up to 12 C during D O 12 and up to 15 C during D O 11 (Wolff et al., 2010) During D O 12, these changes may have taken place within 25 years (Burns et al., 2003; Burns et al., 2004) Even if the magnitude of these changes was ameliorated in lower latitudes, it is still very likely that the intensity and rapidity of these two D O events, in particular, had appreciable effects on the climate across the circum Indian Ocean region and the Horn of Africa Second, a clear relationship is evident between millennial scale climatic instability, seen in high mid and low latitude records, and monsoonal variability. P aleoclimates (PC) and paleo e nvironments (PE) records from the coast of West Africa, North Africa, the Mediterranean Sea, and the circum Indian Ocean region all show that pluvial events, including D O events and sapropels, are associated with coeval increases in African and SW Asian m onsoonal intensity and precipitation. The Moomi cave record and Arabian Sea TOC show SW Asian monsoonal flux during D O events (Schulz et al., 1998; Burns et al., 2003; Burns et al., 2004) Marine core record s from the Gulf of Guinea show West African monsoonal flux during D O events (Weldeab et al., 2007) Unpublished 18 O data from the same location are also tightly linked to temperature changes above Greenland during OIS 3 (Weldeab per comm.). The fluctuations in both the African and SW Asian monsoons appear in sync with each other and, significantly, in sync with te mperature flux above Greenland during H and DO events of OIS 3. Most importantly, however, we must consider the unique position of the Horn of Africa and the highlands in the Horn. Today, the distribution of precipitation across the Ethiopian
39 highlands is controlled by orographic influences (Mohammed et al., 2004) Similar orographic forces must have controlled the spatial distribution of rains across the highlands during OIS 3. Therefore, the SW Ethiopian highlands must have had relatively greater amounts o f precipitation during periods of monsoonal intensity of OIS 3. The Mediterranean records from Soreq cave and the Nile paleohydrology records provide clear evidence to support this idea (Bar Matthews et al., 20 00; Revel et al., 2010) The only way to increase Nile River discharge naturally is at the source locations, in the Ethiopian highlands and Lake Victoria. Third, the increase in monsoonal precipitation during D O events of OIS 3 almost certainly had an effect on local environments of the Horn and hunter gatherer resources. In particular, f luctuations in the monsoon rains would alter the distribution of the vegetation zones around Moche Borago. During D O events, I believe that more diverse and abundan t resources were available around Moche Borago. The montane woodlands would have probably moved toward lower elevations, more marshy grasslands might have appeared on the plains below the site, and Lake Abaya almost certainly was present. Therefore, it i s during pluvial times, and, in particular, D O 12 (46.8 ka) and D O 11 (43.3 ka) that I expect to find changes in stone artifacts at Moche Borago as the local hunter gatherer populations adapted to more abundant resources.
40 Figure 3 1: P aleoclima tic proxy sites across the circum Indian Ocean area that are described in the text
41 Table 3 1: Description of Paleoclimatic Sites in Figure 3 1 No. Location Core number Latitude Longitude References 1 Central Greenland GISP2 72.6 38.5 Grootes et al. 1993; JGR special volume and CD 2 Central Greenland GRIP 72.58 37.63 Dansgaard et al. 1993; Johnsen et al. 1992; JGR 3 Greenland NGRIP 75.1 42.3 Johnsen et al. 2001; NGRIP members 2004; http://www.glaciology.gfy.ku.dk/ngrip/; Johnsen et al. 1992; H ansson 1994 4 Villars Cave, France Vil 9 45.3 0.5 Genty et al. 2003; Grootes et al. in prep. 5 Soreq cave, Israel speleothems 31.7 35 Bar Matthews et al. 1999 6 Lake Lisan, Paleo Dead Sea, Perazim valley out 31.21 35.31 Schramm 1997; Schramm et al. 20 00; 7 Hulu Cave, China PD, MSD, MSL, H82 and YT 32.5 119.17 Wang et al. 2001 8 Kashiru, Burundi, Africa Ka 1 3.45 29.53 Bonnefille et al. 1990, 1992; Bonnefille and Chalie 2000 9 Arabian Sea SO42 57KL 20.9 63.12 Leuschner et al. 2000; 10 Arabian Sea off Oman RC27 14 18.25 57.66 Altabet et al. 2002 11 Arabian Sea off Oman RC27 23 17.99 57.59 Altabet et al. 2002 12 Arabian Sea off Oman ODP site 723B 18.05 57.61 Altabet et al. 1999, 2002 13 Arabia Sea SO90 88KL 23.1 66.48 Shulz et al. 1998 14 Arabi a Sea SO90 111KL 23.58 64.22 Shulz et al. 1998 15 Arabia Sea SO90 136KL 23.12 66.5 Shulz et al. 1998 16 Arabian Sea Piston core 905 10.77 51.95 Klocker et al. 2006 17 Atlantic Ocean MD95 2042; SU81 18 37 10 Sanchez Goni et al. 2008 18 Atlantic Oce an MD99 2331; MD03 2697 42 9 Sanchez Goni et al. 2008 19 Atlantic Ocean MD04 2845 45 5 Sanchez Goni et al. 2008 20 Namib Desert 27 21 Stokes et al. 1997 21 Atlantic Ocean GIK 16776 3.73 11.4 Dupont et al. 2000 22 Atlantic Ocean GIK16856 4.8 3. 4 Dupont et al. 2000 23 Atlantic Ocean GIK16876 2.2 5.1 Dupont et al. 2000 24 Atlantic Ocean GeoB1008 6.58 10.32 Dupont et al. 2000 25 Atlantic Ocean GeoB1016 11.77 11.68 Dupont et al. 2000 26 Atlantic Ocean GIK16415 9.57 19.1 Dupont et al. 200 0 27 Atlantic Ocean GIK16416 9.9 19.38 Dupont et al. 2000
42 Table 3 1 : Continued Note : For added sites and descriptions, see Voelker (2002) No. Location Core number Latitude Longitude References 28 Atlantic Ocean KS84067 4.13 4.12 Dupont et al. 2000 29 Atlantic Ocean KS12 3.87 1.93 Dupont et al. 2000 3 0 Atlantic Ocean KW23 3.77 9.28 Dupont et al. 2000 31 Atlantic Ocean V22 196 13.83 18.95 Lezine and Casanova 1991 32 Atlantic Ocean M12392 1 25.17 16.83 Lezine and Casanova 1991 33 Atlantic Ocean MD03 2707 2.5 9.39 Weldeab et al. 2007 34 N. Soma lia 11 49 Brook et al. 1995 35 S. Somalia 3 44 Brook et al. 1995 36 Lake Abhe 11 42 Gasse 1977 37 Lake Masako M96A 9.33 33.75 Barker et al. 2003 38 Indian Ocean MD85 668 0.02 46.04 Meynadier et al. 1992 39 Indian Ocean MD85 669 2.49 46.92 M eynadier et al. 1992 40 Indian Ocean MD85 674 3.19 50.44 Meynadier et al. 1992 41 Oman margin RC27 14 18.25 57.66 Altabet et al. 2002 42 Oman margin RC27 23 17.99 57.59 Altabet et al. 2002 43 Moomi cave, Socotra Island M1 2 12.5 54 Burns et al. 2003 44 Kilwa cave system 8.93 39.35 Lundblad and Holmgren 2005 45 Songo Songo Island cave system 8.42 39.48 Lundblad and Holmgren 2005 46 Tanga cave system 5.03 39.08 Lundblad and Holmgren 2005 47 Xiangshui Cave X3 110.92 25.25 Cosford et al. 2008 4 8 Xiaobailong Cave XBL 1 24.2 103.35 Cai et al. 2006 49 Medite4rranean Sea MS27PT 31.81 29.47 Revel et al. 2010 50 Sierra Leone Rise CAMEL 1 5.11 21.04 Flores et al. 2000
43 Figure 3 2: Timeline of the Last Glacial showing the boundar ies of Oxygen Isotope Periods and Dansgaard Oeschger events
44 Table 3 2 : List of Dansgaard Oeschger E vents from the Last Glacial to the P resent Event Age Range, BP Temperature I Age Reference Temperature Reference Younger Dryas 11,513 11,613 10 4 Rasmussen et al. 2006 Grachev and Severinghaus 2005 1 14,549 14,735 11 3 Rasmussen et al. 2006 Grachev and Severinghaus 2005 2 22,992 23,588 Andersen et al. 200 6 3 27,314 28,146 Andersen et al. 2006 4 28,401 29,299 12 5 Andersen et al. 2006 Sanchez Goni et al. 2008 5 31,884 33,016 7 5 Andersen et al. 2006 Sanchez Goni et al. 2008 6 33,084 34,296 7 3 Andersen et al. 2006 Sanchez Goni et al. 2 008 7 34,769 36,091 9 ( +3; 6) Andersen et al. 2006 Sanchez Goni et al. 2008 8 37,445 38,895 11 ( +3; 6) Andersen et al. 2006 Huber et al. 2006 9 39,320 40,900 9 ( +3; 6) Andersen et al. 2006 Huber et al. 2006 10 40,593 42,227 11.5 ( +3; 6) And ersen et al. 2006 Huber et al. 2006 11 42,422 44,158 15 ( +3; 6) Svensson et al. 2008 Huber et al. 2006 12 45,854 47,766 12 ( +3; 6) Svensson et al. 2008 Landais et al. 2004b 13 48,215 50,245 8 ( +3; 6) Svensson et al. 2008 Huber et al. 2006 14 5 2,974 55,366 12.5 ( +3; 6) Svensson et al. 2008 Huber et al. 2006 15 54,494 57,006 10 ( +3; 6) Svensson et al. 2008 Huber et al. 2006 16 56,943 59,517 9 ( +3; 6) Svensson et al. 2008 Huber et al. 2006 17 ca. 59,390 12 ( +3; 6) Svensson et al. 2008 Huber et al. 2006 18 ca. 64,045 11 2.5 Estimated Landais et al. 2004a 19 ca. 72,280 16 2.5 Estimated Landais et al. 2004a 20 ca. 76,400 11 2.5 Estimated Landais et al. 2004a Note: This chart follows the definitions of D O events after Wolff et a l. 2010.
45 Figure 3 3 : Generalized wind patterns of the summer monsoon systems in Africa and SW Asia.
46 Figure 3 4 : Contemporary mean annual precipitation rate across the Horn of Africa. Data credit: FAO / UNEP Desertification and Mapping Project
47 CHAPTER 3 A BRIEF HISTORY OF T HEORETICAL IDEAS ABOUT HUNTER GATHERER BEHAVIORAL VARIABILI TYAND ARID CLIMATIC EVENTS DURING THE LA TE PLEISTOCENE This chapter briefly reviews the history of theoretical thought that has framed scholarly inquiry into the re lationship between Late Pleistocene human behavior variability in Africa and environmental change. Since the early 20th century, much of this research has focused on the relationship between arid climatic events and human behavior change. I contend that arid adaptation ideas are popular to this day because of the theoretical history and also because we know with greater certainty that glacial climates were most common during the Middle and Late Pleistocene. Modern Human Behaviors Just as with any other natural organism, humans rely on climates to regulate our environment, subsistence resources, and even comfort. We differ from other organisms, however, in the complex array of modern behaviors developed since the Middle Pleistocene to modify our environm ents to our benefit. These behaviors broadly include new social, technological, and symbolic tools that allowed hunter gatherers the ability to contend with a wider range of environments and exploit resources more effectively and efficiently (McBrearty and Brooks, 2000 ; Marean and Assefa, 2005) Several overviews of modern human behaviors have been discussed about how these behaviors can be identified materially in the archaeological record (McBrearty and Brooks, 2000; Wad ley, 2001; Henshilwood and Marean, 2003; Marean and Assefa, 2005) The main issues in these discussions relate to concerns about methodologies to identify the behaviors in archaeological contexts, as well as the ontology of the behaviors, regarded both i ndividually and as a group package. McBrearty and Brooks (2000) for example, provided the most detailed
48 synthesis to date about behavioral modernity, including the list of traits that may define behavioral modernity in African contexts, the timing and speed of the behavior changes, and the relationship between the African and European archaeological records. Henshilwood and Marean (2003) however, questioned the widely used trait list approach and the empirical tests employed to assess behavioral modernity in the archaeological record. Wadley (2001) argued that intangible symbolic systems a re the indisputable evidence for modern behavior, and she suggested several methods to identify symbolism in material culture. I view modern behaviors, as the y were developed, as a buffer to protect humans from the uncertainties of the natural world. On the one hand, this buffer can be envisioned materially. Such technologies as air conditioners and air heaters create artificial and stable environments. The bow and arrow would also fall into this category. Projectile hunting weapons likely provided a security factor when firing an arrow at a distance from potentially dangerous prey. On the other hand, the buffer can also be conceptualized less tangibly. Ex change systems provide a multi locus insurance for resources. Symbol ic systems also help to make sense of the unpredictable and inexplicable events in the natural world. As modern humans, we have developed new social and technological behaviors to ultima tely detach ourselves from the unpredictable natural world to ensure our resources and the security of our offspring. Late Pleistocene Behavioral Variability in the Archaeological Record H unter gatherers may have adapt ed their behavior to varying ecolo gical conditions brought about by glacial or interglacial phases during the Middle and Late Pleistocene For example, the variable use of marine resources in relationship to sea stand and coastline fluctuations during the Middle Pleistocene in South Afric a shows context dependent hunter gatherer adaptations (Henshilwood et al., 2004; Jacobs et al., 2006; Marean et al., 2007; Fisher et al., 2010) T he highly mobile subsistence strategies developed by Aterian hu nter gatherers in the Saharan
49 grassland during the Last Interglacial (128 75 ka) period of the Late Pleistocene are other examples of the unique behavioral adaptations of people in Africa during this time (Garce a and Giraudi, 2006; Barton et al., 2009) Furthermore, m any of the behavioral changes seen during the Middle and Late Pleistocene appear to be associated with innovations among African hunter gatherer populations to create and maintain social relations hips (McBrearty and Brooks, 2000; Wadley, 2001; Ambrose, 2002; Henshilwood and Marean, 2003; Marean and Assefa, 2005) These relationships may have provided a self maintained and multi locus system to minimize any risk associated with living in diverse or degraded environmental conditions often hypothesized for arid, glacial periods in Africa (Ambrose, 1998a; Ambrose, 1998b; Ambrose, 2002; Ambrose, 2003) Adaptat ion to local environmental conditions may also reflect the development of cultural patterning in stone tool technologies across Africa during the Late Pleistocene (Clark, 1964; Clark, 1988; Brooks et al., 2006) The social and technological behaviors, which are first seen in Later Stone Age (LSA) hunter gatherers during the Late Pleistocene in Africa, are also part in parcel of the greater development of modern behaviors. Generally speaking, LSA behaviors incl ude broadening social and symbolic systems and expressions, expanded habitat and resource use, and new ways of making and using stone artifacts (McBrearty and Brooks, 2000) Cultural pat terning in stone tools may also relate to genetic and cultural isolation due to catastrophic climatic events (Ambrose, 2002; Ambrose, 2003) Arid Adaptation Ideas Since the Early 20th century The earliest known anatomically modern humans (White et al., 2003) and much of the earliest evidence for behavioral modernity are now known to date to glacial MIS 6 (195 128 ka) (McBrearty and Brooks, 2000) The glacial interglacial cycling that followed has been seen by
50 some researchers to be a major influence on modern human behavioral adaptations (Clark, 1988) Scholars in the first half of the 20 th century were aware of a similar pattern of glacial interglacial cycling during the Pleistocene, but they were at a disadvantage from the lack of accurate chronological information available after the advent of radioca rbon dating in the 1950s. In the second half of the 20 th century, radiocarbon and potassium argon dating enabled researchers to emplace arid adaptation ideas within chronological frameworks. The result was the development of a generalized sequence of cul 1900 1940 As early as 1901, Max Blanckenhorn in Neues zur Geologie un Paleontologie Aegyptens, IV denote periods of increased rainfall (Blanckenhorn, 1901) At the same time, other German scholars had developed a theoretical teleconnection between glacial (Flint, 1959) According to these ideas, mid and low latitude heating increased evaporation, which caused increased amounts of snowfall and cooler temperatures in high er latitudes such as in Europe. Conversely, glacial conditions in Europe were also believed to cause increased precipitation in Africa because storms were deflected toward Africa (Flint, 1959) In 1914, Brooks adopted the based scholarship (Brooks, 1914) most significant influence on pluvial research in Africa at this time. Way land used river gravel deposits to establish a series of two pluvial events (Kamasian and Gamblian) and an epi pluvial event in Uganda, which he then used to correlate to inter pluvial (glacial) events in Europe (Wayland, 1929; Wayland, 1930; Wayland, 1934)
51 was highly influential on the work of other early Africanists including Louis Leakey (Leakey and Solomon, 1929) and Erik Nilsson (1929) This is because the pluvial method was more than just a means for relative dating. I t was also one of the first methods to tie perceived cultural changes seen in the African Stone Age archaeological record to a paleoenvironmental record (Wayland, 1929; Wayland, 1930; Wayland and Burkitt, 1932; Wayland, 1934) Key to this chronology was the link between culture change and the environment. Wayland (1934) suggested that perceived rapid cultural changes were associat ed with arid (inter pluvial) events Examples, of these rapidly changing cultural industries include the Chellean, Aurignacian, Capsian, and Stillbay. 1940 Present Day At the First Pan African Congress on Prehistory in 1947, the nomenclature of the Plei (N ilsson, 1949) The biggest development at this time, however, was the advent of absolute dating methods. I n the 1950s radiocarbon dating confirmed the relative timing of the latest pluvial events (Arnold an d Libby, 1949; Libby and Arnold, 1950; Libby, 1951; Flint, 1959) Around this same time, Cole (1954) provided an updated account of the pluvial hypothesis reiterating the links between arid, inter pluvial environments and rapid cultural change in African Stone Age archaeology According to Cole (1954) exhibiting rapid changes in the frequencies o f types of stone artifacts or techniques used to make stone artifacts. This phrase was not a cultural chronological or even a cultural spatial term, such
52 rapid cultural change (inferred from stone tool technology), and for this reason Cole was able to dustries were all dated to inter (1954) pluvial events. Second, and less obvious, is the i cultural change in Africa. It has only been within the last decade, and following much intensive one Age) was found to be neither rapid nor punctuated (McBrearty and Brooks, 2000) By the 1960s, the pluvial hypothesis was discredited due largely to the efforts of Flint (1959) who argued that the pluvial method was unsupported by existing evidence. He also stated that climates were an unsuitable primary basis for a geo stratigraphic sequence. C ultur e ecology became the next dominant t heoretical grounding in African Stone Age archaeological research at this time As part of the broader processualist movement, cultur e ecology studies integrated ethnographic and environmental research within archaeological investigations The purpose of this integration was to study the systemic processes of humans past and present adapting to their environment using an explicit and positivist scientific methodology (Binford, 1962; Binford, 1965; Trigger, 1989) Clark (1964) was among the most ambitious and influential proponents of the early cu lture ecology approach. He advocated a direct relationship between environmental change and human were responsible for the basic differences in the stone cu (White, 1959; Clark, 1964:177)
53 Furthermore, and perhaps more than any other mid 20 th century researcher, Clark championed the link between arid environments and rapid behavioral change by suggesting that the challenges of early humans were wrought by environmental deterioration and met by the development of new behaviors (Clark, 1960; Clark, 1963; Clark, 1965; Clark, 1988) For example, Clark (1960: 309 310) suggested : -times of relatively short duration compared with the intervening wet periods, when new ideas and new forms were able to spread with greater ease throughout the continent and when less f towards improved methods of securing food and more comfortable living quarters lanation of the historical development of human society]. On the other hand, as soon as technical ability permitted [that is to say from the end of the Earlier Stone Age onwards], the long periods of wetter climate made for stability, slow development, an d isolation of groups, and so resulted in a number of contemporary regional cultural variants. Conclusions Today, arid adaptation ideas can be found in the resurgence of special interest discussions concerning African cultural environme ntal refugia (Brandt, 1988:840; Shea, 2006; Brandt et al., 2006; Cohen et al., 2007; Hildebrand et al., 2008) and past climatic catastrophes such as the Toba super eruption in SE Asia (Ambrose 1998a, 1998b, 200 3; Rose and Chesner 1990; but see Oppenheimer 2002; Gathorne Hardy and Harcourt Smith 2003) However, a problem with studying just arid adaptation is that it may be only one perspective of a much larger phenomenon. This phenomenon was recognized by Clark (1988:251) who also implied that Last Interglacial hunter gatherers were as adapted to their environment as were hunter gatherers in glacial periods. Furthermore, the ethnographi c record provides a vital
54 account of how contemporary hunter gatherers adapt to arid or unpredictable environmental conditions. A number of studies have addressed hunter gatherer behavioral variability, and both Lee and Devore (1968) and Kelly (1995) have provided thorough descriptions of this research. In Africa, some of the seminal studies of hunter gatherer living outside of tropical forests include (1977; 1980; 1982) (1968) s (2010) research among the Hadza (Tanzania). Kelly (1995) provided a detailed discussion of the various 20 th Century theoretical approaches in anthropology that have been used in hunter (1980) seminal study showed that hunter gatherer organization around resources can be projected onto, or interpreted from, archaeological assemblages. Yellen and Harpending (1972) have shown that !Kung band aggregation and cultural homogeneity can be attributed to resource availability, which is also observable in the archaeological material culture record. According to Yellen (1977) however, it is the predictability of resources in an area rather than resource diversity that can affect hunter gatherer behavior. In variable and unpredictable conditions, like deserts, Yellen (1977) finds that people living in these areas proffer a flexible behavioral strategy that enables them to adapt to, and exploit, numerous environments and available resources. People living in variable and unpred ictable conditions are also more resistant to change. Behavioral conservatism, therefore, is a product of successful adaptation to unpredictable environments, which contradicts many ideas about behavioral change to arid and glacial times in the Late Pleis tocene archaeological record. In environments where resources are unpredictable and highly variability, ethnographic and archaeological cultures should evidence consistent and rarely changing behaviors and
55 material culture. But, what might be expected during OIS 3? The climate during OIS 3 may have been unstable and resources in the Horn of Africa might have become unpredictable. Under these conditions a more conservative material culture archaeological record could be expected following a period of adaptation to the new environmental conditions. However, the logic here assumes that the environmental changes from good to bad, hence from resource abundance to a lack of resources and unpredictable resources. More intense monsoons during OIS 3 almost c ertainly would have introduced a degree of unpredictability into the environment, but the added moisture may have also helped to increase resource diversity and abundance in the area. Therefore, increased unpredictability coupled with increased resources may have created a situation where it was not only advantageous for hunter gatherers to diversify into new resource niches but it was timely to experiment with new ways to more effectively collect diverse resources. To develop this idea further it is use ful to look towards the available archaeological data, which will be discussed in the following chapter. Here, I have briefly reviewed the history of theoretical thought about arid adaptation ideas in African Stone Age research since the early 20 th centur y. I have focused intently on these ideas because this genre of thought has been such a predominant force in African Stone Age archaeology throughout much of the last century to today.
56 CHAPTER 4 ARCHAEOLOGICAL SITES IN EASTERN AFR ICA DATING TO THE LATE PLEISTOCENE (OIS 4 AND OIS 3) This chapter summarizes the archaeological data about hunter gatherer occupation of the Horn of Africa and East Africa during OIS 4 and OIS 3. The OIS 3 archaeological record, in particular is minimal, being characterized b y large spatial and temporal gaps in the data. Many of these sites are also poorly dated with the result that little data currently are available to discuss how hunter gatherers were living in this region during this period. OIS 3 Archaeological Record: Limitations Two reasons for the dearth of archaeological data in eastern Africa during OIS 3 include chronology and sampling issues and prehistoric human behavior. The first problem, chronology and sampling, relates not only to the historical lack of rese arch in this area and in this period in general (cf. Brandt, 1986) but also to th e lack of dating techniques that have been, until recently, available to date sites of Late Pleistocene age. As both Clark (1954) (1986) reviews of the Horn show, a number of sites are known in the region but many still remain undated. The second reason draws upon current estimates for low population densities across Africa during the Last Glacial (Ambrose, 1998b) and the likelihood that human populations across this region were located in only specific localities during glacia l periods (Brandt et al., 2006; Hildebrand et al., 2008; Basell, 2008) Chronological and Sampling Problems Radiocarbon dating has been widely used to date Holocene and Late Pleistocene sites, but apart from this technique, other methods have not been available to date Late Pleistocene sites until relatively recently. Obsidian hydration used to be fairly common, but it is now known to be problematic, if not unreliable entirely The complicating factor for o bsidian hydration is due to the variability of hydroxyl ions within the obsidian (Mazer et al., 1991; Anovitz et al., 1999;
57 Hull, 2001; Rogers, 2008) Other techniques had minimum age limits in excess of the L ate Pleistocene, such as Potassium Argon dating. Today, new techniques, such as Argon Argon and optically stimulated luminescence dating, provide more options for researchers but even these techniques are dependent upon specific environments and sample ty pes. The use of radiocarbon has sustained Late Pleistocene archaeology in eastern Africa up to the present day. However, even radiocarbon has had limitations when dealing with Late Pleistocene aged materials. Until relatively recently, a ~40 ka cap on t he maximum age limit has been in place for radiocarbon analysis because no calibration curves existed beyond this time period. Radiocarbon ( 14 C) is a product of atmospheric interactions with cosmic rays and geomagnetism, as well as fluctuations in the gl obal carbon reservoirs (Hughen et al., 2004) Several notable dev iations of cosmogenic nuclides ( 14 C, 10 BE, 36 CL) occurred during the Late Pleistocene due to solar activity and geomagnetic flux (Baumgartner et al., 1998; Hughen et al., 2004; Muscheler et al., 2005; Chiu et al ., 2007) Within the last decade new calibrations have been developed using independent dating methods such as U series on marine fossil coral and speleothem records t hat extend the calibration curves up to 50 to 60 ka (Hughen et al., 2006; Chiu et al., 2007; Weninger and Joris, 2008; Reimer et al., 2009) The 40 ka cap not only limited the usefulness of radiocarbon dating, but it also required researchers to publish only uncalibrated radiocarbon ag es from m any Late Pleistocene sites in eastern Africa (cf. Clark, 1954; Mehlman, 1989) To avoid confusion, extra vi gilance is now required when discussing these ages to differentiate between calibrated and uncalibrated dates. M any dated Late Pleistocene archaeological sites in Africa were also excavated prior to the development of improved collection techniques and r adiocarbon pre treatment methods which
58 are now known to help ensure more accurate and precise results (Brandt et al., 2010) Today, it is important to know what material is being sampled for radiocarbon dating. Bone, for example, poses dating problems due to potential contamination and weathering (Hedges and Vanklinken, 1992) The collection of radiocarbon samples can also affect results. Samples taken from sieves at Moche Borago for example, provided widely inaccurate and imprecise results due to contamination (Brandt et al., 2010) Methodological problems also occur in dating many Late Pleistocene sites in eastern Africa. Many L ate Pleistocene sites are dated from a single sample or at the most two or three samples, (cf. Brandt, 1986) 4 at Lake Besaka are the only sites dating to the Late Pleistocene in which each has more than three absolute dates. However, only the topmost deposits of the 11 m sequence which leaves the entire lower two thirds of the sequence to be inferred (Kurashina, 1978; Clark and Williams, 1978; Brandt, 1986) At Porc Epic, questions arise about the accuracy of the ag es, as well as the context of the archaeological deposits (Brandt, 1986) Problems Arising from Late Pleistocene Demography and Settlement Patterns While the lack of Late Pleistocene archaeological data in eastern Africa can be attributed to an absence of research or methods, the prehistoric distribution of people across this landscape may also be a major factor. Human populations across Africa during the Last Glacial were already very low, possibly due to the aftereffects of catastrophic events such as the Toba super volcano eruption (Ambrose, 1998b) These populations would have been spread thinly across the landscape, and current methods may be unable to identify the archaeological signatures of these occupations. Furthermore, human p opulation s may have moved into or out of specific areas during glacial p eriods such as OIS 4 and OIS 2 (Barham and Mitchell, 2008) Among the best known
59 examples of population movements is the widespread abandonment of previously occupied areas in the Sahara during OIS 4, ~74 ka, (Wendorf and Schild, 1992) These populations may have moved east into the Nile valley, but even these occupations appear to have been ephemeral (Van Peer, 1998; Van Peer and Vermeersch, 2000) In Somalia, early Holocene hunter gatherers may have been targeting isolated inselberg formations (large rocky outcrops) instead of living further abroad these features (Brandt, 1986) In Kenya, Late Pleistocene hunter gatherers may have followed specific environmental ecotones (Ambrose, 2001) Humans likely a lso congregated in isolated environmental refugia across the Horn and East Africa during the especially arid periods of OIS 4 and OIS 2 (Ambrose, 1998b; Brandt et al., 2006; Basell, 2008) The lack of Late Ple history, but the lack of sites itself may be a real pattern also OIS 3 Archaeological Record: Available Data Basell (2008) noted approximately 60 sites in eastern Africa that contain Middle Stone Age archaeological materials. Of these sites, only 14 have deposits dated directly to OIS 3 and artifacts that span the major technol ogical change at this time to microlithic tools. These 14 sites across this region equate to only 1 site for every ~420,000 square km (Figure 4 1) To compound the problem, many of these sites have dating or context issues For convenience, I have prov ided a summary in Appendix B of sites from across the Horn, East Africa, and North Africa that date to OIS 3. These sites have archaeological deposits which span the transition to microlithic tools at this time. This summary includes published ages and b rief descriptions of archaeological materials I also provide additional details about the archaeological sites from the Horn and East Africa that date to OIS 3 and that have early microlithic technology.
60 Mode 4 and Mode 5 Stone Tools Microliths are blu nted (backed) on one lateral edge for purposes likely related to hafting or handling the tool. These artifacts often take the form of crescents or geometric trapezoids. The development of microliths during OIS 3 signals a major change in technology and s ubsistence strategies among hunter gatherers at this time. Microlith technology enabled hunter gatherers to build multi component tools to more effectively use raw materials. It is also likely that the development of these artifacts signals the developme nt of other important technologies such as the bow and arrow (Ambrose, 2002) Microliths may even have been used in early social exchange systems (Ambrose, 2002) The earliest microliths date to the South African Howiesons Poort Industry, but current evidence suggests that the Howiesons Poort ended after 6 0 ka (Jacobs et al., 2008) The appearance of microliths in East Africa, 50 to 40 ka, marks the re appearance of the technology, which continued i nto the Holocene (Ambrose, 2002) Microliths are one technological hallmark for the Later Stone Age of eastern Africa (Clark, 1954; Brandt, 1986) Another technological hallmark for Pleistocene and Holocene Later Stone Age (LSA) technology is the use of blade and bladelet technology, which itself is seen as an increasingly more economical use of raw materials than prior stone tool technologies (McBrearty and Brooks, 2000) Following (1971) classification of S tone A ge lithic technolog y, blade and microlithic/bladelet tools are classified as Mode 4 and Mode 5 respectively useful because it considers only technological changes. Middle Stone Age (MSA) and LSA, by comparison, imply both cult ural and technological changes. The shift from MSA to LSA, for example, signifies technological and subsistence changes and also transformations in hunter gatherer cultures, including language, symbolism, and identities (McBrearty and Brooks, 2000; Wadley, 2001) The MSA and LSA also have different characteristics at various times or places.
61 For instance, in southern Africa, some Pleistocene LSA assemblages do not have microliths whereas they are often presen t in Holocene LSA assemblages (Mitchell, 2002) technologies, w hich are common to both Late Pleistocene and Holocene LSA assemblages in eastern Africa and the Horn of Africa. The earliest published age for Mode 4/5 technology is for Mumba rock shelter in Tanzania where a number of chronometric ages have been provide d ( 25 conventional radiocarbon, 4 AMS, 1 amino acid racemization 6 uranium series, and 5 p otassium thorium ) (Mehlman, 1989; Marks and Conard, 2006; Prendergast et al., 2007) However, despite these efforts, t he age of the site is still contentious (Marks and Conard, 2006; Prendergast et al., 2007) Much of the discussion revolves around the dating of Layer V which includes early Mode 5 technology This layer is generally taken to date to ~65 ka but this age is based on a s ingle U Th date (with an error of at least 6,049 years) However, different results can be calculated if all of the various dates from Layer V are considered (n = 11; cf. Prendergast et al. 2007) and then run through a central age model (Galbraith and Laslett, 1993; Van der Touw et al., 1997; Galbraith et al., 1999; Galbraith et al., 2005) Based on the central age model, which considers both the average and standard deviations of a sample, the weighted mean is only 32,961 3,347, with an error approximately half that of the supposedly more accurate U Th age. Unless substantive evidence suggests a certain technique or a range of ages is more accurate than any other technique, the age of Layer V at Mumba must remain inconclusive. The most securely dated early Mode 4/5 deposits are currently found at Enkapune Ya Muto ( Tanzania ), dating >50 to 46 k a. The Enkapune Ya Muto sequence is dated using 35 conventional radiocarbon and AMS ages plus 3 obsidian hydratio n ages (Ambrose, 1998a) The
62 earliest Mode 4/5 deposits (>50 46 ka), however, are not directly dated. Rather, the published ages are estimates based on sediment deposition rates calculated from the obsidian hydratio n chronology (Ambrose, 1998a) Considering the well cited problems with obsidian hydration (Mazer et al., 1991; Anovitz et al., 1999; Hull, 2001; Rogers, 2008) the 14 C a ges on ostrich eggshell s in overlying deposits may be a more confident minimum age estimate from 39 to 37 ka Only four sites in Ethiopia have stratified early Mode 4/5 deposits. Near Dire Dawa, Porc Epic cave has a 2.5 m sequence, which is capped by c emented breccia (Clark et al., 1984; Assefa, 2006) The lithic technology from Porc Epic has been studied closely (Pleurdeau, 2001; Pleurdeau, 2005) but serious concerns linger about the context and dating results at the site (Brandt, 1986) Already mentioned, has an 11 m sequence but only the topmost portion of the stratigraphically highest loam deposit has been dated (Kurashina, 1978; Clark and Williams, 1978; Brandt, 1986) T he Bulbula R iver site in central Ethiopia (near Ziway) has a single conventional radiocarbon date of 27,050 1,540 14 C BP for the Mode 4/5 bearing paleosol (Gasse and Street, 1978; B randt, 1986) Liben Bore in southern Ethiopia has been studied closely (Brandt, 2000; Negash, 2004; Fisher, 2005) however, it is currently undated. Also worth mentioning, locality FeJx at Lake Besaka (centr al Ethiopia) has slightly younger Mode 4/5 artifacts in Terminal Pleistocene deposits dating from 19 to 22 14 C BP (Brandt, 1986) In Somalia the dating and context is no better. The stratified Mode 3 and Mode 4/5 deposits at Midhishi 2 (NE Somalia) are dated by two radiocarbon samples with the oldest, >40 ka, providing only the minimum age for the depos its (Brandt and Brook, 1984) At Gud Gud the deposits are dated to >40 ka by a single radiocarbon sample but the assemblage is also small, being composed of less tha n 50 mostly un diagnostic artifacts (Brandt and Brook, 1 984; Brandt,
63 1986) At Gogoshiis Qabe in southern Somalia, the Mode 3 deposits are stratified under terminal Pleistocene and Holocene assemblages, but the Mode 3 deposits themselves are undated (Brandt, 1988) Conclusions Archaeological data are sparse for the Late Pleistocen e in eastern Africa due partly to the historical record of research in this area. The methods and techniques available to prior researchers may also have contributed to the lack of data today. At the very least, the absence of dating methods and methodol ogical advances commonplace today makes much of the prior findings difficult, but not impossible, to compare with new data. But the lack of sites might reveal something about the nature of human populations in the region or where they were located during the Late Pleistocene. Resolving these issues requires new research at previously studied sites, as well as the analysis of new sites. Toward this goal, the findings from Moche Borago provide an important new record of human occupation in the Horn during OIS 3 of the Late Pleistocene. In Chapter 5, I will introduce the Sodo Wolayta region where Moche Borago is located and describe the prior research which has been conducted at the site.
64 Figure 4 1 : Map of the Horn of Africa (red) and East Africa (ye llow) showing the archaeological sites discussed in the text.
65 CHAPTER 5 MOCHE BORAGO ROCKSHE LTER: LOCATION, ORAL HISTO RY, AND PRIOR ARCHAEOLOGICAL RESEA RCH Mochena Borago Sodwa I n the local Wolayta language of the Wolayta people the traditional designation for the rockshelter takes its name from Moche Borago According to oral history, Moche Borago was among the first owners of the land around the rockshelter and unknown to most outsiders his name also commands a unique place in southern Ethiopian history. Reportedly, Moche, and perhaps his father Borago, was an adviser to the last k ing of Wolayta before he was defeated by Emperor Menelik in 1894 at the battle of Chelsha in Borada (Tema Ase le, Balcha Basa, and Alaro Asele, per comm. ). The unique oral history surrounding this site and nearby rockshelters compl e ment the prehistoric record being revealed in the current archaeological excavations. This chapter introduces the Wolayta region, it s people, and the local history of Moche Borago rockshelter. The second half of this chapter provides an overview of prior scientific research at Moche Borago, including the Southwest Ethiopia Archaeological Project (SWEAP) excavations from 2006 to 2008. Physiography of the Wolayta Region Wolayta is located within the (SNNPR) of Ethiopia. This region is currently one of the most densely populated areas of Ethiopia, which includes a diverse collection of spoken languages and cultures. Sodo is the capital city of Wolayta, found at the southern foothills of Mt. Damota approximately 3 km from Moche Borago. Wolayta itself is located between the lowland Rift Valley and lake country and southern Ethiopian h ighlands. Variations in elevation,
66 temperature, and rainfall here create a series of altitudinal ecological zones that provide an abundant source of plant and animal resources (Lesur Gebremariam, 2008) Ethiopian Rift Valley System Ethiopia is dominated by the tectonic activities of the East African Rift System. The country is positioned just off the epi center of three divergent tectonic plates which form the East African R ift system to the south, the Red Sea to the north, and the Gulf of Aden to the northeast. The intersection of these plate boundaries is the Afar triple junction and the most prominent feature other than the R ift Valley is the depressed and hyper arid Danakil depression in northern Ethiopia and Eritrea. Within the Rift Valley, the divergent action of the African and Arabian plates forms a nort h south series of normal fault grabens tha t have been in filled to form the Ethiopian lake country (Di Paola, 1972) A long the peripheries of the R ift systems tectonic uplift produced the Ethiopian highlands around 75 million years ago, elevating these areas on average >1 500 m above sea level (asl) The Ri ft Valley subdivides the highlands into western and southeastern components in Ethiopia Climate Tropical Convergence Zone (ITCZ) and monsoonal moisture (Kebede, 2004) From October to March, the ITCZ is positioned south of Ethiopia, allowing cold, dry air to flow in from Arabia. During April and May, the ITCZ moves northward drawing in moisture from the Indian Ocean that accounts for ~25% of the total annual rainfall in Ethiopia. From July country drawing mois ture rich monsoonal air masses from the Atlantic Ocean across Ethiopia and the Horn (Kebede, 2004)
67 Highland a reas, such as southern and southwestern Ethiopia, receive greater amounts of rainfall than lowlands due to the orographic influence of the highlands on the monsoon air masses (Livingstone, 1975; Gasse, 2000; Moh ammed et al., 2004) Moche Borago is located near the eastern limit of the southern highlands in a transitional zone to the lowland lake country. The city of Awassa, which is located in the Rift Valley lake country to the east, receives ~120 mm per mont h from July to August whereas the city of Jimma to the west in the highlands receives >200 mm (National Meteorlogical Agency, 2 010) 3 In Obe Jage, north of Mt. Damota, the average annual rainfall is ~1,250 mm (Le Gal and Molinier, 2006) and rainfall around Moche Borago should have a similar average annual rainfall. Physiography and Vegetation Moche Borago is situated ~2,200 m asl on the western flank of the volcano massif Mt. Damota. This mountain is a major fres hwater resource in the area because precipitation at higher elevations of the mountain drains via numerous stream channels into the valleys below (Borena, 2008) The location of Moche Borago makes it accessible to lowland zones and resources to the south an d east, including the Omo gorge and Lake Abaya, as well as highland resources to the west. On clear days, the western ridge of the highlands is visible from the Moche Borago site to the west, and Lake Abaya can sometimes be seen to the southeast. The n atural vegetation of the area has been replaced largely by farming and eucalyptus ( Eucalyptus globulus ). Today, the slopes of Mt. Damota, and much of the surrounding countryside, are heavily cultivated by seasonal crops of domesticated maize 3 The meteorological station at Sodo was closed in 1986. Awassa and Jimma are the nearest major stations.
68 ( Z ea mays) b arley ( Hordeum vulgare) sorghum ( S orghum bicolor) finger millet ( Eleusine coracana) teff ( E ragostis tef) and enset e ( E nsete ventricosum ), as well as numerous other crops (Le Gal and Molinier, 2006) The predominant natural vegetation of the area is Ethiopian Afromontane forest. However, the distribution of these forests is now heavily restricted and fragmented due to farming. Hildebrand (2003) provided a comprehensive summary of vegetation species common to Afromontane forests. In general, Afromontane forests occur from 1,500 to 2,600 m asl. The extant Afromo ntane forests around Moche Borago are dominated by the conifer Podocarpus falcatus and Juniperus procera The lowlands below Mt. Damota consist of patchy Afromontane forest and defined by Hildebrand (2003) which occurs from 500 to 1,500 m asl. This forest zone includes species such as euphorbia ( Euphorbia candelabrum ) and acacia ( Acacia Abyssinia ) known to occur around Sodo. Based on firsthand observati ons, our SWEAP team speculated that the grasslands below Moche Borago may be relict marshlands from wetter periods of the past. This region might have sustained abundant natural resources in the past. The most obvious is freshwater, which would have p recipitated at high elevations of Mt. Damota and collected into stream channels that eventually drained into the lowlands. During periods of aridity today, precipitation still collects on top of the mountain. During previous glacial periods, it is likely that the same phenomenon might have occurred. Afromontane forests include a number of edible wild plants. To date, no study has been published of the extant or archaeological wild plant types from the Wolayta
69 region, though these studies are part of ongoing SWEAP research. Le Gal and Molinier (2006) provided descriptio ns of agro economy from 1940 to present day, but they did not focus on wild plant resources. The best survey of edible highland plant resources in the SW Ethiopian highlands is Hildebrand (2003) However, her study targeted the Be nch Maji area in extreme SW Ethiopia, which is located at a higher elevation than Wolayta and also receives greater annual rainfall. Hildebrand (2003) also noted that Afromontane forest in the far southwest Ethiopian highlands diff er from Afromontane forests elsewhere in the highlands. However, the species diversity of Afromontane forests in Bench Maji should still approximate the species found in Afromontane forests in Wolayta. The wild edible plants found by Hildebrand (2003) to be common in Afromontane forests include seasonally available greens, fruits, oil seeds, and tubers. Certain resources are available only at certain times of the year. Greens, for example, are most widespread today during the sp ring rains (March to May) whereas oil seed is most common in the dry seasons (Hildebrand, 2003) Tubers are known from ethnographic sources to be a primary subsistence resource of African hunter gatherers (Vincent, 1984) While yams are present in Afromontane forests, it is the false banana, known as enset ( Enset ventrocosum ), that is a staple crop today and often regarded as a famine food during drought (Hildebrand, 2003) Besides being a food resource, enset also provides wood, fibers for clothing and binding, and freshwater (Hildebrand, 2003) Language Omotic and Cushitic are the only indigenous language families found in the Wolayta area to day (Lamberti and Sottile, 1997) Amharic a Semitic language, is the
70 primary language today for commerce and trade across Ethiopia and the Wolayta re gion. T he use of Amharic ( SNNPR ) of Ethiopia today can be traced back to the Amhara expansion and subjugation of southern k ingdoms by Emperor Menelik in the late 19 th century. The Wolayta langua ge is part of the Ometo sub branch of the Omotic language family (Lamberti and Sottile, 1997) To the west, northwest, and east of Wolayta are highland Cushitic speake r s (Kambaata, Haddiya, and Sidamo, respectively) (Lamberti and Sottile, 1997) To the south and west the Gamo and Gofa peoples share a similar O motic language with the Wolayta. Wolayta in this dissertation. Wolayta, as the name for the language and culture in the Wolayta region, is plural for Wolamo / Wellamo S ome researcher s (cf. Cerulli, 1956) have argue d that Wola mo is more accurate to use when speaking about language and culture Lamberti and Sottile (1997) however, contended that local people today prefer the term Wolayta The use of this term in this region is interesting because this Omotic name uses the Cushitic singular suffix. Political History U ntil the late 19 th century, the Wolayta region was a feudal kingdom, part of the wi der network of Gibe k ingdoms that included the Sidamo, Kafa, and Jimma states. R ecords indicate that Wolayta was a subsidiary of the Kindgom of Kafa. Trade was an important staple for the Wolayta Kingdom owing to its location at the juncture of the more northerly Cushitic speaking peoples and southerly Omotic peoples (Cerulli, 1956; Lamberti and Sottile, 1997)
71 Beckingham and Huntingford (1954) stated that the Wolayta Kingdom was established in the mid 13 th century with the Damot D ynasty. The second dynasty ( Wolaitamola ) dealt with repeated Islamic invasions (Shinn and Ofcansky, 2004) The foundation of the third dynasty in the mid 16 th century began in Tigray (northern Ethiopia and Eritrea) Kawo (King) Damota (1835 1845) of the Tigray Dynasty shares the same name as the volcan ic mountain on which Moche Borago is located. Local History of Moche Borago Rockshelter According to local history, the rockshelter and surrounding land was purchased by Moche Borago from Ato Wara and 4 was originally from Sodo Ato Moche, and perhaps his father Ato Borago, was an adviser to the Wolayta and Borago were members of the Maka clan of the Dogala branch of the Wolayta. The rockshelter may have b een unnamed before either Moche or Borago received it, meaning that name likely predate s B efore Moche Borago purchased the land around the rockshelter, it might have been used as a sacred place for ritual sacrifices of goat and sheep. In the Wolayta re gion, ritual sacrifice is recorded in Cerulli (1956) who note d that the practice was most common to Tleh, the spirit of the Omo River 5 but it is also present in other ceremonies. It is unclear if Moche Borago ever inhabited the rockshelter R emnant wall foundations have survived on the south side of the rockshelter but these are widely believed to date to the 19 th century when Menelik occupied the area, or to the 1930s and 4 5 This is only an example becaus e Cerulli (1956) mentioned that ritual sacrifice to the spirit Tleh was made in the Omo River.
72 1940s during the Italian occupation (1936 1941) of Ethiopia. Both ideas may be correct since recent archaeological evidence in Moche Borago supports the use of the site during the early 19 th century (e.g. coins) and use of the site within the past few hundred years. Prior Archaeological Research at Moche Borago In 1995 Roger Joussaume of the C entre National de la Recherche Scientifique (National Center for Scientific Research) in Nanterre France, developed the Groupe d'tude de la Protohistoire dans la Corne de l'Afrique (GEPCA ) (the Protohistoric S tudy G roup in the Horn of Africa). The focus of this project was to study the origins and processes underlying plant and animal domestication in the Horn of Africa during the Neolithic and how these developments related to underlying environmental changes (Gutherz et al., 2000a) The project was subdivided into multiple components including teams studying rock art and megalithic monuments in the reg ion and archaeological excavations in Djibouti and Ethiopia of which Moche Borago was one element GEPCA Excavations : 1998 and January February 2000 The first excavations at Moche Borago were undertaken under the leadership of Xavier Gutherz of the Univer sity of Montpelier (France) in 1998 and again in January and February of 2000 (Gutherz et al., 2000a) Between these two field seasons, team took geological samples of the shelter floor and walls and created three 4 m 2 test excavation areas (Figure 5 1) From the geological samples the team was able to show the presence of volcanic deposits in a nd around the site and that the formation of the rockshelter was likely an abscess of softer materials within ancient and highly heterogeneous volcanic lahar deposits. E xcavations in three areas revealed stratified archaeological deposits containing abu ndant pottery, burnt and unburnt bone, and lithics within stratified
73 sandy and silty deposits, including numerous well defined hearth features (Gutherz et al., 2000a) T he lithic artifacts included large amounts of blade and flake debitage (Poisblaud, 2000) The findings in Test Pit 1 were particularly rich and this area was subseq uently expanded to 6 m 2 during the 2000 field season. Excavation in Test Pit 2 revealed fluvial activity and reworked deposits with interstratified volcanic layers. s team excavated a narrow 0.5 m x 1 m sondage from a pproximately 1.5 to 1.8 m below the current cave floor. The sondage excavations stopped after a hard, dense rocky deposit was discovered, which Gutherz and his team believed to be bedrock. Mode 3 lithics were also recovered in G10, suggesting that strati fied Middle Stone Age ( MSA ) deposit s might underlie the Holocene deposits above. Radiocarbon samples taken below the well defined tephra 11 deposit 6 were dated to 28,700 1,100 BP (31,184 906 cal. BP using Cal Pal Hulu). The Holocene sequence above te phra 11 dated from 4,370 70 BP (3,062 124 cal. BP) to 1,480 60 BP (852 67 cal. BP). GEPCA Excavation : November 2000 A third excavation in November 2000 focused on stratigraphic and sedimentological analysis of the Holocene deposits at the site including correlations between various profiles in the excavation areas (Sordoillet and Pouzolles, 2000) Post hole structures were also identified in the stratigraphic profiles of Holocene deposits in Test Pit 3 A ring of post holes was uncovere d during lateral excavations at Test Pit 1 which may date to the first millennium (Jallot and Pouzolles, 2000) A large amount of 6 see Chapter 6).
74 charcoal was recovered near this area, which led Jallot and Pouzolles (2000) to suggest on site metallurg y even though no other evidence, such as slag, supported this theory Research during th e 2000 season also included a micromo rpholog y analysis that found phytoliths, carbonized and uncarbonized plant remains and carbonized seed s in the Holocene deposits (Gutherz et al., 2001) Faunal materials from these deposits which were also preliminaril y studied include d a large percentage of bovine remains but also artiodactyls and carnivore remains including Panthera pardus, Colobus guereza, and Dendrohyrax arboreyeus Crocodiles, fish, and shellfish (species not provided) were also recovered (Lesur, 2000) The November 2000 excavation also provided more information on the ceramic sequence in this area. Analysi s suggested an aceramic Holocene occupation at the site, which is dated via radiocarbon on charcoal to 4,370 70 BC (3,062 124 cal. BP) (Gutherz et al., 2001) Overlying the aceramic Holocene layers were deposits cont aining a variable amount of ceramics, often having curvilinear decorations. In the uppermost deposits traditional ceramic jebena (coffee pot) and (spice jars) fragments were recovered (Jallot and Pouzolles, 2000) GEPCA Excavation : December 2001 In December 2001, GEPCA carried out a final excavation at t he site. During this season Test Pit 1 and Test Pit 3 were connected in order to clarify ambiguities between the stratigraphic profiles of these areas (Gutherz et al., 2001) This correlation also helped to resolve the stratigraphic lateral associations of a rich bone and ash layer discovered in Test Pit 3. This layer contained two thirds of the total faunal material found at Moche Borago and that faunal material has been relatively dated via stratigraphic correlation to approximately the early 5 th millennium BC. Lesur (in Gutherz
75 et al., 2001) note d that the faunal samples are dominated by Bovinae, especially buffalo ( Syncerus caffer ) but a distinct Bovidae comp onent includes Gazella, Tragelaphis, and possibly Kobus. Lesur found domestic cow ( Bos t aurus ) in the upper and most recent layers at the site. (Gutherz et al., 2001) impression of the Holocene fauna is that the fauna reflects a predominant hunting subsistence strategy focused on Bovidae, especially buffalo and that domesticated animals, such as cow, arrived in this area only within the last millennium Other faunal identifications include the remains of prima tes ( Papio cynocephalus, Colobus guereza ), hyena ( Crocuta crocuta ), suides ( Potomocheroerus larvatus and Phacochoerus africanus ), and hyrax ( Heterohyrax brucei ) (Lesur 2001). Several human teeth were also found in the deposits but no known study has been published on this discovery Most faunal remains have been fire altered as the interpretation of the Holocene faunal assemblage suggest s. The diversity of the fauna remains also indicates the exploitation of two distinct biomes during the Holocene : the humid plains along the Weja River below Mt. Damota and a forest biome (Gutherz et al., 2001) A similar subsistence strategy between highland woodlands and lowland grasslands is speculated for Pleistocene hunter gathere rs occupying the site. SWEAP Excavations : 2006 2008 The Southwest Ethiopia Archaeological Project ( SWEAP ) began excavati ng M oche Borago in 2006. This multi national project led by Co PIs Dr. Steven Brandt (SB) (Univ ersity of Florida) and Dr. Elisabeth Hildebrand (EH) (Stony B rook Univ ersity ), was funded by the U.S. National Science Foundation. Research conducted by SWEAP focused on assessing the context of the stratigraphic and archaeological sequences (pre Holocene) at Moche Borago as reported by Gut herz (Gutherz et al.,
76 2000a; Gutherz et al., 2000b; Gutherz et al., 2001) and continuing excavations into Pleistocene deposits across the site in a systematic and knowledgeable manner (Figure 5 2) The origi nal focus for SWEAP research at Moche Borago was human adaptation during arid periods of the Late Pleistocene, including OIS 4 and OIS 2/Last Glacial Maximum (LGM) Brandt et al. (2006) and Hildebrand et al. (2008) hypothesized that humans moved into wetter regions of the Horn, such as southwestern Ethiopia, during the early Last Glacial (~74 60 ka ) and Last Glacial Maximum (~18 ka) because these areas may have sustained greater resource diversity at these times. The confluence of people moving into SW Ethiopia may have affected the genetic, linguistic, and cultural demography in this region, thus spurring the development of Later Stone Age social and technological behavioral in novations. Since 2006, radiocarbon dating both accelerator mass spectrometry (AMS) and conventional exist anymore (see Chapter 6 for more detail). The oldest, currently dated depo sits at the site also fall within early OIS 3, meaning that OIS 4 layers are also unlikely. The sequence is unique, however, because these deposits preserve among the most complete records of OIS 3 occupation in the Horn. To simplify the discussions of t he deposits across the site SWEAP placed all excavation units on site (both GEPCA and SWEAP ) into three main excavation group s: 1) Block Excavation Area (BXA), 2) TU2 (Test Unit 2 area of GEPCA) ~9 m southeast
77 of BXA, and 3) the newer area known as N42, ~ 9 m southeast of TU2 and ~18 m from BXA. N42 is currently the most southerly excavation at the site (Figure 3 2) 7 The SWEAP team rediscovered, exposed, photographed, and mapped using a Total Station all the GEPCA excavation units and stratigraphic profiles. During the initial mapping of the GEPCA excavation units, the SWEAP team noticed that the grid system used by the GEPCA team was offset 30 from t rue n orth most likely done intentionally to accommodate the natural orientation of the site 8 G eodetic control for SWEAP research is based on the UTM system (zone 37N). Unlike the arbitrary grid used by GEPCA, the global UTM system allows local site data to be more easily articulated into regional and global contexts without requiring coordinate tr ansformations in most circumstances. The SWEAP team preserved the orientation of the pre existing GEPCA excavation units but new excavation areas are oriented according to the t rue n orth UTM system. The variation between the GEPCA and SWEAP grids means that cardinal directions which conform to the GEPCA units are not equivalent to those in the newer SWEAP units. Therefore, in the following discussions, any cardinal orientation that is relative to the 30 Nort h angled offset of the GEPCA grid is indica ted with the prefix GG (GEPCA Grid ), for example, G10 GG north profile, wh ereas orientations relative to newer SWEAP units do not have a designate and are aligned to true north (e.g. N42E38 north profile). The alphanumeric designate of the GEPCA excava tion units was also retained (e.g. G10) N ewer UTM based excavation units use a Northing Easting schema (e.g. 7 In this chapter, and subsequent chapters, a unique three letter name is used to define stratigraphic layers within each excavation area, like BWT. For further information on the stratigraphy and descriptions of particular layers, please see Chapter 6. 8 Magnetic declination from 1995 to 2000 was less than 2 at Moche Borago.
78 N42E38) relying on the coordinates of the southeast corner of an excavation unit that are truncated to two significant digits to the left of the decimal. All spatial data, regardless of excavation area orientation, are recorded using UTM coordinates. SWEAP Excavation s: 2006 In 2006, SWEAP focused primarily on new excavations in unit N42E38 while continuing GEPCA excavations into the Pleistocene deposits in the BXA unit G10 and TU2. G10 had been dug originally as a 1 m x 0.5 m sondage to a depth of ~2 m below surface by GEPCA in 1998 (Gutherz et al., 2000a; Gutherz et al., 2001) In 2006, SWEAP excavated the remaining northern half of G10 relatively quickly in order to assess firsthand the stratigraphic and general archaeological sequence in the Pleis tocene deposits at the site. Archaeological materials recovered from this 1 m x 0.5 m area were minimally piece plotted, except for charcoal samples. The SWEAP excavations in nearby units G9, H9, and I10 proceeded at a slower pace and every artifact fou nd in situ was plotted using the t otal s tation, which is normal SWEAP protocol. At the close of the 2006 SWEAP field season excavations in G10 had reached a depth of 1.60 m below the current rockshelter surface 9 but excavations still had not hit the I began a nalysis of lithic material from G10 in July 2006 with assistance from William Wright and Kochito Kero Preliminary results from this analysis showed that the deepest excavated areas from G10 con tained Mode 3 and Mode 4 /5 lithic artifacts. Strata from the upper G10 Pleistocene deposits, however, appeared to have Mode 4 /5 lithics only. 9 This depth is equivalent t o 1.05 m below tephra BWT, which defines the Holocene/Pleistocene boundary in the BXA deposits.
79 In TU2, a dense layer of heavily rolled and abraded pebbles and stone artifacts with in a reddish silty clay laye r (RGX) were found stratigraphically below a compact volcanic tephra that was assumed to correlate to tephra BWT in the BXA area The TU2 sequence appeared deflated relative to the BXA excavations, and sediments here seemed to be more frequently derived f rom volcanic activities. The stratigraphic sequence in the N42 area completely contradicted the strata from the BXA. The excavations here revealed ~1 m of stratified, archeologically sterile volcanic tephra/ash deposits (some with blocky clasts exceedin g 10 cm diameter) overlying another 1 m of dense, stratified volcanic lahar deposits. Only a relatively thin series of three silty sand layers ( layers: UPFC, MPFC, and LPFC) unconformably placed atop the uppermost lahar layer (ULF) contained deposits si milar to those in the BXA unit where archaeological materials might be expected to be found. In late 2006, five c harcoal radiocarbon samples were sent to Dr. Hong Wang at the Illinois Geological Survey. Due to the small size s of charcoal multiple piec es of charcoal from the same stratigraphic layer were aggregated into a single sample. The calibrated ages of these samples suggested that the deposits underlying tephra BWT at Moche Borago dated to O IS 3 ( samples SWAP06 1 to SWAP06 5). These same result s also indicated a possible unconformity dating to the Terminal Pleistocene between the uppermost Pleistocene deposits ( layers: RCA, RGCA, and RGCB) and the BWT tephra which marks the Holocene/Pleistocene transition 10 10 S amples of the BWT tephra were also collected by Dr. Leah Morgan for Ar 40 /Ar 39 dating analysis in 2006. The results of this analysis provided an age of 3.16 MA. The Mt. Damota volcano is known to have been active during the Pliocene (Woldegabriel et al. 1990), and this age may be due to feldspars that were derived by erosion from the pre existing older volcanic sources and do not reflect the true depositional age of BWT ( L. Morgan 2008 per comm. ).
80 SWEAP Excavation : 2007 The 2007 fi eld season focused on resolving many of the questions raised by the prior 2006 field season including the lateral correlations of stratigraphic deposits across the site, the depth of the Pleistocene deposits in the BXA, and the ages of the Pleistocene dep osits across each excavation area. Prior excavations were continued in the BXA (G10, G09, and H09), TU2 (TU2S), and N42 (N42E38). T as described by Gutherz et al. (2000a, 2001) was also reached in 2007 in unit G10 ~2 m below the roc kshelter surface. This stratigraphic layer, which was named PKT, appeared to be more like a lahar deposit than bedrock due to the clast size s texture, color, and homogeneity At the completion of the 2007 field season nearly 2,500 stone artifacts fa unal material, ochre, and charcoal had been plotted from units G10, G9, and H9 11 Every stratigraphic profile in the BXA had also been mapped, described, and sampled. Excavations in G10 reached the sterile PKT lahar. In G9, the excavations ended within t he RCA deposit whereas in H9 the excavations reached stratigraphic layer VDBS. In TU2, both GG east and west profiles were mapped, described, and sampled. Excavations in N42E38 were stopped 2.2 m below surface, still within archaeologically sterile volca nic deposits. The anthropogenic pits/possible paleo fluvial features first noticed in G10 GG north in 2006 were seen also in TU2 and N42. In TU2, the LFX1 lahar had been truncated and the underlying LFX2 lahar deposit had been undercut by fluvial acti on before the YBSX tephra had in filled this area. Similarly, in N42E38, the anomalous 11 The current total following the March 2008 field season is 6,520 plotted pieces of lithics, charcoal, bone, and ochre.
81 UPFC, MPFC, and LPFC silty clay deposits first noticed in 2006 were found to contain size sorted cobbles, also likely fluvially deposited. T he N42E38 silty clay deposi ts unconformably overlie the ULF lahar deposit which is similarly believed to have been eroded fluvially. In December 2007 an additional 13 radiocarbon samples on charcoal (SAWP07 01 SWAP07 13) were processed and analyzed by Dr. Hong Wang at the Illi nois Geological Survey. Unlike the prior five radiocarbon samples submitted in 2006, the new samples were all piece plotted and collected in situ and they were not aggregated into larger sample sizes. Five samples targeted the possible unconformity unde rlying tephra BWT and the age range of the deposits immediately underlying BWT (RCA, RGCA, and RGCB). Five charcoal samples were selected from across the other BXA Pleistocene deposits and two charcoal samples were taken from TU2. The results of this as say confirmed the Terminal Pleistocene unconformity between RCA and BWT at the top of the Pleistocene sequence. Furthermore, the CTTC deposits, among the deepest at the site, were dated to ~53 ka and the remaining samples from BXA confirmed the nearly com plete O IS 3 chronological sequence. The samples from TU2 helped to establish a stratigraphic/chronological correlation between the BWT and YBS tephras in these two areas. SWEAP Excavation : 2008 Following the 2007 field season the chronology and composi tion of the Pleistocene deposits at Moche Borago were sufficiently well known to be classified into five major stratigraphic aggregates following the sequence seen in the FG G10 north profile. These stratigraphic groups are described in Chapter 6
82 In early March 2008 an additional eight radiocarbon samples (SWAP08 1 SWAP08 8) were processed by Dr. Hong Wang at the Illinois Geological Survey. Many of t hese samples were prudently chosen to target the very top of the RCA deposits, which were carefully excavated from units H9 and G9 The rest of the new radiocarbon samples were selected to help define the age of the YBT tephra in the BXA area In late March 2008 I returned to Moche Borago with K ochito Kero and M enassis Girma to continue the excavati on of H9 in the BXA. The 2008 excavations were designed to increase the artifact sample size between the YBS and YBT tephras, dating to ~43 ka These excavations were part of the doctoral field research for this dissertation which was funded by the J. Wi lliam Fulbright F oundation as part of the 2007 2008 Ph.D. R esearch Scholarship P rogram. During the 2008 excavations a semi circular feature of burnt and discolored sediment was found in H9 containing a loose in filling of abundant bone and stone artifac ts. This feature was interpreted to be a series of hearth deposits, repeatedly excavated and refilled with anthropogenic materials during the occupation of the site ~43 ka. The depth of the highly consolidated PKT deposit was also explored more extensiv ely using a 0.5 x 0.5 m test excavation. This endeavor was halted after the test unit reached a depth of 1 m below the upper surface of PKT. Following the 2008 field season four more radiocarbon samples (SWAP08 9 SWAP08 12) were sent to Dr. Hong Wang at the Illinois Geological Survey. These samples increased the total radiocarbon array of the Moche Borago Pleistocene deposits to 29 making this excavation location among the most chronologically secure sites in the Horn of Africa
83 Esay Rockshelter Dur ing the March 2008 field season myself, Kochito Kero, and Menassis Girma also visited the nearby Esay rockshelter, about a 30 minute walk south of Moche Borago. Esay rockshelter is smaller than Moche Borago, being approximately 28.5 m long and 20 m deep. The site is located at the head of a deep, steep ravine such as Moche Borago. However, unlike Moche Borago a shallow stream percolates out of the rear cave wall at Esay, bisecting the site. This site is a useful modern analog for possible paleo fluvia l channels seen at Moche Borago The implications of the observations at Esay will be discussed in Chapter 6 Today, Muslim residents of the area visit Esay rockshelter because the stream at the site is believed to contain h oly water as evidenced by th ree modern fire hearths seen near the entrance to the site in 2008 From a geomorphological perspective, staining on the rockshelter walls and roof (especially around joints and cracks) suggests that either water flow may increase during the wet seasons o r was increased during more humid periods in the past. The current stream channel is headed by a 15 cm joint in the rear rockshelter wall 1.2 m above the current cave floor surface. At the base of this joint is a small pool T he stream channel itself is ~2 m wide and is bordered with small plants and shrubs (sp. unknown) A distinctive algal mat lines both bank s of the stream. The channel base is composed of heavily abraded pebbles that are size sorted and intermixed with obsidian and chert stone tools. The stone artifacts did not appear to be heavily abraded, however, and no diagnostic tool types were observed Conclusion s In total, Gutherz and his team excavated more than 20 sq m at Moche Borago, mainly in the northwest section of the site. The GEPC A excavations exposed a broad
84 lateral area which allowed the team to identify the presence of post holes, fire hearths, and other spatial patterning often not easily seen in vertical excavations. The excavations done by GEPCA also facilitated the predomi nantly Pleistocene focused research of SWEAP, which started in 2006, because much of the Holocene deposits in the various excavation areas had been previously removed. The SWEAP excavation s continued the GEPCA research in the BXA and TU2 areas while con currently starting new excavations in the N42 area. An array of speciali zed analyses and radiocarbon samples now provides a detailed perspective into the depositional history and stratigraphic context of the Pleistocene archaeological materials at the site Furthermore, it is now confirmed that Moche Borago contains O IS 3 archaeological deposits. The progressive radiocarbon strategy used to date these deposits is also unique among similarly aged sites in Africa and Moche Borago is now the best dated O IS 3 sequence in the Horn of Africa In Chapter 6, the stratigraphy and cultural sequence of the Pleistocene deposits will be provided in detail The stratigraphic descriptions will offer context to the stone artifact analyses that will be presented in Chap ters 7 and 8
85 Figure 5 1: GEPCA excavation areas at Moche Borago from 1998 to 2001
86 Figure 5 2: SWEAP excavations from 2006 to the p resent
87 CHAPTER 6 THE LITHO STRATIGRAPHIC SEQUEN CE AT MOCHE BORAGO ROCKSHELTER DURING EARLY OIS 3 This chap ter provides detailed descriptions of the Late Pleistocene litho stratigraphic sequence at Moche Borago which showed a nearly continuous deposition from early OIS 3 to the Holocene when there is a major unconformity in the main BXA excavation area. I pro vide a high level of detail for three reasons. First, very few African archaeological deposits date to OIS 3, so that Moche Borago provides unique information on OIS 3 depositional environments. Second, although well defined interstratifications are visi ble within the deposits that provided consistent and clear litho stratigraphic divisions, a great deal of stratigraphic variability exists between excavation areas, indicating the depositional history at Moche Borago is complex. Third, the detailed descrip tions contextualize the archaeological findings described in Chapters 7 and 8. Since sedimentary and geomorphic evidence suggest that certain time periods were wetter than other periods, at the end of this chapter, I will compare what is currently known ab out the depositional history of the deposits at Moche Borago to the known paleoenvironmental and paleoclimatic sequence across the region. Litho Stratigraphic Units Versus Culture Stratigraphic Units The descriptions of stratigraphic sequences at archaeo logical sites have two main common sources to draw upon: 1) geomorphology and lithology and 2) archaeology. Litho stratigraphic methodology focuses on the depositional history and diagenesis of the lithologic deposits. Culture stratigraphic methodology, however, prioritizes changes in the material culture accrued from human occupation at the site. Both methods provide exclusive.
88 However, they must be defined a priori and separated out a posteriori because litho stratigraphic sequences may not parallel changes in material culture as is the case with the oldest lithologic deposits at Moche Borago, detailed in Ch. 8. In this chapter, I therefore rely on a strictly litho s tratigraphic methodology. Descriptions of the archaeology within any group are presented here only for referential purposes. Methods mechanical processes (e.g. fluvial, aeol ian, root, human, and animal actions) and geochemical processes (e.g. soil formation, leaching, and percolation). The volcanic bedrock and ashes common to the Wolayta area (Di Paola, 1972) create alkaline sediment conditions, which affects bone preservation (Berna e t al., 2004) A number of methods are now employed in modern archaeology to discern the stratigraphic and archaeological history at a site. The SWEAP team relie d on a large and diverse international collaboration of researchers to bring reliable, cutting edge methods to bear on the study of Moche Borago. Stratigraphic Descriptions and Profiles A stratigraphic unit is a single, unique layer of sediment. Strati graphic units can be identified vertically from the layers above and below it, and they can be defined laterally from surrounding sediments. Stratigraphic units that share similar features (e.g. color, texture, consolidation, sediment type) are grouped in to larger litho stratigraphic units or stratigraphic group shares similar sediment characteristics, and group divisions are all well defined by volcanic ash layers.
89 We gave each new stratigraphic unit a unique name consisting of a short combination of letters (e.g. PKT), which sometimes acted as an acronym (e.g. BWT is If a stratigraphic unit was also numbered, it was part of a specific depositional sequence. Descriptions of individual stratigr aphic units were collected during the excavations as they were found. Detailed information was recorded about the topology, topography, and sedimentology of each stratum. Sedimentological information included the color (dry and wet), texture, consolidati on, and inclusions of the layer. Topological information related to the location of the stratigraphic unit relative to other layers above and below and to possible correlated units in other areas T opographical information about the top, middle, and base of every stratigraphic unit was collected using the total station, and was also useful for creating detailed 3D models of the strata. Photographs were also taken of each stratigraphic unit. Stratigraphic profiles were recorded using a total station and concurrently noted down onto paper graphs 12 Point spacing within the profiles was approximately 10 cm. The stratigraphic profile data were integrated into the site multidimensional GIS alongside the lateral topographical strata data and archaeological da ta Multidimensional GIS (mDGIS) Modeling A Trimble TS 305 total station was used on site by the SWEAP team since 2006. A TDS Recon drove the total station and customized drop down menus within the TDS Foresight software and provide d prompts to record t hematically based attribute information for every point taken. All archaeological and faunal materials in H09, G09, 12 E Hildebrand was instrumental in describing, maintaining, and mapping the Holocene and Pl eistocene stratigraphic profiles
90 and I10 13 were plotted in situ during excavations. These data were integrated daily into a m D GIS database built within ESRI ArcGIS 9.3 whi ch was cross checked for accuracy of the recording methods and spatial patterning within the data Stratigraphic survey points collected during the excavations were imported into the mDGIS to create 2.5D elevation models of the horizontal stratigraphy an d vertical wall profiles. The vertical profiles were created in 2.5D vector format using a custom Python script because ESRI ArcGIS does not accommodate vertically geo referenced raster imagery. Planimetric photographs taken of each stratigraphic unit an d level, however, could be geo referenced in ArcGIS and were often draped onto the elevation models to create photo realistic representations of the site. Bulk Sample Analysis Representative samples of each stratigraphic unit were collected into individu was recorded using the total station at the time the sample was collected. Sub samples were exported internationally for various analyses. Inductively Coupled Plasma Mass Spectrometry Inductively Cou pled Plasma Mass Spectrometry was attempted on sediment samples from the site to assess geochemical variability. Sample preparation procedures follow ed Kamenov et al. (2009) The samples were processed and analyzed by Dr. Jonathon Walz, with assistance by Dr. George Kamenov, at the University of Florida Department of Geologica l Science ICP MS laboratory. Seventeen samples were processed in 2008 and an additional 28 samples were analyzed in 2009. Both groups of 13 G10 and TU2 were considered test excavations and only charcoal was plotted.
91 samples were analyzed using an Element 2 (Thermo Finnigan) ICP MS. The 2008 samples were heated prior to analysis w hereas the 2009 samples were not cooked before analysis. No discernible difference occurred between cooking the samples prior to analysis. Quantification of the samples in parts per million was made using the USGS reference standard (J. Walz per comm.) X Ray Florescence The x ray florescence (XRF) analysis was also used to identify spatial geochemical variability within the sediments. Energy dispersive x r ay f lorescence analysis was done using a Bruker T racer III V portable XRF. The T racer III V use d an X ray tube excitation source with a Rhodium target. A silicon detector converts the incoming analog x ray pulses into proportionally sized digital signals. The T racer III V variably measure d incoming x ray energy between 0 and 45 keV. Sediment sampl es were placed into 31 mm diameter polyethylene rings with 0.16 mm ultralene film ends directly atop the T racer sensor. Spectra data were collected systematically fo ed all incoming x rays between 17 keV and 40 keV to reach the sample but the filter focuse d the sensitivity of the measurements only to elements between 5 keV and 17 keV. Spectra were analyzed visually using Bruker S1PXRF ver.3.8.27 and Artax ver.184.108.40.206 software. The software was used to export the raw spectra x ray abundance counts (1,024 rows long with 0.04 keV / row) into individual *.cs v files. These data were subsequently collated into a single database for statistical analysis. The data were normalized using the Rh K raw spectrum abundance values of sample CTTC, which had the lowest Rh K abundance.
92 Statistical analysis of the normalized spectrum abundance data was accomplished using SPSS 17.0. Quantile quantile ( QQ ) plots showed that the data were not normally di stributed. Therefore, bivariate relationships were derived using one tailed correlation correlation coefficient was squared to produce the coefficient of determination, R 2 which was then converted to a percentage (R 2 x 100) to describe the amount of shared variation in ranks between samples. All significant values were reported p <.01. Overall, the bivariate relationship between all unique pairs of samples was direct ional (one tailed) and very high (mean of shared ranked variation between samples = 95.4%; SD = 2.0, p < .01). The high relationship among samples was due to the overall chemical similarity within the deposits. However, some chemical variation was still expected and seen due to heterogeneity within the samples themselves or because of likely post depositional processes that affected the chemistry of the samples differently. Analysis of within and between sample variations helped to identify the minimum value of shared ranked variation between two samples that were derived from the same source. This analysis looked at multiple cases of homogenous strata sampled from different excavation areas (e.g. t ephra BWT sampled from the Block Excavat i on Area and TU 2) and unhomogenized, heterogeneous strata samples taken from a single area (e.g. N42 ULF which ha d numerous inclusions). For each type of strata (homogenous, different area or heterogeneous, same area) the results consistently showed that each sample gr oup shared at least 97% of the ranked variation meaning that R 2 values 97% were most likely derived from the same source. Therefore, all R 2 values described in the
93 text here that were were assumed to derive from the same parent material and thus regarded as the same stratigraphic layer. Magnetic Susceptibili ty Magnetic susceptibility was used to identify geomagnetic changes in the sediments that may be due to anthropogenic or geogenic factors. Samples were sieved, ground, air dried and packed into standard Bartington plastic sample boxes (2.5 cm diameter, 10 cc). Each s ample was subjected to a suit of mineral magnetic tests as per Herries and Fisher (in press) These tests included dual frequenc y and low temperature magnetic susceptibility measurements using the Bartington MS2 equipment and IRM (isothermal remanent magnetization) acquisition curves and backfields, hysteresis loops and thermomagnetic curves applying a Magnetic Measurements variab le field translation balance (VFTB) (Herries per comm.) Radiocarbon H. Wang at the University of Illinois at Urbana Champaign Geochronology Laboratory processed and dated the Moche Borago radiocarbon samples. The standard procedure acid base acid (ABA) pr etreatment was used for accelerator mass spectrometry (AMS) 14 C dating of Moche Borago charcoal samples. The same pretreatment was also applied to the 14 C free wood background sample and working standard wood samples, which included IAEA C5 (Two Creek fore st wood), FIRI D (Fifth International Radiocarbon Inter comparison D wood), and one ISGS 14 C dating lab working standard (Reily AC wood) samples. All samples were boiled for 1 hour in 2M HCl and rinsed to neutrality using deionized water (DI water). The samples were then soaked in 0.125 M NaOH for 1 hour and rinsed to neutrality using DI water. Thereafter, the samples were soaked again in 2M
94 HCl for 30 minutes and rinsed to pH 6 using DI water. The samples were dried overnight in an oven set at 70 C. T hen ~5 mg of each unknown sample and the background and working standard wood samples were placed into pretreated quartz tubes for sealed quartz tube combustion at 8 0 0C with 0.1 g Cu, 0.5 g CuO granules and a few grains of Ag foil. The quartz tubes had b een preheated at 8 0 0C for 2 hours and the CuO granules had been preheated at 8 0 0C one day before usage. The Cu grains and Ag foils were reduced using hydrogen gas under vacuum at 8 0 0C The combustion was programmed for 2 hours at 8 0 0C The samples wer e subsequently cooled from 8 0 0C to 60 0C for 6 hours to allow the Cu to reduce the NxO to nitrogen gas. The purified CO 2 was then collected cryogenically for AMS 14 C analysis P urified CO 2 was submitted to the Keck Carbon Cycle AMS Laboratory of the Uni versity of California Irvine for AMS 14 C analysis using the hydrogen iron reduction method. At the University of Illinois at Urbana Champaign, a split of purified CO 2 13 C values using the Finnegan 252 IRMS (isotope ratio mass spectr ometer) with a dual inlet device. Sample preparation backgrounds were subtracted based on the measurements of 14 C free wood. All results were corrected for isotopic fractionation, according to the conventions of Stuiver and Polach (1977) 13 C values measured on prepared graphite using the AMS. The AMS analysis indicated that all working standards of wood were within two standard deviations and all background wood samples were older than 53,300 14 C yr BP, which led us to believe that charcoal samples with 14 C ages from 48,850 to 41,580 were true ages and all other results were rel iable. The accelerator mass spectrometry results are provided in Table 6 1 (H. Wang per comm.).
95 Litho Stratigraphic Groups of the B lock Excavation Area The Block Excavation Area (BXA) was the main focus for both the GEPCA and SWEAP excavation teams. This excavation area, the most northerly at the site, occupies 24 sq m 2 being mostly within the upper 1 meter of deposits 14 The SWEAP excavations were focused close st on a 3 sq m area between excavation units G10, G09, and H09 and also almost exclusively on t he deeper Pleistocene deposits (Figure 5 2) Half of u nit G10 was first excavated by GEPCA as a 0.5 m x 1 m sondage to test the total depth of the archaeological deposits at Moche Borago. The SWEAP team continued these excavations in the remaining 0.5 m x 1 m of deposits and the depositional sequence became the backbone of the stratigraphic research at Moche Borago including naming conventions, litho stratigraphic groups and correlations. Next, t he stratigraphic units are described in groups De eper, older deposits are explained first and the discussion moves upward within each stratigraphic sequence. A composite profile, which shows the stratigraphic sequences and correlation of deposits between each excavation area, is provided in Figure 6 1. PKT Location: BXA Stratigraphic Units: PKT Age: Unknown The PKT deposit is the lowest within the BXA. It is a mottled pink (5YR6/2) consolidated ash deposit with abundant, poorly sorted clasts. The deposit was heavily consolidated and attempts to excav ate it required a hammer and chisel. The deposits were first described by Gutherz (Gutherz et al., 2001) as possible bedrock. This seems 14 This calculation does not consider the connection made between the GEPCA Test Unit 1 and Test Unit 3 made in 2001. Was this area to be cons idered, the BXA surface area becomes 36 sq. meters.
96 unlikely because PKT appeared to be dissimilar to the parent rock in the rockshelt er walls. The total depth of the PKT deposit is still unknown. In 2008, a 0.5 m x 0.5 m test pit was dug to a depth of 1 m below the top surface of PKT but little change was noted within the deposit. PKT was the only stratigraphic unit at Moche Borago which was not aggregated into a larger litho stratigraphic group due mainly to the unique character istics of this layer The x ray fluorescence spectra of PKT show ed the highest recorded amount of rubidium across all stratigraphic samples from each excav ation area collected from the site. Iron and z irconium concentrations, generally found in high amounts across the site, were also lower in PKT. By comparison, t he iron concentration is much higher within the overlying T Group deposits when compared to PK T suggesting that the T Group did not derive from the PKT deposit. T Group Deposits Location: BXA Stratigraphic Units : VHG, CTTC, CTT, MHD, and DCC Age: >53 to 45 ka Located stratigraphically above PKT was a series of sandy silt (VHG, CTT, MHD) and claye y (DCC) layers that represent ed a period of dense human occupation at the site These layers were collectively aggregated into the T Group stratigraphic unit, which was bracketed by tephra YBT ( above ) and ( PKT ) below. The T Group could be subdiv ided into Upper and Lower, based on sedimentological and chronological grounds. The T Group Lower (VHG, CTT, CTTC) deposits were characterized by hard, gravelly ashes and clayey sands that appeared to be more heavily weathered than overlying deposits. Th e age of the T Group Lower has been minimally dated 53,224 2,662 cal. BP because we have currently dated only a
97 single sample, which came from CTT in the uppermost T Group Lower deposits. The lower deposits which are mainly represented by VHG, remained undated. The T Group Upper deposits (MHD and DCC) w ere characterized by an increased clay content and red coloration of the sediments. The T Group Upper included numerous stratified hearth deposits that showed a banded profile in the excavation unit wa lls. While the hearth deposits represented sedimentological deposition, the anthropogenic process by which they were deposited blurs the boundary between purely litho stratigraphic and culture stratigraphic classifications. The weighted mean age 15 of the T Group Upper was estimated to be 45,164 982 cal. BP using radiocarbon. Layer d escriptions Micromorpholog ical analysis show ed that the contact between PKT and the lowest T Group deposit, VHG, was very sharp. Currently, we are unsure if the contact bet ween PKT and VHG is unconformable. VHG was a layer of moderately abundant gravels in a hard, mottled clay matrix containing microscopic clasts of variably weathered volcanic ashes, charcoal, and reworked clay fillings suspected to be from water in the ca ve. The upper surface of VHG was eroded and capped by CTT, a compacted, localized ash lens in G10 surrounded by a clay berm (CTTC). CTT may have be en disturbed. N umerous irregular clasts of weathered tuff, charcoal grains, iron stained bone fragments, and organic matter were revealed in micromorphology (P. Goldberg per comm.) CTT was likely a basal sub deposit associated with the MHD hearth features found stratigraphically above it. Charcoal recovered from CTT was dated directly to 53,224 15 All weighted mean ages presented here rely on the Central Age model (Galbraith and Laslett, 1993; Van der Touw et al., 1997; Galbraith et al., 1999; Galbrait h et al., 2005)
98 2,662 ca l. BP, which is currently the oldest and deepest absolute age at the site. The underlying 20+ cm of archaeological deposits which are mainly VHG are currently undated. MHD is a series of stratified archaeological hearth deposits that indicate d a repea ted human presence in this section of the rockshelter during early OIS 3, ~53 to 45 ka The MHD deposit was characterized by numerous compacted aggregates of fine, iron rich illuvial clay bands deposited through colluvial activities. The upper surfaces of some of the clay bands appear ed to have been moderately eroded and stabilized (P. Goldberg, per comm.) Stratigraphically above MHD was the slightly thicker layer DCC. DCC was the topmost layer of the T Group. The XRF analysis show ed that DCC ha d the highest iron content 16 of the T Group deposits, including the overlying YBT tephra. The weighted mean radiocarbon age for DCC was estimated at 45,164 982 cal. BP which provides the upper age estimate of the T Group deposits Occupational Hiatus #1 Loca tion: BXA Stratigraphic Unit: YBT Age: ~43 ka The first major occupational hiatus at the site was found within the tephra YBT that overl ay DCC and cap ped the T Group. YBT was a ~20 cm thick yellow brown (2.5YR6/3), prominently graded ash that was variabl y intermixed with soil aggregates (sandy silt) (P. Goldberg, per comm.) Statistical comparisons using whole XRF spectra show ed that the chemical composition of the YBT tephra is very similar to that of the 16 According to XRF analysis, the iron content appeared to increase incrementally within the T Group, starting with the lowest iron abundance values at the base, VHG.
99 BWT tephra dating to the early Holocene, R 2 = 0. observation was further supported by mineral magnetic analysis which also point ed to similar sources for these layers ( A. Herries per comm. ). BWT and YBT share d similar reversible thermomagnetic curves which are also un ique compared to other strata at the site. 17 Currently, the source for BWT and YBT is unknown though it is presumed to be local BWT was correlated chemically via XRF to open air tephra up to 18 km northeast of Moche Borago and Mt. Damota. Charcoal was found in YBT, which may have become carbonized as the hot ash was deposited into the site. Multiple radiocarbon samples from YBT date d this volcanic event between 41 and 45 ka providing a weighted mean, calibrated radiocarbon age of 43,403 1,213 cal. BP. The duration of the YBT volcanic event is still undetermined. Artifact densities were high before and after YBT while the tephra layer itself was effectively sterile (1 stone artifact was recovered in YBT from unit G10). S Group Deposits Location: B XA Stratigraphic Units : OBMB, LVDBS, LMGV, VDBS Age: 44 to 43 ka Above YBT were a series of distinct stratigraphic units which form ed the S Group aggregate in the BXA. The S Group is believed to represent a single, rapid, and continuous depositional con text in between two major volcanic episodes (YBT and YBS) at Moche Borago. This interpretation is supported by XRF, mineral magnetic analysis and radiocarbon results. The XRF and mineral magnetic analysis show ed that each of the 17 The magnetic mineralogy between BWT and YBT differed slightly, but overall test results showed that both layers were dominated by small, stable, single domain ferromagnetic grains.
100 S Group layers ha d sligh tly varying chemistry but these deposits were similar overall 18 The mean R 2 which analyz ed full XRF spectra of all S Group samples was 0.967 (SD = 0.188, p < .01). Frequency dependence of magnetic susceptibility (X FD %) also show ed a continuity of fine grained magnetite across the S Group de posits which was also unlike the magnetic signature of the underlying YBT tephra ( A. Herries per comm ) The dating of the S Group deposits was also precise. Three radiocarbon samples taken from a ~20 cm vertical span within various S Group layers provid e d a tightly constrained weighted mean age of 43,480 443 cal. BP. Layer d escriptions At the base of the S Group ( and directly overlying tephra YBT ) was OBMB, a mottled blend of ash and silty clay. The lateral extent of OBMB is currently being questio n ed because this layer is well represented in G10 but nearly absent elsewhere in the BXA and completely absent in other excavation areas The limited extent of OBMB might indicate that it was incised or removed by natural or anthropogenic process. Micr omorphology show ed that OBMB is a massive, weakly bedded tuff with lenses of soil aggregates rich in organic matter (P. Goldberg, per comm.) Intersecting OBMB and the underlying tephra YBT was LVDBS, which is loosely consolidated dark brown silt. LVDBS was interpreted originally as a pit feature cut into OBMB and YBT due mainly to micromorphological analysis and a lack of abrasion on the abundant lithic materials found within the stratigraphic unit Micromorphology seem ed to support this interpretation and show ed an unsorted heterogeneous mix of charcoal, phytoliths, plant 18 In contrast, correlations to YBT were elementally very dissimilar from the S Group, and particularly for Rare Earth Elements (REE). YBT had much larger zirconium counts with moderately larger counts of rubidium, yttrium, niobium, and molybendum. Strontium appeared all but absent in YBT as opposed to S Group deposits.
101 tissues, and bone fragments which suggest ed colluviation and possible trampling of these deposits (P. Goldberg, per comm.) However, it is equally likely that these processes were mo re indicative of the sample location ( near the top edge of the pit) feature rather than being indicative to the feature itself In 2007 and 2008 new evidence suggested we re look at LVDBS and our original interpretation that it was an anthropogenic pit. The reinterpretation o f LVDBS now suggests that it is a fluvial feature, which is consistent with similarly aged fluvial features found across the site. Furthermore, in nearby excavation unit H09 a much more likely anthropogenic pit was excavated in 200 8 in equivalent S Group deposits 19 and it was very dissimilar to the possible G10 LVDBS pit feature. First, the H09 pit is much smaller than the LVDBS pit, being ~40 cm diameter and 10 cm deep. The LVDBS feature was at least 70 cm wide and more than 30 c m deep. Second, the H09 pit was formed within a relict hearth feature that included an underlying patch of burnt, hardened clay ( layer HCLMGV). N o underlying discolored or hardened sediment in or around LVDBS was found to suggest it was ever excavated as a hearth. The subsequent fill deposits within the H09 pit ( layer LMGV) were loosely consolidated silty clay sediments with abundant lithics, charcoal, and the highest concentration of heavily burnt, friable bone found in the BXA. Here, we can see some s imilarity with LVDBS as indicated by the micromorphology and excavations that is, an abundance of lithics, charcoal, and bone overall. But much of the bone and lithic materials from H09 were oriented vertically near to 90 most likely due to repeated e xcavation and mixing within the feature. This pattern was not seen in LVDBS. 19 The H09 pit was found in LMGV which underlay LVDBS in H09. The base o f the H09 pit was HCLMGV and the next major stratigraphic unit underneath appeared to be OBMB.
102 LVDBS therefore seemingly represent s low energy channel fill rather than an anthropogenic pit. The H09 feature, on the other hand, i s currently believed to represent the use o f an in situ hearth that was dug out repeatedly and re used. Further support for the LVDBS fluvial interpretation was found in the stratigraphically equivalent S Group deposits in TU2 and N42 and clear cut and fill activity was associated within the lowe st S Group deposits and the underlying YBT tephra. Furthermore, stone artifacts found in the active stream channel in nearby Esay rockshelter provide a modern analog to the past fluvial features at Moche Borago. At Esay, stone artifacts found in the shal low stream channel were also not rolled because it is a low energy fluvial environment Therefore, at Moche Borago, a brasion may not be a clear signal for fluvial activity in the BXA. The overlying VDBS deposits consist ed of dark silts with abundant sto ne tools, bone, and charcoal. VDBS appear ed to cap the channel fill, but this conclusion is not yet definitive. Occupational Hiatus #2 Location: BXA Stratigraphic Unit: YBS Age: 43,121 692 cal. BP The second volcanic event, called YBS, overl ay VDBS t o cap the S Group Charcoal taken from the YBS ash directly date d this deposit to 43,121 692 cal. BP. YBS was a series of yellow brown (2.5YR5/3), archaeologically sterile ash layers 10 to 20 cm thick interstratified with darker, thinner lenses. Com pared to other major tephra at Moche Borago, YBS ha d a unique elemental and magnetic signature. The elemental composition of YBS show ed lower i ron and Rare Earth Elements (REE) (especially z irconium), except for s trontium which is much higher in YBS. Sta 2
103 values (p < .01) show ed that YBS and BWT share d only 92% of their ranked variability whereas correlation between YBS and YBT account ed for only 88.6% of the ranked variability Being less than 97%, these R 2 values suggest ed that the YBS tephra did not share the same parent material as either BWT or YBT. Mineral magnetic analysis provide d additional information on the depositional history of the YBS tephra. These data s howed that YB S contain ed titanium, which occur red rarely in sampled Moche Borago sediments (the overlying R Group had slightly elevated titanium abundance, however). The X LF (magnetic susceptibility) values of YBS were higher than other tephra at Moche Borago and simil ar to the magnetic characteristics of most other clay and silt based strata at the site. Currently, the only explanation for the titanium within YBS is that this tephra was deposited originally outside of Moche Borago in a titanium rich environment and th en secondarily weathered and transported into the rockshelter (A. Herries per comm.) R Group Location: BXA Stratigraphic Units : RGCB, RCGA, RCA Age: ~41 ka Positioned stratigraphically above YBS was the R Group composed of the RGCB, RGCA, and RCA depos its. The R Group sediments were among the most homogenous the mean ranked variability between R Group deposits was 96.7% (p < .01; SD = 0.6). The R Group sediments we re characteristically red to reddis h brown clays and silty clays. The distinct red color of the R Group sediments suggest ed a change in the oxidation of the iron within these deposits, perhaps due to sub aerial exposure as a paleosol. Round to sub round e d inclusions var ied in abundance but the inclusions were
104 present throughout the R Group, as were abundant, unabraded archaeological materials. R Group sediments share d similar X LF (magnetic susceptibility) and X FD % with the T Group deposits which might indicate that these sediments were deposits under similar climatic contexts (A. Herries per comm.) Currently, the R Group has 11 radiocarbon age s, which were derived from charcoal in the BXA and TU2. The weighted mean age was 41,159 783 cal. BP. Thi s age accords well with the overall Pleistocene chronology at Moche Borago. Late and Terminal Pleistocene Un conformity Location: BXA Stratigraphic Units: R Group to BWT Age: Terminal Pleistocene/OIS 2 Above the uppermost R Group deposit (RCA) was the te phra BWT, which is described in more detail in the next sub section (see Occupational Hiatus #3 : BXA H Group). This ash was a marker tephra for GEPCA (GEPCA called it Tephra 11 and it remain ed one of the key marker tephras for the SWEAP team The impo rtance of BWT was its age. Although t he GEPCA team recognized that Holocene sediments overl a y BWT, they did not date BWT directly. The SWEAP team collected samples of charcoal within BWT from TU2, G9, and H9 for radiocarbon analysis. These samples provi de d a weighted mean age of 8,051 594 cal BP that confirm ed the early Holocene age of the tephra Originally, the R Group deposits underlying BWT were believed to date to the Terminal Pleistocene. The basis for this hypothesis was a GEPCA radiocarbon sample dated 28,700 1,100 BP which was noted in 2000 report (Gutherz et al., 2000a) It is u nclear if this age was calibrated and, if so, what radiocarbon calibration curve did GEPCA use. The position of th e GEPCA sample (Gutherz
105 et al., 2000a) profile drawings appear ed to be from sediments underlying Tephra 11/BWT (presumably R Group). S ubsequent radiocarbon analysis by SWEAP since 2006 consistently show ed that the R Group was m uch older, ~41 ka. The gap in time from the R Group (~41 ka) to BWT (~8 ka) suggested a major un conformity that effectively spanned the Terminal Pleistocene/OIS 2. Currently, this gap in the sequence is the only known major (i.e. > 10,000 years) un confor mity within the Later Pleistocene deposits at Moche Borago. Equivalent strata in TU2 also revealed an un conformity (i.e. TU2 BWT and RGX). The cause of the un conformity is currently unknown. Volcani c activity and fluvial action are both being consider ed. Fluvial action may be supported by micromorphology, which shows that RCA ( the uppermost R Group deposit ) ha d a highly porous microstructure with voids filled by oriented and laminated reddish orange clay (P. Goldberg, per comm.) The contact with the overlying BWT tephra was also sharp but the base of BWT was weathered and iron stained. Water flowing at the contact between these units may explain the discoloration, but no evidence existed for post depositional movement of archaeological materials. T he 3D plots of archaeological and faunal materials from the R Group deposits in G09, H09, and I10 reveal ed fine vertical stratification consistent with continuous, undisturbed deposition of the recognized stratigraphic layers and hearth lenses. The hearth s, in particular, were denoted by the absence of lithics. If vertical movement of archaeological materials occurred when the Terminal Pleistocene deposits were eroded, we would not expect vertical spatial patterning. Furthermore, vertical movement might create lag deposits of artifacts, especially at the contact with BWT, or size sorting, neither of which was evident.
106 Occupational Hiatus #3/BXA H Group Location: BXA Stratigraphic Units: BWT Age: Early Holocene The other hypothesis to explain the Termin al Pleistocene un conformity is volcani c activity, which can also be associated with increased precipitation So perhaps it is prudent to jointly consider both ideas. The evidence for a major volcanic event which may have removed the Terminal Pleistocene deposits was found in the presence of the BWT tephra and the H Group, which were found in all three excavation areas. Layer Descriptions H Group deposits were found in all current excavation areas. The H Group was best represented in the BXA by stratif ied ash, lapilli, and bombes found in the north GG profile of units J/K 11 13 (GG TU3), TU2, and N42 and also BWT The H Group was unique for the size of pyroclastic materials and also for the apparent lateral size sorting within these deposits. In N42, the H Group was composed mainly of bombe and lapilli clasts, but no clear BWT ash was present. In TU2, larger bombe and lapilli clasts of the H Group gave way to smaller lapilli and ash, and a clearly defined BWT ash layer was also found. In BXA, lapill i and bombes were infrequent and the BWT ash was thicker. Currently, we believe that BWT was one event within the large H Group volcanic series. BWT was always found in association with H Group deposits (overlying). BWT was a 20 cm thick, dense, homoge nous, and archaeologically sterile white ash dating to the early Holocene. BWT was among the clearest marker tephras throughout the site deposits. Tentatively, t his tephra has been correlated chemically via XRF to tephra up to 18 km NE of Moche Borago an d Mt. Damota.
107 Stratigraphic S ummary of TU2 Approximately 8 m east southeast of the BXA, TU2 was re opened by SWEAP in 2006 to observe the stratigraphic profiles described by Gutherz et al. (2001) and int egrate these data into the SWEAP database. Aspects of the stratigraphy were reminiscent of both the BXA and N42 areas. In 2007, SWEAP extended the TU2 excavations 1 m to the south to increase the profile length and archaeological sample size from this ar ea. These excavations yielded high densities of lithics within R Group sediments a s well as additional supporting geomorphic evidence for fluvial channels. L Group Location: TU2 Stratigraphic Units: LFX 1 and 2 Age: >45 kya? At the base of TU2S (the d eepest of the two TU2 units) was LFX1, a lahar deposit composed of numerous sub angular clasts of ash and other detritus suspended within an orange (7.5YR5/3) homogenous matrix. Above LFX1, LFX2 was marked by larger clasts of ash but retain ed much of the coloration and composition of LFX1. Both LFX1 and LFX2 were archaeologically sterile. The XRF correlation analysis (p < .01) using 2 show ed that the TU2 L Group share d more than 97% of ranked variation with the L Group deposits identified in N42 (ULF and LLF). The L Group deposits, however, have not been identified within the BXA. Both LFX1 and LFX2 ha d been truncated and in the case of LFX1, undercut LFX2. The cause of the truncation is believed to be fluvial erosion though detailed ana lyses are pending. A single radiocarbon date of a thin lens (BGC) between LFX2 and the overlying tephra YBSX date d these L Group deposits minimally to 44,888 786 cal. BP.
108 S Group Location: TU2 Stratigraphic Units: BGC Age: ~43 ka The S Group was rep resented in TU2 in only a single, thin lens of clayey sediments (BGC) that contain ed an abundance of gravel inclusions and abraded lithics. The gravel inclusions were poorly sorted and had rounded to sub rounded edges. A single c harcoal sample collected from BGC provide d a radiocarbon age of 44,888 786 cal. BP for this layer. This age was similar to the weighted mean age of the BXA S Group (43,480 443 cal. BP) T he shared post depositional context made it more likely that the intrusive TU2 L Group f luvial channel was contemporary with the BXA S Group fluvial channel (LVDBS). Occupational Hiatus #2 Location: TU2 Stratigraphic Units: YBSX Age: 42 to 43 ka Stratigraphically above LFX2 was tephra YBSX. This thick ash and lapilli lens was in fill wit hin the TU2 L Group fluvial channel. Radiocarbon dating of charcoal from the YBSX tephra provide d a single date of 42,776 526 cal. BP. Taken together with the BGC radiocarbon age, YBSX likely was deposited sometime between 44,888 786 cal. BP and 42,7 76 526 cal. BP This age placed YBSX concurrent to the BXA S Group and tephra YBS deposits. The XRF analysis support ed a correlation between TU2 YBSX and BXA YBS. These layers shar d 97.3% ranked variation (p < .01). YBSX also share d > 97% ranked vari ation with BXA LVDBS, and nearly as much with VDBS (96.4%). R Group Location: TU2
109 Stratigraphic Units: RGX Age: >Early Holocene to <~40,000 Stratigraphically above YBSX was RGX. RGX ha d a similar red color and composition to the R Group found in the B XA. While it is currently correlated into the R Group based on composition, dating, and stratigraphic placement, neither XRF nor ICP MS show ed TU2 RGX and BXA R Group deposits. The mean R 2 value with BXA R Group deposits (RCA, RGCA, RGCB) indicate d only 96% share variation within ranks (SD = 0.9, p < .01). TU2 RGX share d > 97% ranked variability with TU2 Y BSX (97.6%), BXA YBS (97.4%) and LVDBS (97.4%) suggesting that it may have derive d some of its sedimentology from these underlying deposits. The Rare Earth Elements ( REE ) and i ron counts were also similar between RGX and these layers. Furthermore, RGX w as notable for containing high frequencies of heavily abraded lithics. Data also suggest ed that RGX was deposited via fluvial processes. Unlike the probable low energy fluvial event seen in the BXA (LVDBS) fluvial processes in RGX may have been more ene rgetic, as evidence d by the abrasion on the lithics. The RGX channel may have also re deposited lithics from other areas of the site or washed away major portions of the original R Group sequence of sediments in TU2 leaving only the heavier fractions of the deposit behind. Occupational Hiatus #3/TU2 H Group Location: TU2 Stratigraphic Units: HEP1 and 2 Age: Early Holocene Stratigraphically above RGX was the TU2 BWT tephra. The XRF R 2 (p < .01) analysis indicate d that 98.7% of the ranked variability w as shared with BXA BWT providing one the highest correlations between Moche Borago sediments. Independent
110 radiocarbon samples were all similar between BXA and TU2 BWT samples and each of these dates was incorporated into the BWT weighted mean age of 8 ,051 594 cal BP. Two layers were stratigraphically above TU2 BWT that were part of the Holocene H Group volcanic event (HEP1 and HEP2). Both TU2 H Group layers were composed of ash and lapilli. HEP2 appears to have a higher frequency of lapilli than the overlying HEP1. Stratigraphic S ummary of N42E38 The stratigraphy of the southernmost excavation area is currently defined using the sequence in N42E38, a 1 x 1 m pit ~8 m southeast of TU2 and ~16 m southeast of the BXA. N42E38 was excavated > 2 m b elow the surface and the majority of deposits appear ed to be volcanic (both tephra and lahar). N42E38 was unique because nearly all deposits were archaeologically sterile, except for the most recent and topmost 5 to 10 cm of sediment and a series of silt and clay channel fill deposits, known as the N42 S Group Also, N42E38 was unique for being located on a small rise (~20 cm above the surface of TU2 and BWT) near the rear rockshelter wall. The geology of this area is currently poorly understood Howev er, the massive volcanic deposits found in N42E38 may have backed up against the rockshelter wall and built up the shelter floor in this area. It is currently not known w hy these thick deposits are not found in the other excavated areas of the site. Curre nt Basal Sediments Location: N42E38 Stratigraphic Units: MPT, SRS, and DRB Age: Undated MPT is the current base of the N42E38 sequence. This lens was undated and archaeologically sterile. Above MPT was speckled red sand (SRS) with dark reddish
111 brown sand (DRB) above it. SRS is reminiscent of similar ly speckled red sand found stratigraphically between YBSX and LFX2 in TU2 (TU2 SRS) 20 XRF analysis show ed that N42 SRS and MPT share d > 97% ranked variation (p < .01) with TU2 SRS as well as with the overlying TU2 LFX1 and LFX2 deposits (TU2 L Group). T his correlation may be misleading because the N42 L Group deposits (LLF and ULF) also share d >97% ranked variation with the TU2 L Group deposits. Therefore, the N42 basal sediments are currently uncorrelated la terally across the site. L Group Location: N42E38 Stratigraphic Unit: LLF, ULF Age: Undated Two distinct lahar features above the N42 basal sediments constitute d the N42 L Group deposits. The L Group ha d two subdivisions : lower lahar feature LLF and upp er lahar feature ULF. LLF exhibit ed numerous inclusions of semi angular ash and lapilli and other detritus up to ~20 cm diameter within a fairly homogenous orange (10YR4/4) clay matrix. ULF was similar to LLF but ha d more abundant inclusions that were mo re angular in shape. Both ULF and LLF share d >97% ranked variation (p < .01) with the TU2 L Group (LFX1 and LFX2). While the TU2 L Group deposits appear ed to contain smaller sized inclusions than those seen within the N42 L Group deposits, similarities o ccurred overall in texture, composition, color, and hardness that strongly suggest ed these deposits derive d from the same event. If so, then the radiocarbon sample from the top of TU2 LFX2 provide d a minimal date of 44,888786 cal. BP for the N42 L Group. 20 TU2 SRS is a thin, red lens that was found only in isolated areas within TU2S. It did not appear to represent a contiguous stratigraphic layer, and therefore it was not included within the descriptions.
112 The L Group deposits between N42 and TU2 were also likely altered post depositionally by fluvial activity. The top of ULF ha d been truncated sharply and clasts within ULF were visibly sheared off. The east and west N42 profile walls show ed that the truncation occur red on an approximately 30 irregular slope facing the rear cave wall. The north profile walls reveal ed a shallow (<20 cm) semi circular depression at the base of the slope. In filling this shallow depression were the N42 S Group deposits S Group Location: N42E38 Stratigraphic Units: COL, LPFC, MPFC, UPFC Age: Undated. Possibly ~42 to 45 k a based on correlation Overlying the L Group is t he lowermost S Group deposit in N42 a thin lens of manganese rich sediment, labeled COL Original ly, COL was thought to be organic, and the algal mat at the base of the shallow stream in nearby Esay rockshelter was suggested as a likely analog H owever, attempts to radiocarbon date COL showed no organic materials present (Hong Wang, per comm.) and XR F analysis thereafter indicated high amounts of manganese, which could be due to fluvial deposition. As with the S Groups in the BXA and TU2, the N42 S Group was a series of silt and clay channel fill deposits ( layers LPFC, MPFC, UPFC). Each of the lay ers record ed a different phase of stream activity. We can see a distinct fining upward size sorting of abraded cobbles between these layers with LPFC containing the highest density and largest cobble size. Infrequent but heavily abraded, archaeological m aterials were found within these deposits supporting the fluvial hypothesis on the one hand and, on the other hand the correlation to other S Group deposits at Moche Borago. These correlations were further supported by XRF analysis. LPFC, MPFC, and UPF C each share d >97% ranked variation (p <.01) with channel fill LVDBS in the BXA.
113 Occupational Hiatus #3/N42E38 H* Group Location: N42E38 Stratigraphic Units: HEP1 and 2 Age: Undated, possibly Early Holocene based on correlation Stratigraphically above t he S Group were two massive tephras (~1 m total thickness) that constitute d the N42 H Group (HEP1 and HEP2). The N42 H Group contained abundant bombes within a loosely consolidated lapilli and ash matrix. The larger fraction between the two layers was fo und within the lower unit (HEP1). HEP2 was less consolidated and numerous large (5 20 cm) cavities were found between the clasts. A Working Model of the OIS 3 Depositional History at Moche Borago Rockshelter Moche Borago was occupied at least 53 5 1 ,000 y ears ago but occupation also likely occurred earlier since the lowest 20 40 cm of archaeological deposits remain undated Concurrent ice core records showed that across the Northern Hemisphere, high latitude cold and arid conditions during H einrich (H) 6 ( ~60 ka) w ere followed in rapid succession by Dansgaard Oeschger (D O) events 17 (~59 ka) 16 (~58 ka) and 15 (~56 ka) (Grootes et al., 1993; Sowers et al., 1993; Meese et al., 1994; Stuiver and Grootes, 2000; North Greenland Ice Core Project, 2004) The intensity of these D O events (17 15) was declining sequentially after H6, but the magnitude of annual temperature increases above Greenland was still ~9 to 10 C above current conditions (Wolff et al., 2010) As Chapter 2 detailed, the Ethiopian highlands were experiencing wetter conditions concurrent to the D O events and sapropels of OIS 3, and these conditions may have been influenced by regional monsoonal flux. Paleoclimatic and paleoenvironmental records from the coast of West Africa, North Africa, the Mediterranean Sea, and the circum
114 Indian Ocean region all show that pluvial events, including D O events and sapropels, during OIS 3 are associated with coeval increases in African and SW Asian monsoonal intensity and precipitation (Schulz et al., 1998; Bar Matthews et al., 2000; Burns et al., 2003; Burns et al., 2004 ; Weldeab et al., 2007; Revel et al., 2010) The formation of Mediterranean sapropels and the Nile paleohydrology records are linked to increased precipitation in the Nile source region, which is the Ethiopian highlands (Bar Matthews et al., 2000; Revel et al., 2010) ~60 ka to 45 ka The T Group deposits are among the least well understood at Moche Borago, which makes it difficult to do comparisons of these layers to the broader climatic context of the Horn at this time. T he basal layer ( PKT ) is still un correlated to other layers and the chronology of the T Group Lower is only very broadly defined. Also, a ~5,000 year gap exists between the upper error age estimate of the T Group Lower and the lower age er ror estimate s of the T Group Upper. What little contextual information that can be gleaned from the T Group suggests that this group began during a wet phase and ended during another wet phase. The basal deposits of the T Group (VHG, in particular) inclu de d reworked clay fillings indicating that these layers were deposited during humid conditions The T Group Upper deposits (esp ecially MHD) presented numerous aggregates of fine, iron rich illuvial clay bands (P. Goldberg, per comm.), showing that these deposits were also laid down in a humid environment. However, if a strict interpretation of the chronology is adhered to, then t he current age for the T Group Upper is coeval to H5 The dating of the T Group Upper deposits is substantiated by three radi ocarbon samples that show continuity from the T Group
115 Upper into YBT and the S Group. Therefore, if I were to speculate broadly about the T Group, I would place the T Group Upper contemporary to D O 12 (46.8 ka). The age of D O 12 is within the statisti cal margin of error for the current weighted mean age estimates of the T Group Upper, and correlation to D O 12 would also explain the evidence that these deposits collected in wetter conditions. I am hesitant to speculate about the chronology of the T Gr oup Lower at this time because too much uncertainty still surrounds these deposits. 45 ka to 43 ka A fter 45 ka a volcanic eruption laid down the first major OIS 3 tephra found at Moche Borago (YBT). N o evidence suggest s human occupation at the site dur ing the time of the eruption. The volcano deposited ~20 c m of homogenous ash and lapilli across the site which would have made the rockshelter uninhabitable for an unknown duration of time The YBT deposit is characteristic of the pyroclastic volcanic a ctivity found across the western margin of the Main Ethiopian Rift. The majority of volcanoes here share a felsic magma composition that has produced numerous explosive ignimbrite and rhyolite deposits across the region (Di Paola, 1972; WoldeGabriel et al., 1990) The source of the YBT tephra is currently unknown but XRF analysis suggest s that YBT shares the same source as the Holocene aged BWT tephra. The timing of the YBT tephra is linked closely to the overlying S Group and tephra YBS deposits. These layers were apparently deposited within one millennium of each other, ~43 ka. 21 At this time (~43 ka), substantial evidence suggests that climatic 21 YBT is dated to 43,403 1,213 cal. BP The overlying S Group deposits are dated 43,480 443 cal. BP. YBS is dated 43,121 692 cal. BP. YBT and YBS do not share a similar source, according to XRF analy sis.
116 conditions around Moche Borago were among the wettest in our OIS 3 record. This evidence is drawn mainly from fluvial geomorphic features found in each excavation area: 1 ) in the BXA, the S Group LVDBS channel; 2) in TU2, the S Group channel that cut into the L Group; and 3) at N42, the L Group was truncated with overlying S Group channel fill deposits. Between t he N42 S Group and LVDBS in the BXA these deposits share d > 97% ranked variation (p <.01) making the S Group correlation between each excavation area at Moche Borago among the most secure. In the BXA, LVDBS (S Group) represented channel fill deposits that were cut into OBMB and tephra YBT. Sections of YBT were even missing in the BXA, and we currently believe this was cut and fill activity from the LVDBS channel. Even clearer evidence exists for fluvial activity in TU2, though the age is less certain. Fluvial activity appears to have truncated and undercut the TU2 L Group. The age of the fluvial activity here can be limited to the time period 45 to 43 ka. Radiocarbon dating of layer BGC between YBSX (above) and the L Group (below) dates the fluvial cutting to a maximum of ~45 ka. The minimal age of the chann el cutting (~43 ka) can be inferred via stratigraphic correlation between tephra YBSX (TU2) and YBS (BXA). BGC is believed to be remnant S Group deposits that contained an abundance of heavily abraded lithics which may indicate increased stream activity, rolling, and fluvial abrasion here (unlike in the BXA). In N42, the most well defined channel fill deposits at Moche Borago (LPFC, MPFC, UPFC) unconformably overlay t he undated upper L Group deposit (ULF). The lowermost fluvial unit, LPFC, contain ed th e highest density and largest cobble size of the group, and we can see a distinct fining upward size sorting of abraded cobble s
117 Infrequent but heavily abraded archaeological materials were also found within these deposits supporting another instance of hig h energy fluvial deposition on the one hand and, on the other hand the correlation to other S Group deposits. Therefore, from ~ 44 ka to ~ 43 ka, a series of high and low energy fluvial features were present at Moche Borago The source of the water is likely due to percolation from joints within the rear cave walls which is similar to the current stream conditions at Esay rockshelter It is unknown if the channels seen in each excavation area were interconnected from a single source or multiple sourc es. Based on the morphology of the channel features, it appears that the channels were spatially limited phenomenon at the site and the energy of the stream flow was lower around the BXA and greatest at N42. It seems highly unlikely that the channels dist urbed the entire site and I speculate that the source of the freshwater at this time may have been a major incentive for occupation of the site at this time. This same time period coincides almost precisely with wetter and warmer climatic conditions tha t would have occurred during D O 11 (43.4 ka). The D O 11 event is evident in both high latitude ice cores and regional records surrounding the Horn of Africa. Above Greenland, D O 11 is associated with the greatest increase in mean annual temperatures kn own during OIS 3, 1 5 C (Wolff et al., 2010) Across Africa, monsoonal activity may also have increased at this time. West African marine core records indicate that monsoonal precipitation was greater at 43 ka just prior to a dry event at 42.5 ka, though the changes are subtle (S. Weldeab per comm.). The Moomi cave records show a similar, subtle pattern of monsoonal increase ~43 ka (Burns et al., 2003; Burns et al.,
118 2004) Nile paleohydrology records also indicate a broad period of pluvial conditi ons at this time, which show wetter conditions in the Ethiopian highlands (Revel et al., 2010) Conclusion s Fluvial channels are known only in the S Group deposits at Moc he Borago even though other litho stratigraphic units, such as the T Group Upper, may be associated with wet climatic conditions. The presence of the channels suggests that the climatic context of the S Group was significantly wetter than other periods in our OIS 3 records from Moche Borago. The timing of the S Group in the BXA is among the most precise and accurate of all litho stratigraphic groups, and this chronology is further supported by the ages for the YBT tephra (below) and YBS tephra (above). R egional and local evidence for monsoonal intensity increase is not as robust for D O 11 as it is for other millennial scale events. However, based on the available facts, I am confident that the S Group deposits show localized evidence for increased monso onal activity due to D O 11.
119 Figure 6 1: This composite image show s the stratigraphic profiles from the BXA (left), TU2 (center) and N42 (right) excavation areas. Selected calibrated radiocarbon ages are placed in their approximate 2D position alo ng the profile walls. Selected climatic curves are represented at the bottom for comparison to the Moche Borago chronology (Sowers et al., 1993; Grootes et al., 1993; Meese et al., 1994; Stuiver and Grootes, 20 00; Wang et al., 2001; Burns et al., 2003; Burns et al., 2004; Cai et al., 2006)
120 Table 6 1: The AMS R adiocarbon A ssay from Moche Borago Bag Number Sample # XU Strat Unit Level d31C pMC D14C 14C age Cologne CalPal, BP 4711 SWAP07 01 N42E 36 HEP 9 25.5 0.9835 0.0014 16.5 1.4 135 15 133 101 2575 SWAP07 02 G9 RCA 2 23.7 0.3497 0.0007 650.3 0.7 8440 20 9479 12 2200 SWAP07 03 H9 RCA 15 23.7 0.0128 0.0004 987.2 0.4 35010 270 40059 847 2290 SWAP07 04 TU2S OST 6 25.1 0.4156 0.0008 584.4 0.8 705 5 20 7902 30 2596 SWAP07 05 H9 RCA 18 23.3 0.0089 0.0004 991.1 0.4 37930 370 42397 384 3334 SWAP07 06 H9 RGCA 19 22.9 0.0057 0.0004 994.3 0.4 41580 590 45082 801 4039 SWAP07 07 H9 VDBS 26 28.9 0.0076 0.0004 992.4 0.4 39200 440 43351 601 2261 SWAP 07 08 TU2S YBSX/ BWT 11 24 0.0085 0.0004 991.5 0.4 38320 390 42776 526 2105 SWAP07 09 G10 YBT 17 26.8 0.0055 0.0004 994.5 0.4 41830 600 45296 842 2206 SWAP07 10 TU2S BGC 16 26.9 0.0058 0.0004 994.2 0.4 41370 570 44888 786 2117 SWAP07 11 G10 DCC 19 23.7 0.0051 0.0004 994.9 0.4 42400 650 45857 1067 3112 SWAP07 12 G10 DCC 23 22.2 0.0069 0.0004 993.1 0.4 39920 480 43704 636 3123 SWAP07 13 G10 CTT 25 24.2 0.0023 0.0004 997.7 0.4 48850 1420 53224 2662 3135 SWAP08 1 G10 DCC 24 24.3 0.0374 0.0008 962.6 0.8 26400 180 31198 360 2109 SWAP08 2 G10 YBT 18 23.3 0.0026 0.0008 997.4 0.8 47700 2500 52086 3742 2037 SWAP08 3 G10 YBT 17 0.01 0.0008 990 0.8 36960 650 41834 445 4002 SWAP08 4 H9 VDBS 26 23.9 0.0042 0.0008 995.8 0.8 44000 1600 47590 2074 3346 SWAP08 5 H9 RGCA 19 25.2 0.0065 0.0008 993.5 0.8 40500 1000 44162 973 2141 SWAP08 7 H9 RCA 15 24.8 0.0111 0.0008 988.9 0.8 36120 590 41225 574 2237 SWAP08 8 H9 RCA 15 22.2 0.0157 0.0008 984.3 0.8 33370 420 38257 1193 2248 SWEAP 08 9 G10 YB S 25.1 0.008 0.0007 992 0.7 38750 680 43121 692 3834 SWEAP 08 10 G9 RCA 23.6 0.0139 0.0007 986.1 0.7 34360 400 39706 851 3424 SWEAP 08 11 G9 RCA 21.1 0.0097 0.0007 990.3 0.7 37200 560 41956 419 2804 SWEAP 08 12 H9 RCA 25.5 0.0101 0.0007 989.9 0.7 36900 540 41806 398 7241 SWAP09 2 I10 YBT 21.9 0.0509 0.0006 949.1 0.6 23920 90 26831 383 7242 SWEAP 09 1 I10 YBT 26.7 0.0094 0.0005 990.6 0.5 37480 470 40158 396
121 CHAPTER 7 DESCRIPTIVE AND STAT ISTICAL ANALYSES OF THE T GROUP AND S GRO UP ASSEMBLAGES AT MOCHE BORAGO: RAW MATERIALS, CORES, AND LITHIC DEBITAGE Chapters 7 and 8 focus upon the analyses of flaked stone artifacts recovered from the early OIS 3 deposits of excavation unit G10 in the main Block Excavation Area (BXA). As previou sly discussed in Chapter 5, G10 provides a link to prior GEPCA excavations, as half of this 1m 2 unit formed the 1998 sondage that first exposed Moche My analytical sample comes from the remaining 50 cm 2 section of G10, excavated during the 2006 2008 field seasons. Although this unit represents only one of five 1m 2 excavation units into Late Pleistocene deposits so far excavated at Moche Borago, G10 is pe section, as every major Late Pleistocene litho stratigraphic unit is exposed in its profiles. 22 that form my analytical sample come from the T Group deposits, further subdivided into T Group Lower and T Group Upper, and the S Group dep osits stratified below the YBS tephra dated to ~ 43ka. Artifacts found in the later OIS 3 strata above YBS, the R Group, were not included in this dissertation (except for a small sample of points discussed in Chapter 8), as they were originally thought to be of OIS 2 (Terminal Pleistocene) age. 22 artifacts.
122 General Description of the Assemblage s This chapter follows a general lithic reduction sequence in that I begin with a discussion of hammerstone percusors followed by raw materials and nodules, cores and debitage The flaked stone artifact assemblage s from the T Group and S Group contain 2,913 total pieces (Table 7 1). The highest frequency of lithics are found in the T Group (n=2351), with T Group Lower ( n = 1,061 ) and T Group Upper (n = 1,290 ) each having simi lar lithic counts T he S Group (n = 555) has ~ 75 % fewer lithics than the T Group. Only five stone artifacts were found in the YBS tephra and two in the YBT tephra all of which were recovered near the boundaries of the tephra units and are probably intrus ive. Therefore, we consider the YBS and YBT tephras archaeologically sterile, and these artifacts have been removed from th is analysis. Minimum Number of Lithics Faunal specialists have long recognized that the actual minimum number of individual (MNI) specimens represented within a dataset may differ from the actual total number of bone fragments at hand (White, 1953a; White, 1953b) These differences accrue because multiple, non diagnostic bone fragments can originate from the same bone. L ithi c assemblage s also undergo taphonomic process es similar to faunal assemblage s, which can produce a fragmented assemblage On the one hand, the fragmentation might occur at the time the assemblage is accrued. The knapper may not have had enough skill to c ontrol the flaking process properly, which could have resulted in more fragmentary pieces. Raw material could also be a factor. Quartzite, for instance, often shatters during flaking. On the other hand, the post depositional process can also
123 fragment a stone artifact assemblage. If stone artifacts are exposed on the surface, then an increased likelihood occurs that these pieces might be walked upon by humans and/or animals, swept up and re deposited by the occupants, or affected by fluvial or aeolian pr ocesses. In any one of these situations lithic fragments such as medial and lateral pieces, can be derived from a single artifact if it is broken Therefore, such fragments can artificially drive up the total counts if each fragment is counted independ ently a single fractured flake shaped tool, or unshaped tool c an produce a proximal and dorsal fragment, as well as multiple medial and lateral fragments T he t otal lithic counts may therefore exaggerate (as in the case of shaped and unshaped tools and c ores) or underestimate (as in the case of whole flakes) the true character of an assemblage. Adapting the faunal based MNI approach to lithic analysis may be one way to more accurately count and analyze a lithic assemblage. Consequently, I introduce for Minimum Number of Lithics (MNL) which refers only to w hole stone artifacts, conjoined artifacts and proximal flake fragments Medial, distal, and lateral flake fragments and angular waste are therefore omitted from total lithic c ounts when using MNL Therefore, w hen only the MNL is used, the stone artifact count lowers to 1,715 total pieces (Table 7 2 ). The disparities in the total counts between stratigraphic groups are also equalized. The S Group has 532 lithics, the T Gro up Upper has 650 lithics, and the T Group Lower has 526 lithics which has important implications for understanding various artifact classes that are discussed in this chapter and the next chapter Hammerstones and Raw Material Nodules To date, o nly o ne possible hammerstone fragment has been found in G10. This artifact (Bag Number (BN) 3115) (Figure 7 1) came from Level 23 (stratigraphic unit
124 DCC of the T Group Upper) and it is made of basalt. A cavity within the nodule may be the reason that the sto ne broke in half Table 7 3 provides the dimensions and weights for this possible artifact and the lithic raw material nodules that are described in the next section Only 15 possible raw material nodules came from G10, which are nearly, or fully, corti cal. The small number of possible raw material nodules constitutes less than 1% of the cumulative (MNL) S Group and T Group lithic assemblage. However, it is doubtful ost of these artifac ts appear to be heavily aerated obsidian that have numerous cracks and irregularities which would make these pieces unsuitable for knapping Only one piece, a chalcedony nodule (BN 3105, Level 23) is not obsidian. Furthermore, the y are small (Table 7 3 ) and the average weight is only ~1.22 grams each (SD = 1.48 g). Therefore, these pieces could have been collected with other raw materials, but deemed unsatisfactory and discarded with no intention of ever being used (Leakey, 1971) natural stone pieces that were not collected for raw material but for some other purpose Raw Materials F ive broad classes of raw materials were found in the G10 assemblage: obsidian, chert, quartzite, roof spall 23 and basalt. The most common raw material is obsidian, especially a homogenous jet black variety referred to here as ommon B lack O bsidian (CBO). Two sources of obsidian are found wit hin 20 km of the site S. Brandt and I visited these sites in 2007 and 2010. P reliminary results from an XRF analysis of the 23 Naturally spalled rock from the rockshelter walls.
125 obsidian from these sites 24 show strong similarities to CBO at Moche Borago. The elemental spectra of the CBO show characteristic s of very low strontium, with higher counts of other rare earth elements, elevated zinc, and high counts of both iron and zirconium (Figure 7 2 ). In lieu of a complete chemical analysis of all obsidian found at Moche Borago, I use CBO here to refer only t o similarities in the external color (jet black) and appearance (homogenous) of the obsidian pieces Obsidian source work is ongoing. W ithin the G10 S Group and T Group assemblage 90.5% o f the obsidian found is CBO (n = 1,561) Gray obsidian represents 5 .7% (n = 98) of the S Group and T Group assemblage s and both green and brown obsidian varieties constitute just over 0.5% combined (n = 10). Twenty artifacts were made from roof spall and 19 artifacts were made of basalt which amounts to 2.3% of the ent ire dataset Sixteen artifacts were made from m ulti colored varieties of chert and chalcedony which represent 0.9% of the entire assemblage. O nly a single artifact was made of quartzite ( core trimming flake BN 179 4 .21, Level 11 from the S Group deposits ). Obsidian artifacts occur in all occupation layers within the deposits but the non obsidian raw materials have a much more limited distribution. Table 7 4 summarizes the MNL distribution of each raw material class within the major stratigraphic grou ps. Table 7 5 also provides MNL based raw counts of each raw material type per excavation level. T hese tables show that the entire assemblage of stone artifacts made from basalt and roof spall occur within the combined T Group Lower and T Group Upper de posits (n = 19 and n = 20, respectively). A similar distribution exists for the category of chert and 24 This analysis was conducted in the spring of 2010 using a Tracer III V porta ble XRF. Final results of this analysis are pending.
126 93.0% of the chert assemblage is also found in the T Group (n = 14). In fact, the only raw materials not represented in the T Group Lower are quartzite and chalcedony whereas multiple raw material types are absent from the T Group Upper (Table 7 6 ). Several explanations might account for the greater diversity of non CBO artifacts in the T Group Lower. The first, and perhaps the most obvious, explana tion is that these patterns are simply the result of sampling bias. This assemblage is taken only from a single 50 cm 2 excavation unit. In adjacent areas at Moche Borago the distribution of raw material types may be entirely different owing to spatiall y distinct behavior patterns of the people who previously occupied this site. Further excavation and analysis should help to clarify these issues. A second explanation for the greater diversity of non CBO artifacts in the T Group Lower is that these ra w material patterns are evidence ( although from a relatively limited sample size) for behavioral differences of the occupants of the rockshelter between the S Group and T Group. S pecific raw materials may have been preferentially used for certain tool typ es. On the one hand, the earliest microlithic tools in East Africa are hypothesized to be related to an uptake in fine grained raw material use (Ambrose, 2 002) This pattern is even seen in Ethiopia at Porc Epic rockshelter At Porc Epic, o bsidian constitutes only 8.0% of the total assemblage yet the majority of microliths, including 50% of all backed pieces, are made from this particular raw material. The nearest obsidian sources to Porc Epic are about 100 km from the site which shows a clear preference to make these tools with obsidian (Pleurdeau, 2001) On the other hand, little evidence is available from Moche Borago to suggest that non CBO was used preferentially to make shaped and unshaped tools or cores. T he
127 sample siz e of non CBO is very small and t he total count of minimal number of lithics (MNL) of non CBO lithics within each of the three main stratigraphic aggregates never exceeds 10 pieces (T Group Lower). Only 15 non CBO (MNL) lithics were found in the S Group, a nd nearly all these are debitage (Figure 7 3, 7 4, 7 5 ). The r elative percentages of non CBO per artifact class within litho stratigraphic groups are also small I n the case of the S Group the relative percentage of non CBO artifacts is never above 1% per lithic group (e.g. shaped tools, unshaped tools, and cores). Therefore, i f non CBO was preferred for any lithic activity then frequencies and percentages of non CBO within the assemblage should be greater 25 but this pattern is not seen in the current dataset. T he distribution of CBO and non CBO between shaped tools, unshaped tools, and cores is also similar in both the T Group Upper and Lower (r 2 = 0.84 and 0.95, respectively, p < .05). The same distribution pattern is true for debitage. Regressio n analysis shows that the percentages of debitage (including flake fragments) made from CBO or non CBO within each litho stratigraphic group are similar (r 2 = S Group: 0.98, T Group Upper: 0.92, and T Group Lower: 0.94, p < .05). I f non CBO w as being used differently than CBO, differences should be evident in the number of artifacts made from either group of raw materials What these distribution patterns suggest instead is that both raw material groups were being used for a variety of purposes, each with similar waste patterns. The broader implications of these findings will be discussed in more detail at the end of this chapter and in Chapter 9 I t is important to point out that the similarities 25 Unless spatially distinct activity areas occur at the site, which is highly likely.
128 between the use of CBO and non CBO raw materials likely refer to both social and technological reasons. On the one hand, if the development of microlithic tools during this period is linked to the availability and use of fine grained raw materials then the ready availability of CBO around Moche Borago may ha ve meant little technological need to acquire other kinds of non local, fine grained raw materials. On the other hand, because the majority of non CBO was found only on debitage (and still in small amounts) non CBO tools may have been curated more thorou ghly for non technological, perhaps social, reasons. In conclusion, four main points can be made about raw materials and raw material use in the G10 stone artifact assemblage. First, five different kinds of raw materials occur, but obsidian is far and aw ay the most common. Second, the most frequent kind of local source in the Sodo Wolayta area. Fourth, the stone artifacts from the T Group, especially T Group Lower, are mad e on more types of raw materials than the S Group, although this is still a very small percentage. However, no evidence shows that certain raw materials were preferentially used to make specific kinds of shaped and unshaped tools or cores. Cores The cor e sample is not very robust with only 94 pieces Single platform cores are most common in each litho stratigraphic unit. Multiple platform cores are next most common and there are less than ten radial cores, including 3 Levallois cores, in the assemblage s (Table 7 8). Nevertheless, the limited sample of cores provide s interesting insights into the lithic reduction strategies employed at the site during early OIS 3 T wo core characteristics stand out above all others : size and diversity
129 Core Size Sma ll cores are not uncommon from Late Pleistocene sites in eastern Africa Merrick (1975) found in Kenya and Tanzania w ere between only 3.4 cm and 1 .5 cm (see Figure 12:1 400). By comparison, the average length 26 of all cores at Moche Borago is 2.3 cm (SD = 0.9 cm). The largest cores at Moche Borago are flaked radially Radial cores include the L evallois flaking method, discoids, and oth er misc ellaneous radial core types. The mean volume for radial cores is 57.5 cm 3 (SD = 86.4) (Table 7 7 ) 27 Levallois radial cores, however are generally small, with an average volume of 5 cm 3 Single platform and multiple platform cores are also small Single platform cores have a mean volume of 7.4 cm 3 whereas multiple platform cores have a mean volume of 5.1 cm 3 28 S everal possible reasons arise why the cores in general, at Moche Borago are small including distance to raw material sources, size of original raw materials, and intensity of use. Distance to source materials is not expected to be a major factor in this area. M ajor obsidian sources are located within 20 km of Moche Borago and they can be reached within a single day. It is also lik ely that other obsidian sources existed in the 26 Dimensions of cores are notoriously difficult, hence the use of volumetric measures here. However, the 27 While still larger, these mean values for radial cores are skewed by a single large radial core in the T Group Lower. 28 These values are similar to a 1.7 cm equilateral cube. It is important to note here that the volumetric calculations suggest that all sides of the core are equilateral when in fact one or two sides m ay be significantly larger than the third side, depending on the type of core.
130 area due to volcanic activity in the region, but these sources are currently unknown to our research team. The size of the raw materials is also not a primary factor regulating the size of the Moche Borago c ores. In 2007 and 2010, S. Brandt and I observed that the local obsidian sources are broad deep veins ( often in excess of 20 30 cm wide ), located over a broad geographic area. Primary core reduction flakes and large cores are in abundance at these sites Therefore, it seems likely that raw materials were available locally to accommodate bigger cores, but the cores were probably reduced intentionally before they were transported to Moche Borago, or large nodules were broken into smaller ones at the quarr y (or less likely at the site given the small size of site nodules) It is also worth noting that the cores (and most other kinds of stone artifacts) most likely represent pieces that were discarded after being used, that is, the small sizes are an artifa ct of use and not inherent to the original artifact itself Many of the cores at Moche Borago have hinge or step fractures (for example, see Figure 7 6 ) which indicates that these pieces were used extensively, developed defects or became too small, and w ere discarded The average volume of radial cores decreases by 51% from the T Group Lower (M = 9.3 cm 3 SD = 3.7 cm 3 ) to the S Group (M = 4.5 cm 3 SD = 2.9 cm 3 ). Therefore, decrease in core volume may indicate a general diachronic shift toward smaller c ores and, by proxy, smaller shaped and unshaped tools and debitage. Diversity of Single and Multiple Platform Cores Merrick (1975) however, found little change in size between the late MSA and early LSA cores in Kenya and Tanzania. However Merrick ( 1975: 396 398) point ed out
131 other qualitative core characteristics of the East Africa early LSA (i.e. Mode 4/5 technology) that include four points : 1. infrequent radial (discoidal, L evallois, and radial) cores 2. prismatic and pyramidal core types 3. predominance (60 cores 4. predominance of blade flake removals on all core types The core characteristics that Merrick (1975) identified with the early LSA are similar to those characteristics in the Moche Borago assemblage. Here, I will compare the Moche findings. Single p latform c ores A shift in the design of cores emerges from the T Group to the S Group The change is toward an emphasis on the production of single platform cores in lieu of other core types. In the T Group Lower single platform cores are most common (58.8%, n = 10) but relatively high percentages of radial cores (17.7%, n = 3) and multiple platform cores (23.5%, n = 4) (Table 7 8 ) still occur In the T Group Upper the assemblage is composed almost uniformly of single platform cores (75%, n = 9) O nly one radial co re (8.3%) and two multiple platform cores in the T Group Upper (16.7%) materialize S ingle platform cores represent 58.5% (n = 38) of the S Group core types These cores are made from a variety of core blank types including flakes and cobbles. C oncurre nt increases in the frequency and diversity of multiple platform cores (17%, n = 11) also arise Only four radial cores have been found in the S Group representing 6.2%. Therefore, from the T Group to the S Group more single platform cores appear in the
132 assemblage and multiple platform cores also become more common. In contrast, radial cores maintain consistent low frequencies throughout the assemblage. Pyramidal and p rismatic c ores Merrick (1975) found tha t formal pyramidal and prismatic cores types are less common in the early LSA than less formal single and multiple platform cores A similar pattern exists at Moche Borago. O nly four pyramidal/prismatic cores in the entire Moche Borago assemblage were fou nd In contrast, 52 single platform cores within the total assemblage (Table 7 8) appeared. In addition to the radial cores, the small numbers of pyramidal and prismatic cores suggests that core production at Moche Borago followed the more relaxed and in formal designs common to single and multiple platform cores. Core f lake t ype More frequently, e arly LSA cores have blade flake removals, according to Merrick (1975) A similar pattern is again found at Moche Borago The incidence of single platform cores with non elongate flake removals is consistently high in the total assemblage ( 65% n = 34) However, w hen flake removal type is analyzed across litho stratigraphic groups a statistically significant invers e correlation is evident between the increase of blade flake removals and the decrease of ovate flake removals from the T Group Lower to the S Group (r = 0.979, sig(1 tailed) = .066) (Table 7 9 ). This inverse relationship would suggest a subtle shift fro m non elongate flake removals to more elongate flake removals from the T Group to the S Group. In summary, t he qualitative and quantitative characteristics of the Moche Borago (1975) definition of early LSA cores, which are seen here to represent Mode 4/5 core technology. The main characteristics of a Mode 4/5 core assemblage, as found at Moche Borago, include:
133 1. Decreasing amounts of radial cores. The percentage of radial c ores decreases across each litho stratigraphic unit at Moche Borago 2. Increased abundance of single platform cores. The S Group is composed of (non regularly patterned) more single platform cores ( 58 %). 3. Few formal cores designs, like radial, pyramidal, or prismatic cores. Pyramidal/prismatic cores constitute less than 10% in the entire Moche Borago core assemblage 4. More abundant blade flake removals from cores. The incidence of elongate flake scar removals on single and multiple platform cores increas e s in the S Group. Th ese findings suggest that the S Group assemblage, at least, is comparable to the early LSA or, more precisely, to Mode 4/5 technologies. In particular, increasing diversity is apparent in single and multiple platform core types from the T Group to the S Group. One explanation for this pattern is experimentation within the technological system at this time. The presence of L evallois and other radial core types throughout each of the three litho stratigraphic groups suggests that the knowledge base underlying technologies during this time accessed both the prior Mode 3 technology and the newer Mode 4/5 technology The d evelopment of new technological knowledge at this time therefore might include increased experimentation and non fixa tion on standardized core morphologies which may explain the high numbers of informal single and multiple platform cores One way to test this hypothesis would be to look toward the R Group, which overlies the S Group and note the most common cores and t he diversity of that assemblage. Debitage T wo broad categories of debitage are described here. The first category of debitage consists of functional debitage types such as core trimming flakes and burin spalls. T hese kinds of debitage often indicate sp ecialized reduction activities. An abundance of
134 core trimming flakes, for example, might suggest a great deal of in situ core preparation and reduction at a site. Burin spalls denote the manufacture of burin tools which may have been used for engraving/ grooving of wood, bone, or antler. O ther types of functional debitage include bifacial thinning flakes, but these are not represented within the current dataset. The second category of debitage is non functional. Non functional flakes include whole, prox imal, medial, distal, and lateral flake fragments, and angular waste. These lithics most often represent the condition of the assemblage, including whether or not the assemblage is biased due to post depositional processes or human behavior. For example, fluvial activity can bias an assemblage by preferentially removing smaller debitage from the assemblage whereas trampling can fracture whole flakes within an assemblage. The relationship between functional debitage types (i.e. core trimming flakes and bu rin spalls) and non functional debitage types (whole flakes and flake fragments) is one sided in the current dataset. Throughout each litho stratigraphic unit non functional debitage predominates The functional debitage never occurs higher than 10% wit hin any litho stratigraphic unit Also, no clear relationship exists between functional debitage and any other lithic type, such as cores or shaped tools. L inear regression shows no statistical relationship between the mean size of core trimming flakes i n each excavation level (overall, the mean length of CTF is ~3.0 cm, SD = 1.24 cm) compared with mean volumetric dimensions of the cores found in the same levels (r 2 = 0.006, p < .05). This result indicates that the core trimming flakes were removed when the cores were larger and not at the end of their use life cycl e when the cores were smaller and discarded
135 B oth types of debitage (including non MNL fragments) represent 83.4% (n = 2,423) of the combined S Group and T Group assemblage, making debitage t he predominant lithic category. The T Group Lower had 1,003 pieces of debitage (including non MNL) amounting to ~94.5% of that assemblage (Table 7 1 0 ). T he T Group Upper had a slightly higher total count of debitage (n = 1,212, non MNL) but the relativ e percentage of debitage is similar to the T Group Lower (~94%) owing to more shaped and unshaped tools. O nly 208 pieces of debitage were found in the S Group, however, and this number amounted to only 37.5% of that assemblage. N early three fourths of t he S Group MNL debitage is comprised of whole flakes (74%, n = 154). This peculiar debitage pattern suggests taphonomic, activity related or even technological reasons which may have influenced this assemblage. Taphonomic explanations, such as the fluv ial actions known in the S Group, seem unlikely. F luvial actions are mass oriented, hence smaller shaped and unshaped tools, debitage, and cores alike would all be removed. N umerous smaller shaped and unshaped tools and cores in the S Group and the volum es of these pieces are on average smaller per excavation level in the S Group than in the T Group (Table 7 1 1 ). Furthermore, few of the artifacts are abraded in the S Group (and T Group) These results suggest that the S Group debitage is not only in si tu ( a finding corroborated by the micromorphology and other litho stratigraphic research ), but no size sorting of the assemblage has account ed for the predominan ce of whole flakes in the S Group. Comparisons between the number of fragmentary debitage fou nd in the T Group and S Group suggest as much about the process affecting the T Group as they do about the S Group. One interpretation for the greater number of fragmentary debitage in the T
136 Group may be due to these pieces being exposed on the surface lo nger and to the effects of foot traffic and trampling. T he T Group deposits are only slightly thicker in G10 compared to the S Group deposits Yet, the T Group may span as much as 10,000 years (i.e. from ~60 ka ~45 ka) whereas the S Group deposits wer e deposited within ~ 3,000 years (from ~41 ka ~44 ka if the minimum and maximum age errors are considered ). Therefore, more fragmentary debitage in the T Group was found because these stone artifacts were exposed on the surface longer. B ehavioral chang es may also explain the differences between the T Group and S Group debitage patterns if different activities were conducted at the site or if the way certain activities were conducted Raw material changes were not a likely factor, however, b ecause only 15 non CBO pieces were found within the S Group MNL assemblage which include d multiple lithic types. The relatively sparse non CBO pieces, as well as the diverse array of lithic types that these pieces represent, suggest no specialized raw material use i n the S Group that might indicate coincident changes in the way tools were being made or the way certain kinds of materials were used What then is the cause for the peculiar debitage patterns between the T Group and S Group ? This period (early to mid O IS 3) is associated with the development and systematic adoption of Mode 4/5 microlithic and blade based technologies across Africa. We therefore should expect to find changes in flake size and elongation that might relate to the development of Mode 4/5 s tone tool technologies Mode 4/5 technology seem s to be well represented in each litho stratigraphic group at Moche Borago via diagnostic shaped tools such as microliths (see Chapter 8 ) and characteristics of the cores The debitage also appears bladey overall and small. T he
137 average length of all whole flakes is 1.6 cm ( SD = 0.9 cm) and a subtle decrease occurs in the mean lengths of whole flakes across each litho stratigraphic group (Figure 7 7 ). However, t he mean aspect ratios (length/width) of the whole flakes, which indicate elongation, remain largely unchanged across each excavation level (Figure 7 8 ). Furthermore, the weighted mean of the average aspect ratios per excavation level is only 1.24 (SE = 0.13), which would suggest a much more ovate w hole flake planform and not an elongate, bladey planform The unchanging aspect ratios belie the bladey appearance of the assemblage, as well as the subtle changes in whole flake length. Therefore, little actual information may be gleaned from the Moche Borago whole flake data about Mode 4/5 technology in the T Group and S Group if the trend is toward smaller and more elongate flakes. Perhaps because of the commonplace fine grained raw materials (CBO), the stone artifacts were already being made small an d more elongate with little reason to change at this time. Furthermore, looking at flake length or aspect ratio might define the problem too narrowly. Other likely behavioral changes evident in the stone artifact assemblage might relate to the predomina nce of debitage in the T Group and whole flakes in the S Group. For example, while proportionally more debitage emerges in the T Group (~94%, n = 1,040) more shaped tools, unshaped tools, and cores in the S Group occur (~65%, n = 347) (Figure 7 9 ). In f act, 20 8 % more shaped and unshaped tools materialize in the S Group than in the T Group (Figure 7 10 and Table 7 1 2 ) 29 In the S Group, 60% (n = 128) of the shaped and unshaped tools are also made on end struck whole flakes. Less than 5% of these pieces a re made on either proximal, medial, or distal flake fragments 29 Microliths increase 650%, burins 75%, points 111%, scrapers 235%, and unshaped tools 331% (see C hapter 8 ).
138 (c ombined, these categories amount to 13.7% ) (Table 7 13). Therefore, when viewed within a broader context, the debitage patterns in the S Group might relate to the manufacturing process of sha ped and unshaped tools. Furthermore, a low but statistically significant relationship exists between whole flake striking platform type and aspect ratios (r = .156, p (one tailed) < .05). Flakes with pointed striking platforms also have the most elong ate dimensions, suggesting that the production of whole flakes in the S Group was systematic with a set standard of manufacture. However, the average lengths of whole flakes in the S Group are 50% smaller than the average length s of shaped and unshaped t ools (M = 2.4 cm, SD = 0.85 cm) (Figure 7 5). The average aspect ratios of whole flakes (M = 1.4, SD = 0.8) are also less than shaped and unshaped tools (M = 1.6, SD = 0.6) indicating the flakes are slightly more ovate than the shaped and unshaped tools. T he discrepanc ies between the sizes and shapes of the whole flakes and shaped and unshaped tools may imply no relationship between these lithic types Alternatively, t he discrepancy may suggest that the whole unmodified flakes were rejected for various reasons, perhaps for being too small. I find it altogether unlikely that there is no relationship between the predominance of whole flakes in the S Group and the concurrent increase in shaped and unshaped tools. However, based on the current dataset onl y circumstantial morphological and metrical data connect these lithic types. T he predominance of whole flakes suggests an intensified focus on the production of these lithics in the S Group Also, the predominance of whole flakes does not represent a bias ed sample due to actions such as fluvial processes. The frequent use of prepared, especially pointed, striking platforms also shows that the manufacture of these flakes was predetermined. Although the sizes and dimensions of
139 the whole flakes do not match the shaped and unshaped tools these flakes may have been the rejects of the manufacturing process because they were, in fact, too small or not elongate enough at the outset Conclusion s The Moche Borago core and debitage samples show diachronic trends from the T Group Lower to the S Group including the increased diversity of core types in the S Group and a general trend toward the production of more elongate flakes. Raw material not surprisingly, does not seem to be a big factor. CBO predominates ac ross all levels O nly in the lowest litho stratigraphic group (T Group Lower) do more kinds and greater numbers of non CBO raw materials occur, albeit in still very small percentages The diachronic technological patterns seen within the Moche Borago li thic assemblage also appears to parallel patterns which are found in other assemblage across East Africa and the Horn of Africa (i.e. Porc Epic and Kenyan l ate MSA / early LSA sites) (Merrick, 1975; Pleurdeau, 2001) Th ese patterns include increas ing percentages of non patterned single platform cores and decreasing percentages of radial core s. T he use of Mode 4/5 technologies at Moche Borago also does not appear to coincide with a shift in raw material use, which is seen in other sites in the region. The highest diversity and frequency of non CBO raw materials are found within the oldest l itho stratigraphic group at the site Perhaps hunter gatherers living within and around Moche Borago at this time just did not need to expend effort procuring fine grained raw materials from distant sources because they had abundant obsidian nearby. If th is hypothesis is substantiated, then it may provide operational limitations on the use of raw material based models for the development of Mode 4/5 microlith technologies at Moche Borago and the surrounding region
140 Table 7 1: Summary of the major lithi c groups by excavation level and stratigraphic group. Stratigraphic Group Level Shaped Tools Unshaped Tools Cores Debitage Nodules Total 11 33 23 13 6 1 76 12 14 16 9 6 1 46 13 63 53 22 33 0 171 S Group 14 37 28 11 59 0 135 15 2 1 7 35 0 4 5 16 5 5 3 69 0 82 Total 154 126 65 208 2 555 18 7 1 0 222 5 235 19 7 2 1 80 0 90 20 0 0 0 8 0 8 T Group Upper 21 1 1 2 162 2 168 22 7 2 2 227 1 239 23 10 6 4 350 7 377 24 5 3 2 163 0 173 Total 37 15 11 1212 15 1290 25 5 2 1 178 1 187 26 0 0 1 58 0 59 27 3 4 1 188 1 197 T Group Lower 28 3 0 5 218 0 226 29 2 3 2 74 0 81 30 10 5 6 273 0 294 32 2 0 1 14 0 17 Total 25 14 17 1003 2 1061 Total Assemblage = 2,906 Notes: This table includes total lithic counts including medial and lateral flake debitage. The seven lithics found in YBS and YBT have been removed from the total count.
141 Table 7 2 : Summary of the major lithic groups by excavation level and stratigraphic group without medial or lateral flake debitage (MNL). Stratigraphic Group Level Shaped Tools Unshaped Tools Cores Debitage Nodules Total 11 33 23 13 3 1 73 12 14 16 9 5 1 45 S Group 13 63 53 22 14 0 152 14 37 28 11 59 0 135 15 2 1 7 35 0 45 16 5 5 3 69 0 82 Total 154 126 65 185 2 532 18 7 1 0 114 5 127 19 7 2 1 47 0 57 20 0 0 0 4 0 4 T Group Upper 21 1 1 2 77 2 83 22 7 2 2 110 1 122 23 10 6 4 153 7 180 24 5 3 2 67 0 77 Total 37 15 11 572 15 650 25 5 2 1 99 1 108 26 0 0 1 26 0 27 T Group Lower 27 3 4 1 86 1 95 28 3 0 5 106 0 114 29 2 3 2 41 0 48 30 10 5 6 104 0 125 32 2 0 1 6 0 9 Total 25 14 17 468 2 526 Total Assemblage = 1,715
142 Figure 7 1 : BN 3115.1, Level 23. Stra tigraphic Unit DCC (T Group Upper). This photo is a possible hammerstone fragment as evidenced by peripheral pock marking. E vidence also shows grinding on this piece.
143 Table 7 3 : Metric information for nodules and hammerstones. Bag Number Excava tion Unit Level Artifact Type Weight (g) Raw Material Length (mm) Breadth (mm) Thickness (mm) Volume (mm^3) 1796 G10 11 Nodule CBO 19.23 12.08 7.88 1830.511 1799.7 G10 12 Nodule CBO 2254 G10 18 Nodule 0.63 CBO 2571 G10 18 Nodule 0.93 CBO 2572 G10 18 Nodule 0.18 CBO 2129 G10 21 Nodule 1.29 CBO 19.21 13.97 7.37 1977.841 2129 G10 21 Nodule 5.58 CBO 29.8 30.11 10.87 9753.412 2133 G10 22 Nodule 0.52 CBO 23.07 9.8 3.36 759.649 3105 G10 23 Nodule 1.39 CBO 3105 G10 23 Nodule 0.95 Gre y Obsidian 3105 G10 23 Nodule 0.18 CBO 3105 G10 23 Nodule 0.19 CBO 3105 G10 23 Nodule 0.05 CBO 3730 G10 27 Nodule 0.18 CBO 12.77 8.1 2.85 294.7955 3105 G10 23 Nodule 3.88 Chalcedony 3115.1 G10 23 Hammerstone 122.82 Basalt 65.05 47 .16 36.03 110531.3
144 Figure 7 2 : XRF spectra of Common Black Obsidian (CBO). CBO has high elemental counts of Fe and Zr, and most rare earth elements including Rb, Y, and Nb. Sr is characteristically low.
145 Table 7 4: MNL Frequencies and percentag es of the various raw material groups in the T Groups and S Group. Stratigraphic Group Raw Material Group N % of Total Count Within Stratigraphic Group % of Total Count per Raw Material % of Total Assemblage Count S Group Obsidian 530 99.62% 31.87% 30.8 5% Chert 1 0.19% 6.67% 0.06% Basalt 0 0.00% 0.00% 0.00% Roof spall 0 0.00% 0.00% 0.00% Quartzitic 1 0.19% 100.00% 0.06% Stratigraphic Group Total 532 T Group Upper Obsidian 636 97.55% 38.24% 37.02% Chert 2 0.31% 13.33% 0.12% Basalt 13 1.99% 68.42% 0.76% Roof spall 1 0.15% 5.00% 0.06% Quartzitic 0 0.00% 0.00% 0.00% Stratigraphic Group Total 652 T Group Lower Obsidian 497 93.07% 29.89% 28.93% Chert 12 2.25% 80.00% 0.70% Basalt 6 1.12% 31.58% 0.35% Roof spall 19 3.56% 95.00% 1.11% Quartzitic 0 0.00% 0.00% 0.00% Stratigraphic Group Total 534 Total Assemblage Count 1718 Total Count per Raw Material Type Obsidian 1663 Chert 15 Basalt 19 Roof spall 20 Quartzitic 1
146 Table 7 5 : MNL Frequencies of raw material groups within the T Group and S Group deposits. Stratigraphic Group Level Quartzitic Basalt Roof spall Chert Obsidian Total Count per Level S Group 11 1 72 73 12 45 45 13 152 152 14 1 134 135 15 45 45 16 82 82 Total 1 1 530 T Group Upper 18 1 126 127 19 57 57 20 4 4 21 85 85 22 9 113 122 23 3 1 1 175 180 24 1 76 77 Total 13 1 2 636 T Group Lower 25 1 2 107 110 26 1 1 26 28 27 7 2 88 9 7 28 5 3 106 114 29 3 4 1 43 51 30 1 2 3 119 125 32 1 8 9 Total 6 19 12 497
147 Figure 7 3 : MNL proportions of CBO and non CBO raw materials in the S Group. The data are further subdivided by lithic group. Figure 7 4 : MNL proportions of CBO and non CBO raw materials in the T Group Upper. The data are further subdivided by lithic group.
148 Figure 7 5 : MNL proportions of CBO and non CBO raw materials in the T Group Lower. The data are further subdivided by lithic group Table 7 6 : MNL Frequencies of individual raw material types organized by stratigraphic aggregate. Raw Material Types S Group T Group Upper T Group Lower Total Quartzitic 1 1 Basalt 13 6 19 Roofspall 1 19 20 CBO 517 588 451 1556 Green Obsidian 2 2 4 Grey Obsidian 10 44 43 97 Brown Obsidian 1 4 1 6 Light Brown Chert 2 2 Dark Brown Chert 1 1 Grey Chert 1 5 6 Chalcedony 1 1 Red Chert 1 1 2 Pink Chert 2 2 White Chert 1 1 Total 532 652 5 34 1718
149 Table 7 7 : Frequency and volumetric measurements for core types and core groups from each of the three litho stratigraphic units discussed in the text (non MNL) Core type N Mean volume (cm^3) Std. Deviation Radial Cores Discoidal 2 22.3 14.8 Levallois 3 5.1 1.4 Misc. Radial 3 145 242.9 Mean 8 57.5 86.4 Single Platform Pyrimidal / Prismatic 4 6.1 5.7 Single Platform 52 8.6 21.4 Mean 56 7.4 13.5 Multiple Platform Biconical 2 7.6 2.4 Double Platform 2 3 3.2 Oppose d Platform, Same Face 5 5.6 1.9 Mult. Platform 1 8.7 Opposed Platform, Different Face 2 4.6 1.1 Double Platform at right Angles 3 5.3 1.3 Bipolar 1 2.4 Acute Angle 1 5.7 Two Platforms, Same Face at Right Angles 1 3.2 Mean 18 5.1 2 Misc Core Types Irregular 2 1.8 0.6 Fragment 9 2.2 1.9 Core Blank 1 4.3 Mean 12 2.8 1.3
150 Figure 7 6 : Selected single and multi platform cores from the G10 assemblage Arrows denote step or hinge fractures.
151 Table 7 8 : Frequencies of core types in the S Group and T Group Upper and Lower (non MNL)
152 Table 7 9 : Relative percents of core flake removal types per excavation levels (non MNL) Level Flake Blade Levallois Multiple Indet. Total S Group 11 50.00% 33.30% 0.00% 16.70% 0.00% 12 12 44.40% 44.40% 11.10% 0.00% 0.00% 9 13 63.60% 27.30% 0.00% 0.00% 9.10% 22 14 36.40% 63.60% 0.00% 0.00% 0.00% 11 15 28.60% 42.90% 0.00% 14.30% 14.30% 7 16 33.30% 66.70% 0.00% 0.00% 0.00% 3 48.40% 40.60% 1.60% 4.70% 4.70% 64 T Group Upper 19 100.00% 0.00% 0.00% 0.00% 0.00% 1 21 100.00% 0.00% 0.00% 0.00% 0.00% 2 22 50.00% 50.00% 0.00% 0.00% 0.00% 2 23 75.00% 25.00% 0.00% 0.00% 0.00% 4 24 100.00% 0.00% 0.00% 0.00% 0.00% 2 81.80% 18.20% 0.00% 0.00% 0.00% 11 T Group Lower 25 100.00% 0.00% 0.00% 0.00% 0.00% 1 26 100.00% 0.00% 0.00% 0.00% 0.00% 1 27 0.00% 100.00% 0.00% 0.00% 0.00% 1 28 100.00% 0.00% 0.00% 0.00% 0.00% 5 29 100.00% 0.00% 0.00% 0.00% 0.00% 2 30 83.30% 0.00% 16.70% 0.00% 0.00% 6 32 100.00% 0.00% 0.00% 0.00% 0.00% 1 88.20% 5.90% 5.90% 0.00% 0.00% 17 Notes: A statistically significant inverse correlation exists between the increase of blade flake removals from the T Group Lower to the S Group and the decrease in flake removals (r = 0.979, sig (one tailed) = .066).
153 Table 7 10 : P ercentages of the major lithic groups by stratigraphic aggregates for the entire assemblage (upper) and MNL only (lower). Frequencies and Relative Percentages of Lithics based on the Full Assemblage Shaped Tools U nshaped Tools Cores Debitage Nodules Total S Group 154 126 65 208 2 555 27.75% 22.70% 11.71% 37.48% 0.36% T Group Upper 37 15 11 1212 15 1290 2.87% 1.16% 0.85% 93.95% 1.16% T Group Lower 25 14 17 1003 2 1061 2.36% 1.32% 1.60% 94.53% 0.19% Total 216 155 94 2424 19 2908 Frequencies and Relative Percentages of Lithics based on the MNL Assemblage Shaped Tools Unshaped Tools Cores Debitage Nodules Total S Group 154 126 65 185 2 532 28.95% 23.68% 12.22% 34 .77% 0.38% T Group Upper 37 15 11 572 15 650 5.69% 2.31% 1.69% 88.00% 2.31% T Group Lower 25 14 17 468 2 526 4.75% 2.66% 3.23% 88.97% 0.38% Total 216 155 93 1225 19 1708
154 Table 7 1 1 : Average volume (mm 3 ) of cores, shaped and uns haped tools per excavation level and litho stratigraphic group (non MNL) Level N Mean Volume (mm^2) Std. Deviation Group Mean Volume (mm^2) Group Std. Deviation S Group 10 5 4751.06 3124.08 2802.05 2937.27 11 69 3679.08 4385.7 12 39 2730.53 203 4.75 13 138 2566.17 2582.75 14 76 2239.93 2040.49 15 10 2814.58 1847.16 16 13 4142.32 3384.49 T Group Upper 17 1 14007.45 8934.95 22388.49 18 7 1434.98 1309.89 19 8 4090.18 1603.54 21 3 13262.59 16901.61 22 9 9547.99 6567 .7 23 14 5361.14 5697.38 24 7 26476.56 56594.97 T Group Lower 25 6 5514.21 4679.17 14771.93 64984.95 26 1 9689.42 27 4 3875.99 1979.03 28 8 3890.43 2671.18 29 4 8192.59 2806.29 30 16 4223.84 2254.58 32 3 143555.7 244133.5 9 Total 441 4657.08 21645.93 Notes: The calculation of the group mean volume was made using the raw data values derived from the level means.
155 Figure 7 7 : Bar chart showing the mean length of whole flakes per excavation level (MNL) Standa rd deviation and a linear trend line have been superposed for reference. This graph shows a subtle, but evident, decrease in the mean length of whole flake from the T Group Lower (levels 32 25) and the S Group (levels 16 11). Level 11 and 12 are absent f rom this chart due to a lack of whole flakes in these levels.
156 Figure 7 8 : Bar chart showing the mean aspect ratio (length/width) of whole flake per excavation level (MNL) This graphs a slight decrease in aspect ratio from the T Group Lower to the S Group, but the trend is actually almost indiscernible. For comparison, an aspect ratio value of 1 is equivalent to a perfect circle and value of 2 is twice as long as it is wide.
157 Figure 7 9 : MNL frequency of raw material groups from excavation l evel 11 to 32. The patterns of lithic types suggest a turnover from debitage dominance in the T Group (level 32 18) to shaped and unshaped tool dominance in the S Group (level 16 11).
158 Figure 7 10 : Relative proportions of lithic groups in the S Gr oup and T Group stratigraphic aggregates.
159 Table 7 12: Total frequencies and percentages of various shaped and unshaped tool categories (non MNL). Stratigraphic Aggregate Misc Unshaped Burins Microliths Points / Unifaces / Bifaces Scrapers Utilized Mo dified Total S Group Frequency 14 7 15 72 47 61 64 280 Percent 5.00% 2.50% 5.36% 25.71% 16.79% 21.79% 22.86% 0.00% T Group Upper Frequency 5 3 2 20 7 9 6 52 Percent 9.62% 5.77% 3.85% 38.46% 13.46% 17.31% 11.54% 0.00% T Group Lower Frequency 3 1 0 14 7 7 7 39 Percent 7.69% 2.56% 0.00% 35.90% 17.95% 17.95% 17.95% 0.00% Total 22 11 17 106 61 77 77 371 Table 7 1 3 : MNL proportions of the blank types used to make shaped and unshaped tools within each litho stratigraphic unit. E nd Struck Flake Side struck flake Flake Fragments Other Total S Group 60.38% 3.30% 13.68% 22.64% 212 T Group Lower 68.75% 3.13% 9.38% 18.75% 32 T Group Upper 59.57% 6.38% 10.64% 23.40% 47
160 CHAPTER 8 DESCRIPTIVE AND STAT ISTICAL ANALYSIS OF THE T GROUP AND S GROUP ASSEMBLAGES AT MOCHE BORAGO: UN SHAPED AND SHAPED TOOLS As discussed in Ch 7, lithic raw materials, cores, and debitage from the T Group and S Group assemblages suggest diachronoic changes in the frequencies of artifact types and core redu ction strategies. This chapter considers the evidence for temporal changes in unshaped and shaped tools from the T Group and S Group assemblages. The chapter first focuses on the description and analysis of the unshaped tools, followed by the five broad classes of shaped tools: 1) microliths; 2) miscellaneous shaped tools such as becs, awls, and drills; 3) scrapers; 4) burins; and 5) points (unifacial, parti bifacial and bifacial). Overview of the Unshaped and Shaped Tools Unshaped tools (also referred to as informal tools) are flaked stone artifacts that display shallow or irregular retouch along one or more edges without intentional alteration of the original margin alteration is more invasive but the shape of the original un worked edge is still maintained. that alters the original shape of the edge(s). The edge alteration on shaped tools is often semi invasive (5 10 mm) to fully invasive (> 10 mm), regularly spaced, and patterned. Shaped tools such as drills, awls, scrapers, or points are made to suite spe cific functional and/or social tasks. A total of 369 unshaped and shaped whole and broken tools were recovered from the S Group and T Group assemblages. Table 8 1 provides the frequency and percentages of shaped and unshaped tools relative to the total n umber of artifacts in this class. Among all stone artifacts, the highest frequencies of unshaped and shaped tools are found in the S Group (9.6%, n = 279) while T Group tools account for 3.1 % (n = 90). In the S Group alone, unshaped and
161 shaped tools acco unt for 50.5% of the assemblage whereas in the T Group these tools only account for 3.9%. The higher percentage of unshaped and shaped tools in the S Group parallels the percentage of cores from the S Group. Table 8 2 shows the relative percentages of s haped and unshaped tools per stratigraphic group, which indicated that there was a fairly consistent distribution of scrapers, points, and unshaped tools within each litho stratigraphic group. These findings suggest that although the total number of unsha ped and shaped tools increases in the S Group, the relative proportions remain largely unchanged between the two groups. Microliths are the only tool class not found within all three litho stratigraphic groups. When a Minimum Number of Lithics (MNL) fi lter (see Chapter 7) is applied to the dataset, the total amount of unshaped and shaped tools drops to 201, which is 54% lower than the non MNL counts (Table 8 3). More unshaped and shaped tools are still found in the S Group, but the proportions of diffe rent tool types change. For example, the frequency of points in the S Group drops from 72 to 29, representing a 59.7% decrease. In comparison, the frequency of points in the T Group Upper decreases by 80% and by 50% in the T Group Lower. Unshaped Tools The frequency and relative percentage of utilized and modified unshaped tools are provided in Table 8 4. MNL counts show that unshaped tools represent 30 to 40% of all tools in the T Group and 35% of all tools in the S Group. The high percentage of uns haped tools suggest that during early OIS 3 the occupants of Moche Borago during the T Group and S Group may have engaged in similar, and perhaps opportunistic, tasks that required relatively quick and informal tools. Unifacial dorsal edge damage is fou nd on 72% of all unshaped tools (Table 8 5). Inverse (ventral) edge damage or retouch accounts only for 12% of all unshaped tools. Retouch and
162 edge damage occurs primarily on one or both lateral edges, but less commonly on the distal edge and almost neve r on the proximal edge. Incidences of step fracturing are also low, and 57% of dorsal flake scars are neither stepped nor hinged (Table 8 6). The low amount of step and hinge flake fractures suggest that the unshaped tools, as a whole, may have been used only briefly and moderately before they were discarded. Modified and utilized unshaped tools were made on similar types of flake blanks. End struck whole flake blanks account for 71% of modified tools and 84% of utilized unshaped tools (Table 8 7). F lake fragments are the next most common blank type, while side struck flake blanks represent only 5% and 1% of the modified and utilized unshaped tools, respectively. The size and elongation of unshaped tools vary between modified and utilized tools and also between the T Group and the S Group. Utilized tools are longer than modified tools by 13%. Furthermore, the longest unshaped tools are found in the T Group Upper compared to the S Group (Table 8 8). However, the aspect ratios of unshaped tools, whi ch is illustrated in Figure 8 1, shows a clear trend towards elongation from the T Group Lower to the S Group. Miscellaneous Shaped Tools There are only ten miscellaneous shaped stone artifacts like becs, awls, and outil ecaille. Table 8 9 provides basic metric and descriptive information about these artifacts. All of these artifacts are made on CBO and all but one piece is found in the S Group, which is BN 3105.11, an irregularly shaped artifact from the T Group Upper. Only a single possible drill like implement has been identified (BN 1794.15) due to dihedral lateral retouching. There are five possible awls, one bec, and two possible outil ecaille. Since most of these stone artifacts come from the S Group it is possible that they indicate more diverse kinds of on site activities at that time than during the T Group.
163 Microliths Microliths first appear in the T Group Upper. The S Group and T Group deposits contained 17 whole and fragmented microliths, representing ~5% of the total unshaped and shaped to ols assemblage. Additional microliths have been recovered from the S Group and T Group deposits in units H9 and G9, but they have yet to be analyzed. The sample size is limited, but we can make a few diachronic observations. Table 8 10 provides basic me tric information for each microlith recovered from G10. Photos and illustrations for most of the microliths are provided in Figure 8 2, and each piece is described in detail in Appendix C. All the microliths in the current sample are manufactured using c ommon black obsidian (CBO). They are all elongate and most are straight backed with obverse backing. The average microlith aspect ratio is 2.1 (SD = 1.0), which shows that these artifacts are twice as long as they are wide. Straight backing accounts for 29.4% (n = 5) of the total microlith assemblage, and many of these pieces are obversely backed (41.2%, n = 7). Bidirectional flaking is the next most common type of backing (23.5%, n = 4). 30 The backing angle is also relatively steep on each piece, whic h would be expected for backed tools. The average backing angle is 81.3 (SD = 13.3). Also, no statistically significant relationship occurs between the backing angle and backing type (r = .202, sig(1 tailed) = .454, p < .01), but obversely backed micro liths, in general, have the steepest edge angles (84.6, SD = 11.44). Finally, no statistically significant relationship occurs between the microlith type and type of backing ( r = .202, sig(1 tailed) = .227, p < .01). 30 Bidirectional flaking differs from alternating flaking pattern s because bidirectional flaking is less standardized. But an alternating flaking pattern is a series of inverse flake scars followed by a series of obverse flake scars in a repetitive fashion along the backed edge of the tool.
164 The current sample is too small to reveal any statistically significant quantitative patterning in the microliths from the T Group to the S Group. However, two pieces of qualitative evidence suggest that microlithic technology at this time may have been in a state of flux. The first piece is artifact BN 2571.1 from Level 18. While most microliths are straight backed, BN 2571.1 is a crescent, which shows that Mode 4/5 geometric microlith designs had been developed by ~45 ka. Geometric microliths become common in Ethiopia in the Holocene M ode 4/5 technologies (Brandt, 1982) and BN 2571.1 is one of the earliest Mode 4/5 crescents known in the area. Furthermore, in overlying S Group deposits, another microlith is made on a Levallois flake (BN 1818.6, Level 16), which is s ignificant because the artifacts shows that prior Mode 3 Levallois flaking techniques were used in combination with Mode 4/5 tool design. Scrapers All the scrapers are made from CBO. More than three quarters of all scrapers or scraper fragments are foun d in the S Group (77%, n = 47), compared to just seven each in the T Group Upper and Lower deposits (Tables 8 11 and 8 12). Several scrapers were fragmented, and the dataset has been filtered for MNL (Table 8 13 and 8 14). Metrical data for all scrapers is provided in Table 8 15. End scrapers are the most common scraper type before and after MNL. The most end scrapers are found in the S Group representing 41.4% (n = 17) of that assemblage whereas there are two scrapers in the T Group Upper (40%) and onl y a single end scraper in the T Group Lower (25%) (Tables 8 13 and 8 14). Notched scrapers are the next most common type, but are significantly reduced in frequency, representing 22% (n = 9) of the S Group and 20% (n = 1) of the T Group Upper. No notched scrapers were present in the T Group Lower. Side scrapers are also poorly represented, accounting for 12% (n = 5) of the S Group, 0% of the T Group Upper,
165 and 25% (n = 1) of the T Group Lower. Other categories of scrapers, such as scraper side combinati ons, represent a combined 28.0% (n = 14) of the total assemblage. The scrapers are all fairly elongate with aspect ratios exceeding 1.0 (Table 8 16), but there are minimal diachronic differences in elongation. The most elongate scrapers are found in th e T Group Upper (M = 1.7, SD = 0.6), and the least elongate scrapers are found in the T Group Lower (M = 1.3, SD = 0.2). End scrapers are also ~10% less elongate than the mean elongation values calculated from the entire assemblage (Table 8 17). Most s crapers are retouched on the dorsal face (Table 8 18). End scrapers are retouched most commonly along the distal edge (52.6%), as are scrapers with multiple retouched edges (e.g. end and side, end and double side). Side scrapers are retouched with equal frequency along either the left lateral or both laterals (33.3%) (Table 8 19). The working edge angles for side and end scrapers are all fairly steep with an average of 73.9 (SD = 4.6). Edge angles appear lowest in the T Group Lower (68.5, SD = 0.7) a nd highest in the S Group (77.2 (SD = 2.7) (Table 8 20). However, these observations are made from only four artifacts in the T Group Lower. Only two end scrapers are in the T Group Upper. Straight working edges are also most common among end scrapers (45%, n = 9) and convex working edges are next most common (30%, n = 6) (Table 8 21). Serrated edges or irregular edges, such as burs or use wear defects, occur in relatively low frequency within the entire scraper assemblage, which suggests that these a rtifacts were not used intensively before being discarded. Of the nine notched pieces, all but one are found in the S Group, and of those pieces, 55% (n = 6) come from Level 13. Notching is most common along the right lateral edges of flakes (66.7%). T he notches are also typically obversely flaked (75%). Table 8 22 provides basic
166 metrics on the notched pieces. The average notch width is 10.5 mm (SD = 2.5), and the average notch depth is 3.8 mm (SD = 0.8). Burins Only 12 burins were identified, and ba sic descriptive information for each piece is provided in Table 8 23. Two noteworthy specimens are illustrated in Figures 8 3 and 8 4. Burins occur in each litho stratigraphic group, but they are most common (67%, n = 8) in the S Group. All the burins ar e made of common black obsidian (CBO) except for BN 1818.12, which is made of gray obsidian. Of these burins, 83% (n = 10) are made on end struck flakes. The burins are also often elongate with a mean aspect ratio of 2.2 (SD = 0.9). However, T Group bur ins appear slightly more elongate (M=2.6, SD = 1.2) than S Group burins (M=2.0, SD = 0.6). Angle burins (42%, n = 5) and dihedral burins (33%, n = 4) are the most common types. Dihedral and angle burins are found in both the T Group and S Group. The aver age width of the burin edges is 6.2 mm (SD = 1.8) and angle burins have a more oblique edge angle (M=86, SD = 9.3) than dihedral burins (M=54,SD = 2.9). Two of the burins are especially noteworthy with detailed descriptions as follows: BN 1802.98 BN 1802.98 is a backed flake but with a series of burin spalls on the distal left lateral edge. The tool is made on an end struck whole flake. The backing is found along the right lateral edge. The distal edge has also been snapped off, likely to create an oblique striking platform for the burin spalls. Two spalls were removed along the right lateral edge to create an angle burin at the intersection of the right lateral and distal faces. It appears that the burins spalls overlie the backing retouch. Le vel 13 G10 S Group Figure 8 3 BN 4906.2 BN 4906.2 is a dihedral burin. The tool is made on the distal section of an end struck flake. It appears that the flake blank was snapped (lateral lateral), and a single burin spall was r emoved from the left lateral edge. This spall was then truncated with two additional obverse flake removals along the proximal left lateral edge. The obverse flake removals may have been to reinforce this face as a striking platform. Subsequently, three or four more burin Level 30 G10 T Group Lower Figure 8 4
167 spalls were removed from the proximal right lateral edge using the former burin surface as the striking platform. The angle of the resulting dihedral burin is 52. Poi nts Points are a major component of most Late Pleistocene archaeological assemblages in East Africa and the Horn. Numerous studies have described the diverse inter regional and intra regional variations of points (Cole, 1954; Clark, 1954; Clark, 1970; Kurashina, 1978; Brandt, 1986; Mehlman, 1989; Pleurdeau, 2001; Brooks et al., 2006) and some authors even see cultural patterning emerge from point types (Clark 1988). Morphologically, points are flake or flake fragments that have convergent lateral edges to make a pointed end. A point can be created using prepared core techniques which do not require retouching, or using unifacial, parti bifacial, or fully bifacial retouch. Points are problematic for two reaso ns. The first is determining function. On the one hand, points may be part of projectile technologies, such as spears or the bow and arrow that relied on hand thrusted, hand thrown, or mechanically projected implements. On the other hand, points may hav e been used as knives, drills, or other kinds of implements that emphasized the convergent morphology of the tool and the sharp lateral edges. Macrowear analysis, including size and morphology studies, may not be able to differentiate between points that were used for different functions because they can produce similar macroscopic use wear. Microwear analysis may be able to suggest point functions, but these studies rely on in depth experimental analysis of specific raw materials and point types. 31 The se cond problem is that the lack of well dated and continuous archaeological sequences across East Africa and the Horn provide little information about how point type and function 31 Microwear analysis is on g oing for Moche Borago (V. Rotts per comm.)
168 change over time and how point types between regions relate to each another. F or instance, in (1954) typology of Late Pleistocene bifacial and unifacial points was based on artifacts from open air sites or sites that had short, discontinuous sequence s. The longest sequence Clark was able to use came from Porc Epic (Ethiopia), which is now known to have a major unconformity and problematic dating. Consequently, Clark (1954; 1988) and others such as Pleurdeau (2001) were still unable to accurately describe temporal changes in point technology and style during the Late Pleistocene, and especially the change from Mode 3 to Mode 4/5 technologies. Moche Borago, however, provides the H orn with the first well dated sequence of Late Pleistocene points, though the assemblage is still small overall. The T Group and S Group have Mode 3 Levallois points, as well as Mode 4/5 unifacial, parti bifacial, and bifacial foliate points. 32 While the size of the assemblage limited the breadth of the analysis, there were qualitative patterns that relate to technology and typology, which are described here. Functional and quantitative differences are still being analyzed. The points are divided into two sub sections: The first provides metric information for the T Group and S Group points, while the second focuses on a subset of the best examples of unifacial and bifacial points from the T Group, the S Group, and R Group. The R Group points are inclu ded to discern the broader diachronic trends within the point assemblage as a whole. Relevant contextual information for the R Group is provided as needed. Summaries of technological and typological changes in the points are provided at the end of this c hapter. 32 There is no evidence for point preforms in the current assemblage.
169 The Point Assemblage The 105 points or point fragments (non MNL) from the T Group and S Group represent 28.6% of the total unshaped and shaped tools. Whole and point fragments are more common in the S Group (68.5%, n = 72) compared to the T Gro up Upper (18.1%, n = 19) or T Group Lower (13.3%, n = 14). Compared to other unshaped and shaped tools, the proportions of points in each litho stratigraphic group range from 25.7% ( n = 72) in the S Group to 38.5% in the T Group Upper (Table 8 1). The majority of points in the total assemblage are flaked unifacially with proportions ranging from 57% in the T Group Lower to 73% in the T Group Upper (Tables 8 24 and 8 25). Point fragments are also more common than whole points (Table 8 26), which may ind icate that the points were used and then discarded. Further evidence for point use is found in the type of fragments seen in the assemblage. Proximal point fragments (15%, n = 16) are twice as frequent as either medial or distal fragments. At Moche Bora go, many of the points are made on flakes with the proximal edge (i.e. the platform) located at the widest part of the tool, which is the butt, opposite the pointed edge. The proximal edge becomes contained in the haft when the point is oriented platform/ base. Therefore, if a point is broken during use, the proximal edge still in the haft may be more likely to be removed and discarded on site when the tool is repaired. However, orienting point platforms down also means that the thicker width around the b ulb of percussion might prevent a secure haft. Bulbar thinning is found on some points, but more than three quarters (78.9%) of points at Moche Borago do not show evidence for bulbar thinning (Tables 8 27 and 8 28). 33 33 Bifacial flaking is not included in these tables.
170 Basic metric data for whole points is provided in Table 8 29, which shows a slight decrease in the mean lengths of points from the T Group Lower (M = 31.3 mm, SD = 14.1) to the S Group (M = 25.9 mm, SD = 9.1) (Figure 8 5). Mean aspect ratios, however, do not change much, although the whole points in the T Group Lower are slightly more elongate than either the T Group Lower or S Group points (Figure 8 6). Point Descriptions This section provides detailed technological and typological descriptions of a sample of 20 Moche Borago points 34 Th e sample is derived primarily from G10 (n = 15), but points from H9 (n = 3), G9 (n = 1), and TU2 (n = 1) are also added to increase the sample size. R Group points are also included to broaden the diachronic depth of the analysis. Each point is illustrat ed in Figure 8 7. Detailed descriptions of each point are also provided in Appendix D. Technological Summary of the Moche Borago Points Two different technological traditions are represented within the Moche Borago points, as shown in Figure 8 8. Mode 3 points, characterized by the use of Levallois flaking methods, were found in both the T Group and S Group, but there are only four Mode 3 points in the entire assemblage (BN 3114.1, 3132.1, 1799.12, 1818.1). Each of these artifacts has a triangular pla tform, radial or convergent dorsal scars, and facetted flake blank platforms, which indicate that they were removed from prepared cores. BN 3132.1 is unique in that the dorsal scar pattern is opposed to the striking platform such as the Nubian Type 1 tech nique (Vermeersch, 2001) Mode 4/5 points are most clearly characterized by unifacial, parti bifacial, or fully bifacial retouching techniques. Intensive retouch is found on each piec e, and the retouch pattern 34 The points included in this sample were chosen because they were whole, but in some cases like Toora points, the flake blank is snapped, which indicates that these tool are either fragmentary or the tools were made on blanks that were snapped intentionally.
171 is most commonly semi circular to fully circular. Bifacial points are usually retouched more finely on one face. Furthermore, the planform shapes of points vary according to retouch type. Unifacial points appear to have the mo st triangular planform shapes and bifacial points have the most ovate planforms. One interpretation for the ovate bifacial point shape may be that these points were used and reworked more heavily prior to being discarded. BN 1818.1 appears to be a mix of both Mode 3 and Mode 4/5 techniques. The piece is made on a transverse flake with L evallois elements, including a radial dorsal scar pattern and facetted platform, which indicates that the flake blank was derived from a prepared core. The flake has b een retouched along the proximal and distal edges (which are equivalent to the lateral edges of the point) to create a sub triangular teardrop planform. The retouching, shape, and size of BN 1818.1 are similar to other unifacial and parti bifacial Mode 4/ 5 points, including BN 1693.3, 1802.90, 3477.3, 3575.3, and 2043.1. Each of these pieces was also made from end struck whole flakes, and each piece shares with BN 1818.1 bilateral or semi circular dorsal retouch that forms straight convergent lateral edge s that expand outward before contracting on a rounded point butt. Bulbar or basal thinning is evident on six parti bifacial points. Five of these points are retouched ventrally along the butt edge, as well as the ventral right lateral edges of the tool. 35 The right lateral ventral retouching always extends halfway up the axis of the tool. The purpose of the right lateral retouching, if significant, is currently unknown. Notches are present on four pieces (BN 4889.1, 2268.2, 1794.36, 1799.13), but it i s uncertain if all of the notches relate to hafting. The notching is most commonly shallow, unifacial, and present on only one lateral edge (possibly the flake blank right lateral). On BN 35 Orientation via the tool should not to be confused with the orientation of the fla ke blank.
172 4889.1, bilateral notches are located equidistantly approximately one half to two thirds from the point base. The left lateral notch is formed by a single large facet with several smaller auxiliary flake facets. The right lateral notch is formed by a single flake facet, but it also exhibits an abraded edge. The locati on of notching is similar across each of the four pieces. Six points (BN 1693.3, 3477.3, 3575.3, 4258.1, 4889.1, 1818.1) also have finer and more intensive retouching along the lateral edges, approximately one half to two thirds from the point base. The r etouching is so intensive on four points (BN 1693.3, 3477.3, 3575.3, 1818.1) that the lateral edge became slightly concave. In another two instances (BN 1818.1 and 4258.1), the lateral angle at the tip may have become more oblique due to more intensive re touching from the tip to midsection. These retouching patterns are believed to be related to hafting, and further evidence for hafting is found in BN 4889.1, which has bilateral notching at approximately the same location where the differential retouch on the other pieces begins. Therefore, if these points were hafted, then it is possible that the hafts obstructed the basal section of the point, meaning that only the exposed part of the point nearer the tip was retouched. Bifacially flaked points appear to be retouched much more intensively than either unifacial or part bifacial flaked points. Bifacial points BN 1799.16 and 1799.13 are especially retouched and their planforms are uncharacteristically ovate with more oblique tip angles. The other two bi facial points (BN 2116.2 and 4854.1), however, are slightly more elongate, and in the case of BN 2116.2, clearly have a teardrop planform, which might imply that BN 1799.16 and 1799.13 may have been used more heavily than BN 2116.2 and BN 4854.1 before bei ng discarded. In summary, the technological analysis of the Moche Borago whole points shows two technological traditions represented in the point assemblage. Mode 3 points are less frequent and are characterized by Levallois flaking techniques. Mode 4/ 5 points are made using either
173 unifacial, parti bifacial, or fully bifacial techniques, and notching and basal/bulbar thinning techniques can also occur. BN 1818.1 appears to be a mix of both Mode 3 and Mode 4/5 techniques. Typological Summary of the Mo che Borago Points Based on qualitative similarities observed in planform shapes (e.g. triangular, ovate, or teardrop) and styles of manufacture (e.g. notches, bulbar thinning, end struck flake blanks), I have been able to discern four possible typological groups within the Moche Borago points, which are named here for convenience. Figure 8 9 shows the distribution of the typological groups and the relationships between each group. Groups of points are first described individually and synchronically befor e being ordered chronologically to discuss typical point changes diachronically. Group 1: Levallois: Levallois is a method of flake manufacture, which gives this group more affinity to aspects of technology than typology, but the L evallois technique is also a shared characteristic among several of the points. There are three pieces in the Levallois Group: BN 3114.1; 3132.1; 1799.12. This group is defined by use of the Levallois point flaking method, which includes the use of facetted platforms and conv ergent flaking/dorsal scars on prepared cores. Levallois points are also not intensively retouched, which means that BN 1818.1 is not included within the Levallois Group even though it is made on a transverse Levallois flake. Group 2: Irra Xoketta Poi nts: Irra Xoketta 36 characteristic shape of this point type. Irra Xoketta points are identified by two primary characteristics: their shape and retouch type. Irra Xoketta points are often made on end struck whole flakes. Most Irra Xoketta points appear to be retouched more intensively on a single 36
174 surface of the tool, typically the flake blank dorsal surface. Bulbar/basal thinning is also common, which results in characteristic moderate to intensive ventral reto uch, as well as a biconvex/lenticular profile. The retouching pattern is often bilateral (lateral lateral) along the lateral edges of the tool and semi circular along the periphery of the point butt. This method of retouching creates a distinctive shape of the point, which is typically like a teardrop, but with subtle differences in the shape of Irra Xoketta points: Irra Xoketta Type I points (BN 3477.3, 3575.3, 2116.2, 1818.1, 1693.3, 2043.1) have a sub triangular planform with straight convergent lat eral edges, expanding outward that terminate at a rounded butt. Finer retouching from approximately one half to two thirds up from the point butt often creates a slightly concave lateral edge form. Irra Xoketta Type I points are frequently made on end st ruck whole flakes. Bulbar/basal thinning is common on both dorsal and ventral surfaces along the proximal edge, which doubles as the base of the point. The proximal edge retouching forms a distinctive rounded butt shape that distinguishes Type I points fr om Type II points. Irra Xoketta Type II points (BN 1887.1, 2072.1) have straight or slightly convex lateral edges, but they do not appear to expand outward as obliquely as Type I points. As a result, Type II points often appear more elongate, even foliate Type II points are also commonly made on end struck flake blanks, and bulbar/basal thinning is common. Unlike Type I points, Type II points are also characterized by more angular, trapezoidal butts. Group 3: Tona Points ( BN 4854.1, 1802.90, 1799.16, 1799.13, 4258.1): Tona points are bifacial with retouch that is characteristically finer and more intensive on one surface. Most of the Tona points within the current sample also seem to be heavily re worked. Thus, the majority
175 of the Tona points tend to be typically smaller than Irra Xoketta points with sub triangular or ovate platforms. Tona points are characterized by bilateral (tip butt)/convergent flaking of the most heavily retouched point surface. The other surface often has a single large basal flake scar, which may originate from the original flake blank ventral surface. Finer retouch is common along the tip and basal edges. Group 4: Toora Points ( BN 4889.1, 3574.1, 1794.36, 2268.2): Toora points are characterized by their shared triangular planform which is often produced by snapping (lateral lateral) the flake blank and retouching the distal lateral edges. Dorsal retouching is bilateral, and it seems to be most commonly lateral lateral oriented, though BN 3574.1 has bilateral tip butt ret ouch. Basal thinning is not present on Toora points. All four points within the current sample exhibit strong indications of hafting, including unilateral or bilateral shallow notching approximately one half to two thirds up from the tool and differentia l retouch patterns. In most cases, retouching of the dorsal surface (and in one case, the ventral surface: BN 2268.2) is finer beyond approximately one half to two thirds up from the point butt creating more oblique tip angles. Diachronic Technological a nd Typological Point Patterns Mode 3 points belonging to the Levallois Group are currently known only from the oldest and deepest strata at Moche Borago, which is the T Group Lower. The Mode 3 L evallois technique is still present in overlying S Group depo sits, but at this time it is also associated with Mode 4/5 retouching and morphology (on both points and microliths). For example, one distinct typological characteristic of Mode 4/5 Irra Xoketta points is the use of bilateral dorsal scars to produce a me dial ridge that may be related to prior Mode 3 techniques. More than 70% of Irra Xoketta points have a medial ridge, which increases the thickness of the points, although its functional value is unknown. Bifacial thinning is also evident in some points, such as BN 4854.1
176 and 1799.13, which shows that the technical means to remove medial ridges was present at that time. Mode 3 points in the assemblage (BN 3114.1 and BN 3132.1) also have medial ridges due to Levallois convergent and radial flaking techniq ues. Furthermore, the only point in the assemblage, which shares Mode 3 and Mode 4/5 characteristics (BN 1818.1), also has a medial ridge because it was removed from a prepared core that created the semi radial/convergent dorsal scar pattern. On other Ir ra Xoketta points the dorsal scars were made after the flake blank was removed from the core. Therefore, the presence of a medial ridge on Irra Xoketta points, via the dorsal flaking technique, may represent a purposely maintained technical tradition that was carried over from prior Mode 3 technologies. Otherwise, we might expect to see more bifacial thinning techniques, such as on BN 4854.1 and 1799.13, which make thinner points that are easier to haft. The use of Mode 3 and Mode 4/5 techniques to make microliths, such as BN 1818.6, suggests that continuity between technologies was not limited to points alone. In BN 1818.6, symmetry of the two technological traditions occurs; the former Mode 3 L evallois technique to produce the flake and the latter Mode 4/5 backing to create the tool. Therefore, it does not seem likely that a discrete break occurs between technological modes at Moche Borago. Gradual changes among types, such as the morphological variations that define Type I and Type II Irra Xoketta po ints, may also reflect similar processes of continuous technological change that resulted in subtle typological variations over time. Conclusions How do the points articulate with other aspects of the lithic assemblage? Based on the qualitative diffe rences in point shape and design, there may be continuous technological changes that resulted in subtle typological changes in point morphology over time. Future quantitative
177 analysis on a larger sample of points is expected to provide statistically signi ficant point changes over time. In contrast, the other shaped tools, unshaped tools, and cores reveal differences in the types and frequencies of these artifacts between the S Group and the T Group. On the one hand, the points show subtle changes and i nferred continuity, and on the other hand, the other artifacts suggest discontinuity or rapid change. How do these observations co relate? First, it is important to account for cultural traditions at this time, including social and symbolic practices tha t may not be easily identified in the archaeological record. Clark (1988) suggested that hunter gatherers in the Horn of Africa had developed cultural traditions and regional identities by the Late P leistocene, which included unique point types. Many of the points in the Horn are small and have foliate shapes, and this regional pattern is substantiated by newer studies at sites such as at Aduma (Ethiopia) (Yellen et al. 2005; Brooks et al., 2006) and Porc Epic (Pleurdeau, 2001; Pleurdeau, 2005) this regional cultural tradition, and air (e.g. Aduma) sites or sites with questionable dating (e.g. Porc Epic). Many of the escribed by Pleaurdeu (2001) also show strong similarities to the Moche Borago points, which may indicate similar ages, even though their actual age is questionable. Therefore, the points show that the Moche Borago T Group and S Group assemblages were probably made by groups of hunter gatherers with similar cultural traditions, which may also relate to a broader cultural tradition found across the Horn. Second, while cultural patterns may have remained largely unchanged that does not mean t change. For example, residency time at the site may have influenced the breadth of on site activities, and
178 longer stays might account for more diverse stone artifact types in the S Group versus shorter stays and fewer artifact types in the T Group. Sho rter stays during the T Group may also account for the lack of curated tools from these levels. Simply put, the lack of formal shaped tools coupled with large amounts of debitage might indicate that the shaped tools were being repaired or modified on site at this time for the purposes of being used elsewhere. In lieu of using their shaped tools then, the occupants of the site relied on informal and expedient unshaped tools during their time at the site. Therefore, while the points may imply cultural or at least, technological continuity between the T Group and the S Group, these observations are not at odds with data about the other shaped tools, unshaped tools and cores. Stone artifacts other than the points appear to be influenced by different facto rs than the points, namely site use and mobility, which may be due to ecological changes in the area between the T Group and S Group.
179 Table 8 1: Frequency (upper) and percentages (lower) of the unshaped and shaped tools per total number of artifacts i n this class. Backed Scrapers Points Burins Misc. Shaped Unshaped Tools Total S Group 15 47 70 8 14 125 279 T Group Upper 2 7 19 3 5 15 51 T Group Lower 0 7 14 1 3 14 39 Total 17 61 103 12 22 154 369 Backed Scrapers Points Burins Misc. Shaped Unshaped Tools S Group 4.07% 12.74% 18.97% 2.17% 3.79% 33.88% T Group Upper 0.54% 1.90% 5.15% 0.81% 1.36% 4.07% T Group Lower 0.00% 1.90% 3.79% 0.27% 0.81% 3.79% Note: These values include tool fragments. Table 8 2: Percentage of unshaped a nd shaped tools per stratigraphic aggregate. Backed Scrapers Points Burins Misc. Shaped Unshaped Tools Total S Group 5.38% 16.85% 25.09% 2.87% 5.02% 44.80% 279 T Group Upper 3.92% 13.73% 37.25% 5.88% 9.80% 29.41% 51 T Group Lower 0.00% 17.95% 35.90% 2. 56% 7.69% 35.90% 39
180 Table 8 3: MNL frequency (upper) and percentages (lower) of the unshaped and shaped tools in the S Group and T Group. Backed Scrapers Points Burins Misc. Shaped Unshaped Tools Total S Group 8 41 29 8 9 57 152 T Group Upper 2 5 7 3 3 10 30 T Group Lower 0 4 4 1 3 7 19 Total 10 50 40 12 15 74 201 Backed Scrapers Points Burins Misc. Shaped Unshaped Tools S Group 3.98% 20.40% 14.43% 3.98% 4.48% 28.36% T Group Upper 1.00% 2.49% 3.48% 1.49% 1.49% 4.98% T Group Lower 0.00% 1.99% 1.99% 0.50% 1.49% 3.48%
181 Table 8 4: Frequency and percentage per stratigraphic group of unshaped tools based on non MNL counts (above) and MNL counts (below). Modified Utilized Total S Group 64 51.20% 61 48.80% 61 T Group Upper 6 40. 00% 9 60.00% 10 T Group Lower 7 50.00% 7 50.00% 8 Total 77 50.00% 77 50.00% 79 Modified Utilized Total S Group 26 45.61% 31 54.39% 57 T Group Upper 4 40.00% 6 60.00% 10 T Group Lower 4 57.14% 3 42.86% 7 Total 34 45.9 5% 40 54.05% 74
182 Table 8 5: Percentages of edge damage on unshaped tools (MNL) based on the location of edge damage. Dorsal Left Lateral Right Lateral Both Laterals Distal Distal Combo. Proximal Combo. Semi Circular S Group 13.5% 13.5% 17.6% 5.4 % 5.4% 0.0% 1.4% T Group Upper 1.4% 0.0% 4.1% 2.7% 1.4% 1.4% 0.0% T Group Lower 2.7% 0.0% 0.0% 0.0% 1.4% 0.0% 0.0% Ventral Left Lateral Right Lateral Both Laterals S Group 2.7% 2.7% 2.7% T Group Upper 1.4% 0.0% 0.0% T Group Lower 0.0% 2.7% 0.0% Bifacial Left Lateral Right Lateral Both Laterals Distal Combo. Indet. Lateral Combo. S Group 4.1% 1.4% 1.4% 0.0% 0.0% 5.4% T Group Upper 0.0% 0.0% 0.0% 1.4% 0.0% 0.0% T Group Lower 0.0% 0.0% 0.0% 0.0% 1.4% 1.4% Edge Damage Face Percentage Total Do rsal 71.6% 53 Ventral 12.2% 9 Both 16.2% 12 Total 74
1 83 Table 8 6: Edge damage type on unshaped tools (MNL). Dorsal Ventral Bifacial/Parti bifacial Simple Stepped Combo Simple Stepped Combo Simple Combo Total S Group 56.9% 1.7% 19.0% 8.6% 0.0% 0 .0% 8.6% 5.2% 58 T Group Upper 70.0% 0.0% 10.0% 10.0% 0.0% 0.0% 10.0% 0.0% 10 T Group Lower 42.9% 0.0% 0.0% 14.3% 14.3% 14.3% 14.3% 0.0% 7 Total Percent 57.3% 1.3% 16.0% 9.3% 1.3% 1.3% 9.3% 4.0% 75 Group Percent 74.7% 12.0% 13.3% Table 8 7: Relative percentage of blank types used for modified and utilized unshaped tools (non MNL). Level Side Struck Flake End Struck Flake Flake Fragments Levallois flake Levallois Point Indet. CTF Burin Total Modified Tools S Group Percentage 6.30% 68 .80% 15.60% 6.30% 3.10% 64 T Group Upper Percentage 83.30% 16.70% 6 T Group Lower Percentage 85.70% 14.30% 7 Modified Tools Group Percentage 5.20% 71.40% 13.00% 1.30% 6.50% 2.60% 77 Utilized Tools S Group Percentage 1.60% 83.60% 6.60% 1.60% 1.60% 3.30% 1.60% 61 T Group Upper Percentage 88.90% 11.10% 9 T Group Lower Percentage 85.70% 14.30% 7 Utilized Tools Group Percentage 1.30% 84.40% 6.50% 1.30% 1.30% 1.30% 2.60% 1.30% 77
184 Table 8 8: Mean aspect ratio and length values for whole modified and utilized unshaped tools (MNL). Modified, n = 34 Aspect Ratio Length S Group Mean 1.71 23.9 Std. Deviation 0.84 11.21 T Group Upper Mean 1.82 33.35 Std. Deviation 0.85 13.83 T Group Lower Mean 1 .44 28.54 Std. Deviation 0.75 8.7 Utilized, n = 40 Aspect Ratio Length S Group Mean 2.2 26.2 Std. Deviation 0.9 7.51 T Group Upper Mean 1.71 36.51 Std. Deviation 0.92 7.57 T Group Lower Mean 1.28 33.4 Std. Deviation 0.34 5.31 Figur e 8 1: Average aspect ratio of unshaped (modified and utilized) whole flake blanks between the three litho stratigraphic groups.
185 Table 8 9: Basic metrical data and descriptions for miscellaneous shaped tools in the S Group and T Group (MNL). Bag Number Excavation Unit Level Stratigraphic Aggregate Artifact Type Raw Material Length (mm) Breadth (mm) Thickness (mm) 1794.15 G10 11 S Group Drill CBO 35.18 29.47 12.12 1794.27 G10 11 S Group Outil ecaille CBO 27 27.85 8.69 1794.29 G10 11 S Group Awl CBO 29.87 13.44 5.44 1794.39 G10 11 S Group Awl CBO 19.4 9.17 5.1 1799.17 G10 12 S Group Outil ecaille CBO 27.36 27.33 6.56 1802.102 G10 13 S Group Bec CBO 30.86 14.19 9.67 1802.136 G1 0 13 S Group Awl CBO 17 11.65 6.25 1802.82 G10 13 S Group Awl CBO 26.34 16.18 10.32 1805.69 G10 1 4 S Group Awl CBO 15.69 13.85 3.42 3105.11 G10 23 T Group Upper Irregular CBO 27.68 23 3.75
186 Table 8 10: Basic metrical data and descriptions of the microliths found in the S Group and T Group. Bag Number Excavation Unit Level Stratigraphic Aggregate Artifact Type Raw Material Length (mm) Breadth (mm) Thickness (mm) Type of Backing Angle of Backing 1799.15 G10 12 S Group Straight Back CBO 25.22 11.01 3.77 Bidirectional 88 1799.42 G10 12 S Group Straight Back CBO 21.92 15.68 3.93 Inverse 95 1802.2 G10 13 S Group Microlith Fragment CBO 11.64 6.61 4.43 Obverse 100 1802.4 G10 13 S Group Collateral Backed CBO 27.18 7.34 5.42 Bidirectional 68 1802.54 G10 13 S Grou p Orthogonal Truncation CBO 15.71 9.21 7.06 Inverse 68 1802.98 G10 13 S Group Backed Blade CBO 26.27 18.37 4.92 Obverse 65 1805.54 G10 14 S Group Curved Back CBO 21.85 12.25 6.9 Bidirection 90 1805.58 G10 14 S Group Backe d Blade CBO 20.42 13.43 6.45 Natural 60 1805.73 G10 14 S Group Microlith Fragment CBO 15.11 8.1 5.32 Natural 90 1805.75 G10 14 S Group Microlith Fragment CBO 13.54 10.13 5.5 Obverse 82 1805.77 G10 14 S Group Microlith Fragment CB O 15.46 8.97 2.66 Obverse 85 1805.78 G10 14 S Group Straight Back CBO 18.25 10.75 3.34 Obverse 78 1805.83 G10 14 S Group Curved Back CBO 11.04 10.14 2.39 Obverse 95 1818.5 G10 16 S Group Collateral Backed CBO 20.35 5.25 4.68 Bid irectional 55 1818.6 G10 16 S Group Straight Back CBO 31.81 15.78 3.87 Inverse/Alt 88 2571.1 G10 18 T Group Upper Crescent CBO 30.1 7.83 2.93 Obverse/Alt 88 3105.12 G10 23 T Group Upper Straight Back CBO 13.34 7.39 3.11 natural 87 Note : Mi croliths listed are complete unless otherwise noted in artifact type.
187 Figure 8 2: Microliths from the BXA T Group and S Group deposits.
188 Table 8 11: Frequency of scrapers and scraper fragments by type within each level and stratigraphic aggregate. End Scraper Side Scraper End and Single Side End and Double Side Double Side Opposite side Dorsal/ Ventral Notch and End or Side Scraper Convergent Scraper Notched Scraper Irregular Fragment Total S Group 11 8 2 0 0 0 1 2 1 14 12 1 0 1 1 0 0 0 0 3 13 7 2 1 3 0 4 6 0 23 14 2 1 0 0 1 0 3 0 7 Total 18 5 2 4 1 5 11 1 47 T Group Upper 18 2 0 0 0 0 2 19 0 0 1 0 0 1 21 1 0 0 0 0 1 22 0 1 0 0 0 1 23 0 0 0 0 1 1 24 0 0 0 1 0 1 Total 3 1 1 1 1 7 T Group Lower 25 0 1 0 0 0 0 0 1 27 1 0 0 1 0 0 0 2 28 0 0 0 0 0 0 1 1 29 0 0 0 0 0 1 0 1 30 0 0 1 0 1 0 0 2 Total 1 1 1 1 1 1 1 7 Note: Counts are not filtered for MNL.
189 Table 8 12: Relative percentage of scraper types and scraper fragments per level. End Scraper Side Scraper End and Single Side End and Double Side Double Side Opposite side Dorsal/ Ventral Notch and End or Side Scraper Convergent Scraper Notc hed Scraper Irregular Fragment S Group 11 57.10% 14.30% 7.10% 14.30% 7.10% 12 33.30% 33.30% 33.30% 13 30.40% 8.70% 4.30% 13.00% 17.40% 26.10% 14 28.60% 14.30% 14.30% 42.90% Total 38.30% 10.60% 4.30% 8.50% 2.10% 10.60% 23 .40% 2.10% T Group Upper 18 100.00% 19 100.00% 21 100.00% 22 100.00% 23 100.00% 24 100.00% Total 42.90% 14.30% 14.30% 14.30% 14.30% T Group Lower 25 100.00% 27 50.00% 50.00% 28 100.00% 29 100.00% 30 50.00% 50.00% Total 14.30% 14.30% 14.30% 14.30% 14.30% 14.30% 14.30% Note: Percentages are not based on MNL counts.
190 Table 8 13: MNL frequen cy of scrapers and scraper fragments by type within each level and stratigraphic aggregate. End Scraper Side Scraper End and Single Side End and Double Side Double Side Convergent Scraper Notched Scraper Notch and End or Side Scraper Irregular Total S Group 11 8 2 0 0 1 2 0 13 12 1 0 1 1 0 0 0 3 13 6 2 0 3 3 5 0 19 14 2 1 0 0 0 2 1 6 Total 17 5 1 4 4 9 1 41 T Group Upper 18 2 0 0 0 2 19 0 1 0 0 1 23 0 0 0 1 1 24 0 0 1 0 1 Total 2 1 1 1 5 T Group Lower 25 0 1 0 0 1 27 1 0 0 1 2 30 0 0 1 0 1 Total 1 1 1 1 4
191 Table 8 14: Relative percentage of scraper types and scraper fragments per level based on MNL counts. End Scraper Side Sc raper End and Single Side End and Double Side Double Side Convergent Scraper Notched Scraper Notch and End or Side Scraper Irregular Total S Group 11 61.50% 15.40% 7.70% 15.40% 13 12 33.30% 33.30% 33.30% 3 13 31.60% 10.50% 15.80% 15.80 % 26.30% 19 14 33.30% 16.70% 33.30% 16.70% 6 Total 41.50% 12.20% 2.40% 9.80% 9.80% 22.00% 2.40% 41 T Group Upper 18 100.00% 2 19 100.00% 1 23 100.00% 1 24 100.00% 1 Total 40.00% 20.0 0% 20.00% 20.00% 5 T Group Lower 25 100.00% 1 27 50.00% 50.00% 2 30 100.00% 1 Total 25.00% 25.00% 25.00% 25.00% 4
192 Table 8 15: Basic metrical data and descriptions of the scrapers (non MNL) fo und in the S Group and T Group. Bag Number Level Stratigraphic Group Raw Material Scraper Type Length (mm) Breadth (mm) Thickness (mm) Angle Left Angle Right Angle Proximal Angle Distal 1794.10 11 S Group CBO Side Scraper 32.63 14.57 14.18 89 1794.1 8 11 S Group CBO End Scraper 29.95 17.72 12.56 95 1794.20 11 S Group CBO End Scraper 34.58 23 10.19 71 1794.24 11 S Group CBO Convergent Scraper 23.87 12.58 9.4 85 65 1794.28 11 S Group CBO End Scraper 20.28 14.49 6.59 45 78 1794.33 11 S Grou p CBO Irregular 11.79 9.74 4.48 78 78 61 76 1794.34 11 S Group CBO End Scraper 10.99 12.68 7.43 91 1794.35 11 S Group CBO End Scraper 16.01 16.01 5.26 98 1794.38 11 S Group CBO End Scraper 26.27 20.09 9.96 83 1794.41 11 S Group CBO Side Scrape r 30.91 12.29 6.22 70 68 1794.51 11 S Group CBO Notched Scraper 16.39 19 4.45 73 78 88 1794.54 11 S Group CBO End Scraper 18.37 12.37 4.76 85 1794.71 11 S Group CBO End Scraper 23.49 20.08 4.28 88 1794.76 11 S Group CBO Notched Scraper 23.77 1 4.43 6.08 56 90 1799.29 12 S Group CBO End and Double Side 22.24 15.89 6.43 58 65 70 1799.32 12 S Group CBO End and Single Side 24.11 14.32 5.7 68 66 1799.45 12 S Group CBO End Scraper 17.93 22.49 6.84 83 1802.104 13 S Group CBO End Scraper 23. 78 12.18 6.15 75 1802.112 13 S Group CBO Notched Scraper 24.71 18.98 7.64 52 1802.115 13 S Group CBO End Scraper 20.48 13.37 5.42 80 1802.116 13 S Group CBO End Scraper 16.62 17.25 8.03 82 1802.118 13 S Group CBO End Scraper 19.28 11.92 6.0 6 88 1802.123 13 S Group CBO Side Scraper 19.3 15.42 7.2 1802.124 13 S Group CBO Notched Scraper 18.98 14.03 3.17 61 60
193 Table 8 15: Continued. Bag Number Level Stratigraphic Group Raw Material Scraper Type Length (mm) Breadth (mm) Thicknes s (mm) Angle Left Angle Right Angle Proximal Angle Distal 1802.126 13 S Group CBO End and Double Side 19.64 11.7 7.6 64 84 88 1802.127 13 S Group CBO Convergent Scraper 17.23 10.44 5.6 69 90 86 1802.25 13 S Group CBO Convergent Scraper 37.08 15.57 8.1 4 71 55 35 1802.26 13 S Group CBO Notched Scraper 25.82 14.11 4.38 70 68 1802.301 13 S Group CBO End Scraper 31.77 13.68 4.6 1802.34 13 S Group CBO Convergent Scraper 18.39 15.97 7.03 75 80 1802.41 13 S Group CBO End Scraper 15.16 12.75 5.16 80 1802.57 13 S Group CBO End and Double Side 18.53 11.88 4.84 82 67 93 1802.65 13 S Group CBO Side Scraper 13.95 19.45 8.56 75 1802.67 13 S Group CBO End and Single Side 13.66 11.76 5.93 68 80 1802.68 13 S Group CBO End Scraper 36.73 20.53 8.0 1 40 1802.80 13 S Group CBO Notched Scraper 27.48 21.18 6.69 70 1802.87 13 S Group CBO Convergent Scraper 32.8 17.07 8.25 75 1802.88 13 S Group CBO End and Double Side 23.97 22.78 6.46 72 71 75 1802.91 13 S Group CBO Notched Scraper 42.58 17. 09 6.3 68 65 1802.92 13 S Group CBO Notched Scraper 22.68 14.93 3.81 44 52
194 Table 8 15: Continued. Bag Number Level Stratigraphic Group Raw Material Scraper Type Length (mm) Breadth (mm) Thickness (mm) Angle Left Angle Right Angle Proximal Angle D istal 1805.24 14 S Group CBO End Scraper 16.23 13.7 5.63 82 1805.5 14 S Group CBO Notched Scraper 37.19 24.02 5.72 68 81 1805.57 14 S Group CBO Side Scraper 14.43 12.99 4.9 85 98 1805.70 14 S Group CBO Notch and end or side scraper 19.25 12.17 6 .77 83 60 75 1805.74 14 S Group CBO Notched Scraper 14.72 10.92 3.87 81 78 1805.80 14 S Group CBO Notched Scraper 23.23 15.22 6.27 81 1805.82 14 S Group CBO End Scraper 23.63 13.24 4.28 82 3574.1 25 T Group Lower CBO Side Scraper 38.16 26.29 9 .7 56 74 3730.3 27 T Group Lower CBO End Scraper 17.32 15.49 5.86 68 3730.4 27 T Group Lower CBO Double Side 32.02 21.58 8.75 50 55
195 Table 8 15: Continued. Bag Number Level Stratigraphic Group Raw Material Scraper Type Length (mm) Breadth (mm) Thickness (mm) Angle Left Angle Right Angle Proximal Angle Distal 4109.6 28 T Group Lower CBO Fragment, Indet. 15.36 16.5 4.68 71 4100.2 29 T Group Lower CBO Notch and end or side scraper 27.36 24.56 7.65 55 56 4105.2 30 T Group Lower CBO Oppo site side dorsal / ventral 31.31 23.98 5.06 47 50 4906.7 30 T Group Lower CBO End and Double Side 24.45 18.8 7.41 82 74 85 2254.3 18 T Group Upper CBO End Scraper 16.28 14.37 5.07 70 2254.7 18 T Group Upper CBO End Scraper 30.2 16.8 8.47 80 21 16.8 19 T Group Upper CBO End and Single Side 34.65 17.57 7.84 83 75 2129.3 21 T Group Upper CBO End Scraper 23.44 20.65 7.48 67 2133.3 22 T Group Upper CBO Side Scraper 32.51 25.57 15.29 68 3105.12 23 T Group Upper CBO Irregular 50.93 19.99 6.1 4 3575.7 24 T Group Upper CBO Notched Scraper 31.54 25.84 6.41 77
196 Table 8 16: Mean aspect ratios of all scrapers types (MNL) subdivided by litho stratigraphic group. Aspect Ratio Mean Std. Deviation N S Group 1.53 0.45 41 T Group Upper 1.7 3 0.58 5 T Group Lower 1.34 0.17 4 Table 8 17: Mean aspect ratios of MNL end scrapers only subdivided by litho stratigraphic group. Aspect Ratio Mean Std. Deviation N S Group 1.43 0.41 17 T Group Upper 1.47 0.47 2 T Group Lower 1.12 1
197 Ta ble 8 18: Distribution of retouch location per scraper type (MNL). Dorsal Ventral Bifacial Parti bifacial Total End Scraper 95.00% 5.00% 20 Side Scraper 100.00% 6 End and Single Side 100.00% 2 End and Double Side 100.00% 5 Convergen t Scraper 50.00% 25.00% 25.00% 4 Notched Scraper 80.00% 10.00% 10.00% 10 Irregular 100.00% 1 Double Side 100.00% 1 Notch and End or Side Scraper 100.00% 1
198 Table 8 19: Location of retouch per scraper type. Left Lateral Right Later al Both Laterals Distal Distal Combination Proximal Proximal Combination Semi Circular Circular Total End Scraper 5.30% 52.60% 21.10% 10.50% 5.30% 5.30% 20 Side Scraper 33.30% 33.30% 16.70% 16.70% 6 End and Single Side 50.00% 50.00% 2 En d and Double Side 80.00% 20.00% 5 Convergent Scraper 50.00% 50.00% 4 Notched Scraper 25.00% 37.50% 12.50% 25.00% 10 Irregular 1 Double Side 100.00% 1 Note: End scrapers are retouched more frequently on the distal rather than the proximal, and side scrapers are retouched on either mal nd at least one other edge.
199 Table 8 20: Angle of the working edges for end and side scrapers. End Scraper Side Scraper Distal Proximal Left Mean Std. Deviation N Mean Std. Deviation N Mean Std. Deviation N S Group 79.69 13.75 13 77.5 22.31 4 74.37 10.60 8 T Group Upper 75.00 7.07 2 T Group Lower 68.00 1 69.00 17.01 3 Note: Derived from MNL filtered data. Table 8 21: Frequency of working edge form for end scrapers (above) and side scrapers (below). End Scraper Working Edge Form Level Straight Convex Serrated Straight Convex Irregular Total 11 2 5 0 1 0 8 12 1 0 0 0 0 1 13 2 1 1 2 0 6 14 2 0 0 0 0 2 18 1 0 0 0 1 2 27 1 0 0 0 0 1 Total 9 6 1 3 1 20 Total Percentage 45.0% 30.0% 5.0% 15.0% 5.0% Si de Scraper Working Edge Form Level Straight Convex Concavo Convex Straight Convex Total 11 1 0 0 1 2 13 1 1 0 0 2 14 0 1 0 0 1 25 0 0 1 0 1 Total 2 2 1 1 6 Total Percentage 33.3% 33.3% 16.7% 16.7% Note: The data are based on MNL coun ts.
200 Table 8 22: Basic metric information on notched flakes from the S Group and T Group. Bag Number Level Length (mm) Width (mm) Thickness (mm) Retouch Class Notch Width (mm) Notch Depth (mm) Notch Location 1794.76 11 23.77 14.43 6.08 Obverse 7.68 1.2 Distal Dorsal 1794.51 11 16.39 19 4.45 Inverse 11.25 1.95 Ventral Rt. Lat 1802.92 13 22.68 14.93 3.81 Obverse 7.06 2.14 Dorsal Rt. Lat 1802.124 13 18.98 14.03 3.17 Obverse Dorsal Both Laterals 1802.26 13 25.82 14.11 4.38 Obverse 7.1 2.71 Dorsal D istal Rt. Lat 1802.80 13 27.48 21.18 6.69 Obverse 10.2 2.62 Distal Dorsal 1802.112 13 24.71 18.98 7.64 Inverse 7.44 2.42 Ventral right 1802.91 13 42.58 17.09 6.3 Part Bifacial 15.99 3.63 Prox Left Lat Ventral 1805.80 14 23.23 15.22 6.27 Obverse Dorsa l Right Lateral 1805.5 14 37.19 24.02 5.72 Obverse 10.25 Dorsal Left Prox 1805.74 14 14.72 10.92 3.87 Obverse Dorsal Distal and Right Lat 3575.7 24 31.54 25.84 6.41 Obverse 17.14 3.31 Dorsal Right Lateral Note: All pieces are made of CBO. All artif acts listed are made on whole flakes except for BN 1802 .26 and 1805.74.
201 Table 8 23: Basic descriptive information of burins. Bag Number Level Strat Group Artifact Type Raw Material Length (mm) Breadth (mm) Thickness (mm) Number of Burin Spalls Width of B urin Cutting Edge (mm) Angle of Burin 1794.5 11 S Group Other CBO 27.22 16.09 6.4 1 7 65 1802.99 13 S Group Angle CBO 28.75 29.64 9.25 1 9 75 1802.134 13 S Group Angle CBO 34.83 11.96 5.25 1 4 89 1802.98 13 S Group Angle CBO 46.42 18.41 5.89 1 1805. 11 14 S Group Angle CBO 40.23 22.48 9.92 1 7 97 1805.19 14 S Group Dihedral CBO 30.21 13.57 8.22 1 9 57 1805.5 14 S Group Dihedral CBO 37 24 6 1 1818.12 16 S Group Dihedral Grey Obs. 31.98 12.3 11.2 3 5 52 2116.1 19 T Group Upper Other CBO 44.46 10.0 1 7.43 1 6 2116.4 19 T Group Upper Angle CBO 36.6 20.1 5.28 1 4 83 2133.11 22 T Group Upper Multiple CBO 50.69 23.2 10.79 4 6 95 4906.2 30 T Group Lower Dihedral CBO 20.39 10.1 4.83 6 5 52 Note: All tools are whole, but some burins are made on flake f ragments.
202 Figure 8 3: Angle burin from Level 13 with possible backing along the right lateral edge. Figure 8 4: Dihedral burin recovered from Level 30.
203 Table 8 24: Frequencies of point retouch type. Level Unifacial Parti Bifacial B ifacial Total S Group 11 10 3 1 14 12 3 1 4 8 13 18 7 1 26 14 14 5 1 20 15 2 0 0 2 16 2 0 0 2 Total 49 16 7 72 T Group Upper 18 2 0 2 19 3 1 4 22 4 1 5 23 2 2 5 24 3 0 3 Total 14 4 19 T Group Lower 25 3 0 0 3 28 1 0 1 2 29 0 1 0 1 30 4 2 1 7 32 0 1 0 1 Total 8 4 2 14 Note: Data derived from non MNL counts. Table 8 25: Percentages of point retouch type. Unifacial Parti Bifacial Bifacial S Group 68.1% 22.2% 9.7% T Group Upper 73.7% 21.1% T Gr oup Lower 57.1% 28.6% 14.3% Note: Data derived using non MNL counts.
204 Table 8 26: Completeness of the points (non MNL) within the T Group and S Group. Whole Broken Indet. Proximal Fragment Medial Fragment Distal Fragment Total S Group 29 22 11 2 6 70 T Group Upper 7 4 3 5 1 19 T Group Lower 4 3 2 2 3 14 Total 40 29 16 9 10 103 38.8% 28.2% 15.5% 8.7% 9.7% Table 8 27: Frequency of bulbar thinning on points. Level Absent Marginal Marginal to Semi Invasive Semi Invasive Invasive Tot al S Group 11 5 0 1 0 0 6 12 2 1 0 0 0 3 13 11 1 1 1 1 15 14 8 0 1 0 1 10 15 2 0 0 0 0 2 16 2 0 0 0 0 2 Total 30 2 3 1 2 38 T Group Upper 18 0 0 0 0 19 2 0 0 2 22 1 0 0 1 23 1 0 0 1 24 0 1 1 2 Total 4 1 1 6 T Group Lower 25 1 0 0 1 28 0 0 0 0 29 1 0 0 1 30 2 1 1 4 32 0 0 0 0 Total 4 1 1 6 Note: Pieces with no evident bulb of percussion have been omitted from this table.
205 Table 8 28: Percentage o f bulbar thinning found on points. Level Absent Marginal Marginal to Semi Invasive Semi Invasive Invasive S Group 11 83.3% 16.7% 12 66.7% 33.3% 13 73.3% 6.7% 6.7% 6.7% 6.7% 14 80.0% 10.0% 10.0% 15 100.0% 16 100.0% Total 78.9% 5.3% 7.9% 2.6% 5.3% T Group Upper 18 19 100.0% 22 100.0% 23 100.0% 24 50.0% 50.0% Total 66.7% 16.7% 16.7% T Group Lower 25 100.0% 28 29 100.0% 30 50.0% 25.0% 25.0% 32 Total 66.7% 16.7% 16.7% Note: Pieces with no evident bulb of percussion have been omitted from this table.
206 Table 8 29: Basic metric information for whole points. Bag Number Level Group Point Retouch Type Raw Material Phy sical Condition Length (mm) Breadth (mm) Thickness (mm) 1794.25 11 S Group Unifacial Point CBO Fresh 48.16 22.54 6.18 1794.44 11 S Group Parti Bifacial Point CBO Fresh 21.6 13.78 4.98 1794.48 11 S Group Parti Bifacial CBO Fresh 16.75 14.53 4.12 1794.49 11 S Group Bifacial Point CBO Fresh 38.06 20.62 8.62 1794.52 11 S Group Unifacial Point CBO Fresh 36.63 16.51 6.69 1799.12 12 S Group Parti Bifacial Point CBO Fresh 31.36 24.68 8.23 1799.13 12 S Group Parti Bifacial CBO Fresh 24.16 18.43 4.69 1799.16 12 S Group Bifacial Point CBO Fresh 24.99 21.99 6.96 1802.105 13 S Group Parti Bifacial CBO Fresh 36.4 16.26 5.4 1802.108 13 S Group Unifacial Point CBO Fresh 31.39 17.62 8.43 1802.121 13 S Group Unifacial Point CBO Fresh 18.18 14.46 5.04 1802.135 13 S Group Parti Bifacial CBO Fresh 20.02 17.76 5.93 1802.85 13 S Group Parti Bifacial CBO Fresh 30.92 22.12 6.25 1802.86 13 S Group Unifacial Point CBO Fresh 23.95 27.55 9.11 1802.9 13 S Group Unifacial Point CBO Fresh 20.63 29.27 8 1805.37 14 S Group Uni facial Point CBO Fresh 22.47 12.29 4.95 1805.55 14 S Group Unifacial Point CBO Slightly Abraded 19.65 10.29 5.68 1805.62 14 S Group Unifacial Point CBO Moderately Abraded 23.37 14.41 8.31 1805.63 14 S Group Unifacial Point CBO Fresh 11.76 10.7 4 1806.1 15 S Group Unifacial Point CBO Fresh 12.7 11.02 4.83 1806.9 15 S Group Unifacial Point CBO Fresh 23.46 16.81 7.69 1818.1 16 S Group Unifacial Point CBO Fresh 45.11 27.85 7.17 1818.4 16 S Group Unifacial Point CBO Fresh 8.88 12.41 3.65
207 Table 8 29: Continued. Bag Number Level Group Point Retouch Type Raw Material Physical Condition Length (mm) Breadth (mm) Thickness (mm) 2116.2 19 T Group Upper Bifacial Point CBO Fresh 30.92 19.31 5.96 2116.3 19 T Group Upper Parti Bifacial CBO Fresh 25.37 36.5 7 .25 2116.7 19 T Group Upper Unifacial Point CBO Fresh 31.94 20.21 7.63 2133.2 22 T Group Upper Parti Bifacial CBO Fresh 49.68 34.69 13.72 3132.1 29 T Group Upper Parti Bifacial Point CBO Fresh 51.03 20.57 6.98 3477.3 24 T Group Upper Parti Bifacial CBO Fresh 40.17 22.2 8.25 3575.3 24 T Group Upper Parti Bifacial CBO Fresh 33.61 19.98 8.89 4564.3 25 T Group Lower Unifacial Point CBO Slightly Abraded 27.6 17.29 3.27 4105.7 30 T Group Lower Parti Bifacial Point CBO Fresh 17.66 12.06 5.03 4854.1 30 T Gr oup Lower Parti Bifacial CBO Fresh 29 18.46 6.38
208 Figure 8 5: Histogram of the mean lengths of whole points per stratigraphic aggregate.
209 Figure 8 7: Points from the Moche Borago OIS 3 deposits. These points come from the BXA and TU2 area s, as well as the T Group, S Group, and the R Group.
210 Figure 8 8: Point technological traditions represented at Moche Borago.
211 Figure 8 9: Point typology for Moche Borago. The points represented are ordered in a roughly chronological sequence.
212 CHAPTER 9 SYNTHESIS This dissertation has studied human behavioral change during a known period of climatic flux, Oxygen Isotope Stage 3 (OIS 3), dated 59.4 to 27.8 ka. My study focused primarily on the frequencies and kinds of flaked stone artifacts tha t were used at the site from ~60 ka to ~43 ka to infer changes in hunter gatherer subsistence strategy and site use. Based upon these findings, and drawing upon a top down, multi scalar methodology that described global and hemispheric climatic changes tha t likely affected regional and local environments, I have suggested that humans living in and around Moche Borago may have altered their behaviors in response to ecological changes brought on by monsoonal flux during OIS 3. In Chapter 1, I posed four rese arch questions: 1) What is the evidence at global, continental, and local scales for climatic fluctuations during early and middle OIS 3? 2) How might these climatic changes have affected the local ecology around Moche Borago? 3) If evidence exists at various s cales for climatic and environmental fluctuations during OIS 3, how did the mobility patterns, subsistence strategies, and social organization of hunter gatherers vary in response to paleoenvironmental instability? 4) If evidence exists at various scales for climatic and environmental fluctuations during OIS 3, how did hunter gatherer stone tool technology vary in response to paleoenvironmental fluctuations? I now attempt to provide succinct answers to these questions by drawing upon the discussions and analy tical results presented in the previous chapters.
213 What is the E vidence at G lobal, C ontinental, and L ocal S cales for C limatic F luctuations d uring E arly and M iddle OIS 3? How M ight t hese C hanges H ave A ffected L ocal E cology A round Moche Borago? Numerous m ulti proxy data sources, such as speleothems, marine cores, and Nile paleohydrology records, indicate that the African and SW Asian monsoons reacted similarly and synchronously to high latitude climatic changes above Greenland Nile paleohydrology records in particular, provide a direct link to monsoon driven climatic changes in the Ethiopian highlands (Revel et al., 2010) At Moche Borago, fluvial activity peaks ~43 ka, wh ich is contemporaneous to Dansgaard Oeschger 11, when monsoons are expected to be more intense. The catalyst behind monsoonal flux in the Horn appears to be related to broader climatic changes across the Northern Hemisphere at this time. Beginning in OI S 4, ~75 ka, Northern H emisphere climates became increasingly unstable (Sanchez Gone and Harrison, 2010; Wolff et al., 2010; Hessler et al., 2010) Heinrich events may have slowed or shut down the North Atlant ic thermohaline circulation, which regulates oceanic temperatures and, by proxy, terrestrial temperatures and precipitation, creating cooler and more arid conditions (Hemming, 2004) Cooler sea surface temperatures would have reduced the land sea thermal gradient and s hifted the location of the Inter Tropical Convergence Zone, minimizing the intensity of monsoonal winds across Africa and SW Asia (Vidal and Arz, 2004; Tierney et al., 2008) In contrast, the rapid warming ass ociated with Dansgaard Oeschger (D O) events would have increased Northern Hemisphere insolation which intensified the land sea thermal gradient and strengthened the monsoons from West Africa to Western Asia (W ang et al., 2001; Burns et al., 2003; Burns et al., 2004; Cai et al., 2006; Cosford et al., 2008; Revel et al., 2010) Numerous multi proxy records show contemporaneous African and SW Asian monsoonal variability, which is in sync with high latitude temp erature anomalies above Greenland, such as D O events. Marine core records from the Gulf of Guinea show African monsoonal flux during
214 D O events (Weldeab et al., 2007) Unpublished new data from the same location demonstrate that the monsoons are tightly linked to high latitude temperature anomalies during OIS 3 (S. Weldeab per comm.). Sapropel events seen in the Soreq cave speleothem records reveal similar variations in the Africa n monsoons across the eastern Mediterranean region (Bar Matthews et al., 2000) In the Indian Ocean, the Moomi cave record from Socotra Island, as well as Arabian Sea marine cores, provides detailed accounts of SW Asia n monsoonal flux during D O events (Schulz et al., 1998; Burns et al., 2003; Burns et al., 2004) The highlands of Ethiopia, and in particular the SW Highlands, are uniquely positioned to receive both West Af rican and SW Asian monsoon al precipitation, and this area would have experienced incre ased monsoonal precipitation during D O events. Nile paleohydrology records are linked to climatic fluctuations affecting the source waters in the Ethiopian highlands, a nd these records show increases in monsoonal activity in the Ethiopian highlands during OIS 3 (Revel et al., 2010) The orographic effects of the highlands on monsoonal air masses (Mohammed et al., 2004) may have even increased rainfall in SW Ethiopia, as those air masses similarly do today. At Moche Borago, S Group or similar deposits in each of the three excavation areas show clear geomorphic evidence that the local area was we tter ~43 ka. In the BXA, the LVDBS deposits of the S Group represent channel fill deposits of a fluvial feature that cut into tephra YBT. Sections of tephra YBT were even missing in the BXA, which is likely due to cut and fill activity from the BXA chann el. In TU2, fluvial activity truncated and undercut the TU2 L Group. In N42, size sorted channel fill deposits unconformably overlie t he L Group deposit (ULF) in that area.
215 Stratigraphic correlations and radiometric dates from the BXA and TU2 showed tha t the S Group fluvial features in those two excavation areas were contemporaneous. Geochemical and stratigraphic correlations link these features to fluvial features also found in N42. The YBT (underlying) and YBS (overlying) tephras bracket the S Group d eposits in the BXA and TU2, providing the maximum and minimum weighted mean age estimate of the fluvial activity: 43,403 1,213 cal. BP to 43,121 692 cal. BP respectively. The weighted mean age of radiocarbon samples on charcoal taken directly from th e BXA and TU2 S Group deposits fit the YBT YBS chronology, 43,480 443 cal. BP Therefore, from ~ 44 ka to ~ 43 ka, a series of high and low energy fluvial features were present at Moche Borago The source of the water was likely due to percolation from joints within the rear rockshelter walls A similar phenomenon is seen at Esay rockshelter today, which has an active stream channel in that site The morphology of the channel features and condition of associated stone artifacts suggest that the energy of the stream flows on site were lower around the BXA and greatest at N42. Size sorting or abrasion is not indicated on the stone artifacts in the BXA, and these stone artifacts are currently believed to be in situ The presence of the freshwater at this time may have been a major incentive for occupation of the site. The timing of the fluvial activity at Moche Borago coincides closely with wetter and warmer climatic conditions that would have occurred during D O 11, 43.4 ka. The D O 11 event is eviden t in both high latitude ice cores and regional records surrounding the Horn of Africa. Above Greenland, D O 11 is associated with the greatest increase in mean annual temperatures known during OIS 3, 1 5 C (Wolff et al., 2010) Across Africa, monsoonal activity may also have increased at this time. West African Atlantic marine core records indicate that African monsoonal precipitation was greater at 43 ka just p rior to a dry event at 42.5 ka, though the
216 changes are subtle (S. Weldeab per comm.). The Moomi cave records show a similar, subtle pattern of monsoonal increase ~43 ka (Burns et al., 2003; Burns et al., 2004) Nile paleohydrology records also indicate a broad period of pluvial conditions at this time, which show wetter conditions in the E thiopian highlands (Revel et al., 2010) In summary, a chain of evidence suggests that high latitude climatic changes affected low latitude climates and environments. In the Horn of Africa, climatic changes impacted the monsoon systems, which account for much of the annual moisture in the area. At Moche Borago, on site fluvial activity peaked ~43 ka. The timing of this fluvial activity is concurrent to the well dated an d well studied D O 11 event, which had the second hig hest temperature flux during OIS 4 and OIS 3 (+15C) (Wolff et al., 2010) As with other D O events during this time, D O 11 probably had a significant impact on monsoon intensity across the Horn and I currently believe that the increased fluvial activity at Moche Borago ~43 ka is linked to the D O 11 event. If E vidence E xists at V arious S cales for C limatic a nd E nvironmental F luctuations d uring OIS 3, H ow D id the M obility P atterns, S ubsistence S trategies, and S ocial O rganization of H unter G atherers V ary in R esponse to P aleoenvironmental I nstability? The frequency and type of stone artifacts represented in the Moche Borago assemblages change from the T Group to the S Group litho stratigraphic units. The T Group assemblage is composed largely of debitage, and more kinds of raw materials may have been collected and used during the T Group than in the S Group. T hese raw materials may indicate that hunter gatherers were traveling farther across the landscape, collecting different raw materials, and staying at the site for only brief periods of time. In contrast, the S Group assemblage is almost completely made fr om locally available common black obsidian (CBO). The S Group has more stone artifacts and more diverse types of stone artifacts, which might be due to longer occupation time in the rockshelter. The lack of fragmentary debitage in the S Group may indicate residential activities like waste management.
217 Today, Moche Borago is located within the Afromontane woodlands on the volcano base of the volcano and Southern Rift Valley, as well as lacustrine resources from nearby Lake Abaya and other lowland resources in the Omo gorge. However, lake core and pollen records from across eastern Africa indicate that the montane woodlands would have moved to higher elevations d uring arid phases of the Terminal Pleistocene and Holocene due in part to decreased monsoonal precipitation at lower elevations (Maitima, 1991; Hildebrand, 2003; Kiage and Liu, 2006) Holocene lake level recor ds from the Ziway Shala basin north of Lake Abaya also reveal that lake levels across the region would have fallen during arid phases as a result of decreased monsoonal rains (Gillespie et al., 1983; Lamb et al. 2000; Chalie and Gasse, 2002; Benvenuti et al., 2002) The distances separating each ecological zone would have increased during arid phases of weakened monsoonal precipitation. Therefore, during arid periods, it may have been more advantageous for hu nter gatherers to live in lowland woody grasslands below Mt. Damota, which may have been more centrally located to other ecological zones and resources in the area. Precipitated freshwater probably still collected atop the mountain and drained into the su rrounding valleys during arid periods similarly to drought conditions in the region today (Borena, 2008) The stone artifact assemblages from the T Group may date to one such period of aridity though more precise dating and geomorphological work still need s to be done. Current evidence suggests that the T Group Upper deposits were laid down in a humid environment (P. Goldberg, per comm.), and I think it is likely that the T Group Upper deposits may date to D O 12, which began 46.8 (Svensson et al., 2008; Wolff et al., 2010) But, the absence of fluvial channels
218 throughout the T Group deposits suggests that the climatic conditions were unlikely to be as wet as the S Group, even if some of the deposits do date to D O 12. The T Group stone artifacts appear to have accumulated at the site via repeated brief trips where only limited activities were undertaken. I f the site was only briefly occupied throughout the T Group, due to marginalization or no in situ freshw ater supply, then the range of on site activities represented in the stone artifacts might be smaller than in the S Group when extended occupation was more likely. The T Group assemblages are dominated by debitage (> 85%) and there are 130% fewer cores an d 200% fewer shaped and unshaped tools in the T Group compared to the S Group. Furthermore, there are only seven scrapers and four burins in the T Group compared to 47 scrapers and 8 burins in the S Group. The most diverse kinds of raw materials are also found in the T Group (and especially T Group Lower), which includes multiple kinds of obsidian, chalcedony, basalt, quartzite, and chert. The sources of these raw materials remain speculative, however, but the volcanic geology of the immediate area makes it highly unlikely that the chalcedonies and cherts were available locally. T he relatively sparse amount of cores, shaped tools, and unshaped tools in the T Group suggests 1) there were relatively fewer activities on site than in the S Group, due to the lack of burins, scraper, or other specialized tools and 2) that if shaped and unshaped tools or cores were brought into the site during the T Group then they were likely modified and taken away This pattern of curation would also help to explain the la rge amount of T Group debitage: Stone tools were being repaired or modified on site and then taken away. The more diverse raw material types in the T Group may indicate that more raw material sources were encountered and used at this time as hunter gather ers were foraging for resources farther across the landscape due to increased separation between ecological zones (cf. Ambrose, 2002)
219 In contrast, during time periods when monsoonal intensity increased, hunter gatherer mobility around Moche Borago may have decreased. More resources may have been available locally, especially if the montane woodlands moved toward lower elevations. Mo che Borago may have been located between highland woodlands and lowland grassy woodlands, with supplemental resources coming from Lake Abaya and the Omo Gorge. Another source of in situ freshwater, which is evident in the S Group geomorphology, would have been an added attraction to occupy the site for longer periods of time. During the S Group, Moche Borago may have offered an ideal balance between protection from the elements, convenience to resources (including locally available obsidian), and ecotone resource diversity and abundance. Most of the stone artifacts in the S Group appear to be made of CBO, assumed to be locally available. Two sources of CBO are known within 15 km of the site. The large amounts of shaped tools, unshaped tools, and cores indicate that frequent activities were performed on site during the S Group. Tools such as burins and drills may have been used to work wood or bone for hunting or foraging implements such as spear shafts or digging sticks. Scrapers show that animal hides were processed on site. Numerous points suggest that at this time hunting forays were also based out of the site. In addition, evidence shows that the people living at the site at this time managed the site by possibly removing waste by products of sto ne tools production. The size of the S Group assemblage is 50% smaller than the T Group Upper or Lower. Hypothetically, more intensive occupation of a site should generate more stone tools debris, but less debris might indicate that the site was also act ively maintained. It is unrealistic to assume that behaviorally modern humans their families and their children living in Moche Borago wanted to walk on razor sharp stone debris. The rockshelter was their home. Therefore, one possible explanation wort h further
220 study, is that the smaller S Group assemblage may indicate that occupants at the site were collecting and redistributing lithic debitage outside of the cave. Further evidence for site management may be found in the lack of S Group fragmentary de bitage. More than three fourths of the S Group debitage is comprised of whole flakes, which may have been produced to make shaped and unshaped tools. No evidence suggests that the lack of fragmentary debitage is due to fluvial sorting or other natural pr ocesses. Therefore, while whole flakes may have been collected for shaped and unshaped tools, the fragmentary debitage could have been intentionally discarded. In summary, the Moche Borago stone artifact assemblage shows patterns that give evidence to h unter gatherer site use, mobility, and subsistence strategy. The debitage dominated T Group assemblage contrasts with the abundant shaped tools, unshaped tools, and cores in the S Group. The emerging picture of hunter gatherer life ways may be due to var ying amounts of aridification and separation between ecological zones and resources. The site may have been marginally located to numerous ecological zones during arid periods, and hunter gatherers infrequently relied on the rockshelter at these times, pe rhaps as they ranged far and wide across the landscape for resources. Wetter times meant that woodlands moved into lower elevations. The Rift Valley lakes probably also expanded, and the site may have been centrally located to numerous resources. An on site source of freshwater may have lured people to this location. S Group stone artifacts attest to a wide range of on site activities at this time, which relate to foraging and hunting activities, as well as domestic activities such as hide scraping. If E vidence E xists at V arious S cales for C limatic and E nvironmental F luctuations d uring OIS 3, H ow D id H unter G atherer S tone T ool T echnology V ary in R esponse to P aleoenvironmental F luctuations? Hunter gatherers living in and around Moche Borago may have altered their way of life and their technologies in accordance with suspected ecological changes caused by monsoonal
221 flux and broader climatic instability. In addition to suspected changes in subsistence economy and mobility, these hunter gatherers began relying on Mode 4/5 stone tool technology that had similarities to Mode 3 techniques and knowledge. The timing of the newer Mode 4/5 technology at Moche Borago and probably at other sites, such as Enkapune Ya Muto in Kenya and Mumba rockshelter in Tanzani a, coincides with unstable climatic conditions during OIS 3. These erratic conditions would have affected monsoon systems across the region and the distribution of ecological zones and natural resources. The switch to Mode 4/5 technology was an adaptatio n to the social and economic requirements of the OIS 3 hunter gatherers as climates and environments fluctuated across the region and over time. Modern hunter gatherers living in unpredictable environments such as deserts employ flexible behavior strateg ies that enable them to adapt to, and exploit, numerous available resources in varying ecological zones (Yellen, 1977) During OIS 3, climatic instability across the Horn may have created environmental conditions where resource availability and location were similarly unknowable and unpredictable. More abundant or diverse resour ces during wetter periods of OIS 3 may have made collecting resources from known sources easier. But, the variable conditions would have required repeated re learning of the natural environments, thus finding new resources would have cost valuable time an d energy (Kelly, 1995) The development at this time of new tool types (e.g. microliths) for new technologies (e.g. the bow and arrow) with li ghtweight, portable, and multi component designs would have allowed hunter gatherers to range more freely across the landscape to find and utilize diverse resources. E vidence suggests that certain technological developments, which are found more frequentl y in the S Group, begin in the T Group Upper 37 when several new types of stone 37 Analysis is ongoing to assess culture stratigraphic divisions within the assemblage, but in this dissertation I relied strictly on litho stratigraphic breaks to avoi d confusion.
222 artifacts first appear, including Mode 4/5 Type I Irra Xoketta points and microliths But Mode 3 L evallois cores and points are also still present. Both Mode 3 and Mode 4/5 sto ne artifacts in the T Group Upper suggest that hunter gatherers were adjusting their stone tool technologies at this time, but still retaining tried and true methods. Technological development toward smaller and multi component stone artifacts, such as mi croliths, suggests that hunter gatherers at this time desired less wieldy tools that could be repaired and applied flexibly and to various applications. If the T Group Upper dates more precisely to D O 12, as I have speculated, then it means that hunter g atherers may have begun developing new technologies as early as the T Group Upper to contend with variable and unpredictable climatic and environmental conditions associated with monsoonal flux during OIS 3. The mobility patterns of these people during th e T Group Upper were still fairly mobile, however, perhaps due to long distances between resources. While new tool types appear in the T Group Upper, the prior Mode 3 technology does not disappear completely, and its sustained presence in the S Group ind icates that the technological innovation of Mode 4/5 tools was a prolonged and complex process. Several tools (e.g. BN 1818.1 and 1818.6) show that Levallois techniques were used to make Mode 4/5 tools in the S Group, and Irra Xoketta points may also deri ve aspects of their unique morphology from Mode 3 technology. The mixing of techniques implies that hunter gatherers had not completely abandoned the older knowledge base (Mode 3), but were incorporating it into the new system, possibly i n innovative ways and contexts. Ongoing technological innovation might also account for the increase in less formal single platform core designs within the assemblage. Single platform cores become the dominant core type in the T Group Upper, and this pattern is similarl y seen in the S Group. The single platform cores are made casually, and they appear to be based on informal designs, which is unlike
223 pyramidal or prismatic core types commonly seen in later periods. It is possible that old technological traditions were b eing revised or abandoned at this time, while new technological traditions were still in formative stages. Therefore, the informality of single platform core designs may be because technologically or culturally specific formal core designs had not yet bee n settled upon at this time. In summary, ~45 ka, T Group hunter gatherers at Moche Borago began to develop and use new types of Mode 4/5 stone artifacts, such as microliths and Irra Xoketta points. Warmer climates due to D O 12 may have brought wetter c onditions to the area, though these conditions were still likely drier than during the later periods represented by S Group deposits. The development of Mode 4/5 technologies at this time could have been a coping mechanism to the destabilized environments and unpredictable resources brought about by the monsoonal fluctuations. The designs of the new tools may have prioritized efficiency and portability, which would have been essential if the locations and types of resources in an area were changing. Indi vidual pieces used in multi component tools could be repaired separately and the use of microlithic technology was lightweight. By ~43 ka, added monsoonal moisture may have increased resource diversity around Moche Borago, especially if the montane woodl ands moved toward lower elevations and the distances to other resources, like Lake Abaya, shortened. C oupled with an on site freshwater supply, and Moche Borago may have been an ideal residential base to utilize highland woodland, lowland grassland, and n umerous other resources. The more intensive occupation of the site at this time (S Group) gives a more in depth glimpse into hunter gatherer life ways and behaviors, and stone tools suggest on site activities related to hunting, foraging, and domestic tas ks.
224 But the environment would have still been in a state of flux due to increased monsoonal rains and the locations and types of resources may have been largely unpredictable. Stone artifacts from this time suggest that hunter gatherers continued to uti lize Mode 3 technological knowledge and skills along with the newer Mode 4/5 system. Even though more diverse resources may have been available at this time, which likely affected mobility patterns, the long term unpredictability of the resources and their locations may have sustained technological innovation. The continued innovation of Mode 4/5 tools into the S Group, therefore, suggests that the development of Mode 4/5 tools, like microliths, at this time was not a direct product of increased mobility, but rather a by product that enabled hunter gatherers to become more mobile, dependent on the circumstances. Thus I believe that the development and continued innovation of Mode 4/5 technology found at Moche Borago shows that these people were adapting p rimarily to the variable environments and unpredictable resources in early OIS 3. Moche Borago in Broader Perspective Three variables place Moche Borago into its regional context: time period, environment, and culture. The timing of the earliest Mode 4/ 5 technology across eastern Africa is perhaps the most contentious of the three variables due to dating issues at other sites like Mumba (Mehlman, 1989; Prendergast et al., 2007) and Enkapune Ya Muto (Ambrose, 1998a) rockshelters,. However, the ages of the Mode 4/5 deposits from these sites suggests to me that OIS 3 is emerging as the consensus for this region (see Chapter 4). Last Glacial environments and climates may have heavily influenced hunte r gatherers living in both highland and lowland environments of the Horn and East Africa. Ambrose (2002) for example, has hypothesized th at unpredictable environments in savanna environments of East Africa during the Last Glacial required hunter gatherers to develop more mobile subsistence strategies and tool kits and also exchange systems. However, lowland climates in East Africa
225 may have been even drier than highland Ethiopia at that time, and detailed paleoenvironmental records for the Horn and East Africa are generally lacking. Therefore, comparisons between highland and lowland sites must be made cautiously. Throughout this disserta tion, I have argued that the development of Mode 4/5 technology occurred during early OIS 3 at Moche Borago. While I am confident in my conclusions, I am also aware of the limitations of my data, which is only an early OIS 3 stone artifact assemblage. Cu rrently, no earlier in situ archaeological deposits are known at Moche Borago. It is equally plausible that Mode 4/5 technology could have developed earlier during OIS 4 in response to monsoonally driven climatic fluctuations and environment instability, if my theory is correct. D O events occurred in OIS 4 as early as ~76 ka (D O 20) (Wolff et al., 2010) D O 19, dated ~72.3 ka, was associated with the biggest temperature anomaly of the Late Pleistocene above Greenland (+16C) (Landais et al., 2004) and it could have had a significant impact on monsoonal patterns across the Horn and East Africa. But my main point is that whether or not Mode 4/5 was developed during OIS 4 or OIS 3, the key factor may have been that hunter gatherers were adapting more toward regional cl imatic instability and resource unpredictability rather than to a broad period of aridification, resource depletion, and ecological niche narrowing. The idea of regional cultural traditions in the Horn and East Africa during the Late Pleistocene can be tr aced back to Leakey (1931; 1936) and Clark (1954; 1988) who maintained that artifact morphology and frequency, especially in points, reflected distinctive styles that were expressions of cultural var iability. The Type 1 and 2 points from Moche Borago, which in the and culturally distinct.
226 Local evidence for regional cultural traditions around Moche Borago m ay be seen in stone artifacts predating and postdating the YBT eruption at the site. Similarities between point typologies before and after YBT show that the reoccupation of Moche Borago after YBT was by hunter gatherers who shared a culture with prior oc cupants at the site. Irra Xoketta, Tona, and Toora points each occur in the T Group Upper and the S Group. Microliths and single platforms cores, which occur for the first time in the T Group Upper, are also similarly seen in the S Group. If our estimat es of the YBT eruption are accurate, then YBT may have been on par with the Mount St. Helens eruption of 1980 in that it took over a decade for plant and animal communities around Mount St. Helens to make a comeback (del Moral and Lacher, 2005) YBT may have had a similar effect on the area surrounding Mt. Damota. The lack of artifacts from within the YBT ash at Moche Borago suggests that the eruption was very likely devastating, destroying local flora and fauna resources and forcing people o ut of the area. Entire hunter gatherer groups may have even been killed. Therefore, the similarity of stone artifact styles from layers, which were dated before and after the YBT eruption, may indicate that these hunter gatherers were part of a broader r egional cultural tradition, or at least a tradition that surpassed the local influence of the YBT eruption. Larger stone artifact assemblages from Moche Borago and coeval assemblages from other sites in the area are expected to provide further useful info rmation to test this hypothesis in the future. As a topic of future study, the Irra Xoketta points and associated artifacts from Moche Borago should be compared in more depth to the assemblages from other local and regional sites. For example, based on a r ather subjective comparison of shape, size, and design attributes, I find that Irra Xoketta points are very similar to points found in central Ethiopia and Somalia formerly (Clark 1954) At Hargeisa in Somalia, Clark (1954:193)
227 believed that Somaliland Stillbay points derived from earlier Levallois based techniques, and the points were frequently made o n end struck flakes, exhibited bulbar thinning, and had sub triangular to foliate shapes with rounded butts characteristics that are common to Irra Xoketta uni facial to fully bifacial points, 38 most of which are made from the Levallois method, are association with backed tools, (Kurashina, 1978) A similar assemblage is found at Porc Epic in east 39 (Clark et al., 1984) age and typology (Pleurdeau, 2001) These points are also found with the earliest microliths at Porc Epic, which parallels the situation at Moche Borag Conclusions This dissertation has provided the first detailed description of a well dated and continuous early OIS 3 archaeological assemblage in the Horn of Africa. The stone artifact assemblages from Moche Borago dating to early OIS 3 s how changes that may reflect alterations in hunter gatherer subsistence economy, mobility, site use, and technology. To explain the changes in flaked stone artifacts, I have explored the possibility that hunter gatherers were adapting to local environment al fluctuations, which may relate to broader climatic and monsoonal variability during OIS 3. There has been very little research in general directed at understanding hunter gatherer behavioral adaptations to OIS 3 paleoclimates in Africa. Consequently, many of the ideas 38 instead. 39 Referred to as Ensemble III in Pleurdeau (2001).
228 proposed in this dissertation remain untested, and some untestable, due to the lack of available data. Much more research is required in order to answer such basic questions as: What is the current actual ratio of precipitation deliver ed from the African and SW Asian Monsoons in SW Ethiopia? Did this ratio ever vary in the past? Were there latitudinal gradients in the Horn affecting the contributions from the African and SW Asian monsoons (Kebede et al., 2009) ? Did the monsoon systems vary similarly and synchronously to broader climatic changes or were there other factors which influenced monsoonal intensity du ring OIS 3? Were the orographic effects of the Ethiopian highlands that today act upon monsoonal moisture similar in the past? Furthermore, how did highland ecological zones in SW Ethiopia change during OIS 3? (2003) study provides useful hypotheses about Holocene paleoenvironmental changes in the western highlands of SW Ethiopia. To date, however, there are no archaeological studies from SW Ethiopia, or for that matter anywhere in the Horn, which describe environme ntal fluctuations during OIS 4, OIS 3, or OIS 2. And, what few lake cores and pollen assemblages that are available, they are poorly dated beyond 40 ka. How did hunter gatherers living in the lowland areas of the Horn adapt to climatic flux during OIS 3? If these areas were already drier than highland areas, how did fluctuations in monsoonal rains affect lowland ecology? My interpretation of the flaked stone artifacts from the T Group suggests that while hunter gatherers developed new technologies to p erform activities required of unstable climates and variable resources, aridity and distance between resources may be a primary factor driving mobility patterns. Therefore, if lowland areas were generally drier
229 than the highlands during OIS 3, then hunter gatherers in these areas may have continued to rely on a high degree of mobility unlike their S Group counterparts in the highlands. Many of these questions should be resolvable with continued research. At a local level, the SWEAP team is continuing to w ork at Moche Borago and surrounding areas on a variety of related issues. During the recent 2010 field season, SWEAP researchers visited several other open air and rockshelter sites in the area, which can be compared to the archaeology of Moche Borago. N ew radiocarbon samples have also been submitted, which we hope will improve the current dating of the site and area. We are also exploring the use of other dating techniques like luminescence in volcanic sediments. Lithic microwear and faunal analyses wi ll also provide cogent data on the function of stone artifacts, the local wildlife during early OIS 3, and hunting strategies. On a regional level, new lake cores from Ethiopia and sediment cores from the Indian Ocean and Gulf of Guinea (Atlantic Ocean) m ay provide new data about paleoenvironmental and paleoclimatic changes during the Late Pleistocene, including monsoonal fluctuations. Research at additional archaeological sites across the Horn, like at Aduma (Middle Awash, Ethiopia) or around the Bulbula / Ziway area in central Ethiopia could yield comparative data to Moche Borago about lowland hunter gatherer adaptations during the Last Glacial. Above all, I hope that this dissertation provides useful information to other researchers studying Late Plei stocene hunter gatherer behavioral adaptations in Eastern Africa and at continental and global scales.
230 APPENDIX A LIST OF ABBREVIATION S AND DEFINITIONS US ED IN THE TEXT 14C Carbon 14 used for radiocarbon dating. AMH Anatomically Modern Humans AMS Accel terator Mass Spectrometry BP Before Present (i.e. 1950) BXA Block Excavation Area at Moche Borago D O Dansgaard Oeschger event (i.e. D O 12) East Africa Kenya, Tanzania, Uganda, Rwanda, Burundi GEPCA Groupe d'tude de la Protohistoire dans la Corne de l'Afrique (English: Protohistoric study group in the Horn of Africa) Horn of Africa Ethiopia, Eritrea, Somalia, Djibouti ITCZ InterTropcial Convergence Zone ka kilo anum Last Glacial Geological period lasting from 73.5 ka to 14.7 ka LGM Last Glacial Maximum (27.2 23.5 ka) LSA Late Stone Age MNL Minimum Number of Lithics MSA Middle Stone Age mtDNA Mitchrondrial DNA North Africa Morocco, Algeria, Tunisia, Libya, Egypt, Sudan, Chad OIS Oxygen Isotope Stage PE PC Paleoenvironment and Paleoclimate SST Sea Surface Temperature
231 SWAP Sodo Wolayta Archaeological Project SWEAP SouthWest Ethiopia Archaeological Project
232 APPENDIX B LIST OF ARCHAELOGICA L SITES FROM THE HOR N OF AFRICA, EAST AF RICA, AND NORTH AFRI CA, DATING TO THE LAST GLACIAL PER IOD W HICH HAVE DEPOSITS C ONTAINING MODE 3 AND MODE 4/5 LITHICS Country Site Name Locality Total Depth of Deposits Stratigraphic Context Age Calibrated Age Depth (cm) / Stratigraphic Association Kenya Lukenya Hill GvJm62 Area 2: 4.4 meters Alluvial fan deposit s adjacent to Lukenya Hill 21,535 980 25924 1294 220 230 GvJm46 ~30,000 Prolonged Drift Weakly developed paleosols stratigraphically above sand and silt deposits from aggrading stream channel. Deposit capped by disconformable Makali a ash, dated ~30 ka >40,000 Prospect Farm Locality 1 4 ~15 meters Interstratified volcanic ash layers 21,800 32,500 46,500 53,100 ~50,000 Kisese II Upper Mode 5 assemblage (Levels 1 12) 18,190 300 21852 440 Spit 9 at base of microlithic rich deposits Lower Mode 5 assemblage (Levels 13 36) 31,480 1,350 36235 1619 Level 14
233 Country Site Name Locality Total Depth of Deposits Stratigraphic Context Age Calibrated Age Depth (cm) / Stratigraphic Association Kenya Enkapune Ya Muto 5.54 meters Loam 16,300 1000 DBL1.2 29,300 750 DBL1.2 35,800 550 DBL1.2 37,000 1,100 DBL1.3 39,900 1,600 DBL1.3 >26,000 RBL4.1 41,400 700 RBL4.1 29,280 540 R BL4.1 + 4.2 35348 2183 DBL1.3 46406 2758 GG1.3 32,548 1247 RBL4.2 Tanzania Mumba >10 meters Stratified sand and loam deposits 33,460 900 a Level IV 20,995 680 to 65,680+6049 / 5426 a Level V Naisusu Beds Olduv ai Gorge Aeolian tuffs 17,550 1,000 21048 1206 Aeolian tuffs 42,000 1,000 Ethiopia Porc Epic ~2.5 meters cemented breccia capped by unconformable Holocene dripstone 61,202 958 60 200 cm; ensemble 3 / 4a 61,640 1,083 77,565 1,575 K'one G4, Area A ~10 meters Loam 14,670 200 Upper Loam Bulbula River Not Excavated Stratified paleosols 27,050 1,540 paleosol Liben Bore >2 meters Silty clay UNDATED
234 Country Site Name Locality Total Depth of Depos its Stratigraphic Context Age Calibrated Age Depth (cm) / Stratigraphic Association Somalia Midhishi 2 0.7 meters 7 lithostratigraphic units >40,000 uppermost Mode 3 bearing unit 18,790 340 Deposits overlying Mode 3 Gud Gud 2 meters 1.4 m eters of archaeologically sterile deposits overlying thin occupation horizon >40,000 Gogoshiis Qabe ~1.8 meters UNDATED
235 Country Site Name Locality Technique / Material Lithic Assemblage Comments References Kenya Lukenya Hill GvJm62 14C on apatite Mode 5 Kusimba (2001) cautions 14C ages are inaccurate because of modern carbon contamination. Ambrose (1988) suggests GvJm62 and GvJm46 date to ~39,000 ka. Kusimba 2001 GvJm46 14C on apatite Mode 5 Prolonged Drift Mode 3 Artifact densit y very low estimated only 50 lithics / 0.01 cubic meters Merrick 1975; Ambrose 2001 Prospect Farm Locality 1 4 Obsidian hydration Mode 5 Michels et al. 1983; Ambrose 2001 Mode 3 Mode 3 Kisese II 14C on OES Mode 5 More abundant micro liths than lower deposit Inskeep 1962; Barut 1997 14C on OES Mode 5 Microliths rare. Assemblage similar to Nasera and Mumba rockshelters were early Mode 5 is scraper dominated with few backed pieces
236 Country Site Name Locality Technique / Material Lithic Assemblage Comments References Kenya Enkapune Ya Muto 14C on charcoal Mode 5 / Sakutiek Age reported for the earliest Mode 5 at EYM is >46 ka but this is based is an estimate based on sediment deposition rate using obsidian hydration ages. Ambr ose 1998, 2001 Mode 3 / Endingi Obsidian hydration Mode 5 / Sakutiek Mode 5 / Nasampolai Mode 3 / Endingi Tanzania Mumba 14C on tufa and achatina shell Archaeologically sterile Mehlman 1989; Prendergast et al. 2007; Marks and Conard 2006 14C and Th 230 on bone apatite, OES, shell Mode 3 and Mode 5 Assemblage is heavily biased and existing analyses are based on a limited sample from the original Kohl Larson excavation. N aisusu Beds Olduvai Gorge 14C on OES Mode 3 and Mode 5 Dating inconsistency between radiocarbon and Argon/Argon Leakey et al. 1972; Barut 1997 40Ar / 39Ar on biotite Manega 1993
237 Country Site Name Locality Technique / Material Lithic Assemblage C omments References Ethiopia Porc Epic Obsidian hydration Mode 3 and Mode 5 Context and dating are problematic and there is a large disparity between radiocarbon and obsidian hydration age estimates. Some of the MSA deposits are also eroded by fluvial a ctivity. Assefa 2002; Pleurdeau 2001; Clark et al. 1984; Michels and Marean 1984) K'one G4, Area A 14C Mode 3 and Mode 5 11 meter sequence and only the topmost section of the stratigraphically highest loam deposit has been dated. Lithics in the upper Loam include backed Levallois flakes but the deposits may be secondary context. Kurashina 1978; Clark and Williams 1978; Brandt 1986, Brandt and Brook 1984
238 Country Site Name Locality Technique / Material Lithic Assemblage Comments Refere nces Ethiopia Bulbula River 14C Mode 5 Site is unexcavated. Lithics include non geometric microliths and blades made on obsidian. Gasse and Street 1978; Gasse et al. 1980; Brandt 1986 Liben Bore Mode 3 with increasing abundance of Mode 5 higher i n the sequence Deposits appear to span the Mode 3 to Mode 4/5 transition, however, the site is currently undated. Brandt 2000; Kinahan 2004; Fisher 2005; Negash 2004 Somalia Midhishi 2 14C on charcoal Mode 3 Brandt and Brook 1984; Brandt 1986; Gresha m 1984 Mode 3 and Mode 5 Gud Gud 14C on charcoal Likely not Mode 3 Assemblage is less than 50 pieces. Brandt 1986; Brandt and Brook 1984 Gogoshiis Qabe Mode 3 and Mode 5 Mode 3 deposits are stratified under terminal Pleistocene (Eibian) an d Holocene (Bardaale) assemblages, but the Mode 3 deposits themselves are undated Brandt 1986
239 Country Site Name Locality Total Depth of Deposits Stratigraphic Context Age Calibrated Age Depth (cm) / Stratigraphic Association Egypt Bir Sahara BS 11 a nd BS 15 Upper Lake: basal unit of sandy silt 30,870 1,000 >44,700 Bir Tarfawi white limestone unit 32,780 900 40,710 3,270 >41,450 middle of grey silt 44,190 1,380 Taramsa Hill Aeolian sands 55,500 3.7 aeolian deposits surrounding location of Taramsa I skeleton Sector 89/03 38,100 1,400 Nazlet Safaha 37,200 1,300 Nazlet Khater .Nazlet Khater I Gravel and wadi deposits UNDATED Nazlet Khater 2 Nazlet Khater 3 Nazlet Khater 4 Sand and gravel deposits 31,320 2,310 30,980 2,850 30,360 2,310 33,280 1,280
240 Country Site Name Locality Total Depth of Deposits Stratigraphic Context Age Calibrated Age Depth (cm) / Strati graphic Association Egypt Sodmein >4 meters 25,200 500 MP1 / Top of Layer D >30,000 MP2 >45,000 MP3 / Interface between Layers F and F >44,500 MP4 118,000 8,000 MP5 Libya Jebel Gharbi Wadi Ghan Three main units (from bottom to top): 7 12 m alluvial gravel capped by calcrete; 3 6 m of red aeolian sand grading into loess; 3 4 m light pink aeolian sand grading into loess UNDATED Ain Zargha Colluvial silt overlain by aeolian sands. Sequence topped by two more l ayers of colluvial silts with interstratified paleosols Lower paleosol within the upper colluvial silt layer
241 Country Site Name Locality Total Depth of Deposits Stratigraphic Context Age Calibrated Age Depth (cm) / Stratigraphic Association Libya Jebel Gharbi Ain Shakshuk 3 main stratigraphic units representing aeolian sediments deflated from nearby wadi 27,310 320 Upper Deposit 24740 140 30,870 200 25500 400 44,600 2,430 Charcoal bearing layer in mid dle of sequence 43,450 2,110 Lowest Deposit Haua Fteah 13 meters 28,500 800 33,100 400 >35,950 43,400 1,300 47,000 3,200
242 Country Site Name Locality Total Depth of Deposits Stratigraphic Context Age Calibrated Age Depth (cm) / Stratigraphic Association Morroco Mugharet el 'Aliya eolian sands 56,000 5,000 t Layer 6 47,000 5,000 Layer 6 51,000 5,000 Layer 9 81,000 9,000 Layer 10 Grotte de Pigeons (Taforalt) ~1 0 meters 20,200 37,400 Phase C 40,600 52,300 b Phase 03 Top 67,500 79,200 b Phase 03 Middle 46,600 88,900 b Sequence 05 Upper 77,400 112,900 b Phase Group F Dar es soltan DeS 1 7.5 meters ~49,000 57,000c G4 ~62,000 140,000c G3 ~72,000 100,000c G2
243 Country Site Name Locality Technique / Material Lithic Assemblage Comments References Egypt Bir Sahara BS 11 and BS 15 Mode 3 / Aterian several deposits at the site are deflated Wendorf 1977 Bir Tarfawi 14C on Melanoides Mode 3 / Mousterian several deposits at the site are deflated and some show dessication cracks Mode 3 / Aterian Taramsa Hill Weighted mean estimate of multi aliquot OSL ages Mode 3 char acterized by Nubian Levallois technique Vermeersch et al. 1998 find the "Late Middle Paleolithic" assemblage to be transitional to the systematic production of blades Vermeersch et al. 1998; Vermeersch et al. 1995 Sector 89/03 AMS on charcoal "Transitio nal" Mode 3 / Mode 4/5 Van Peer 1998 Nazlet Safaha "Transitional" Mode 3 L evallois Van Peer 1998 Nazlet Khater .Nazlet Khater I Mode 3 Nubian Style Levallois Vermeersch et al. 1982 Nazlet Khater 2 Mode 3 Nubian Style Levallois Nazlet Kh ater 3 Mode 3 Nubian Style Levallois Nazlet Khater 4 14C on charcoal Mode 3 overlaid by Mode 4 Shows persistence of Levallois technology into Upper Paleolithic in Egypt
244 Country Site Name Locality Technique / Material Lith ic Assemblage Comments References Egypt Sodmein 14C Mode 5 Abundant blades with few retouched tools Moeyersons et al. 2001; Mercier et al. 1999 14C Mode 3? / Possible Aterian 14C Mode 3 Disconformity between layers F and G characterized by th ick lens of ground based hymenoptera nests. Assemblage includes classic levallois and Nubian levallois as well as some truncated, facetted pieces 14C Mode 3 Assemblage similar to MP3 weighted mean estimate of thermoluminesence ages derived from flint blocks Mode 3 Assemblage predominantly Nubian Levallois
245 Country Site Name Locality Technique / Material Lithic Assemblage Comments References Libya Jebel Gharbi Wadi Ghan Alluvial gravel and calcrete contain Mode 3; lowest loess unit contai ns Aterian and Mode 5 lithics Garcea and Giraudo 2006; Barich and Garcea 2008 Ain Zargha 14C on charcoal Mode 5 and Aterian Undated Mode 3 lithics in lower colluvial silt layer. Ain Shakshuk 14C on charcoal Mode 5 14C on charcoal Mode 5 14C on charcoal 14C on charcoal Mode 3 / Aterian Haua Fteah Mode 5 Ages are considered underestimates. Low artifact density during Mode 3 to Mode 4/5 transition. Rapid appearance of Mode 4/5 has led to speculation the tech nology was not indigenous to the region. McBurney 1967; Moyer 2003 Mode 3 and Mode 5 Mode 3 Mode 3 Mode 3
246 Country Site Name Locality Technique / Material Lithic Assemblage Comments References Morroco Mugharet el 'Aliya E SR on tooth enamel Mode 3 / Aterian Wrinn and rink 2003 Mode 3 / Mousterian Low artifacts counts. Level deposits may be disturbed Low artifact counts. Instursive artifacts from higher deposits Grotte de Pigeons (Taforalt) OSL and 14C Mode 3 with Mode 4/5 near the top of the deposit Bouzouggar et al. 2007 OSL, TL, 14C Mode 3 Mode 3 / Aterian Mode 3 Dar es soltan DeS 1 Multi Grain OSL Mode 3 / Aterian Barton et al. 2009
247 APPENDIX C DETAILED DESCRIPTION S OF THE MICROLITHS AT MOCHE BORAGO Level 12 BN 1799.15 : Medial flake fragment with semi collateral backing on the right lateral. The flake blank has a parallel dorsal scar structure and the right lateral edge is modified lightly. B N 1799.42: Proximal flake fragment with straight backing on the right lateral edge. The flake blank has a dihedral platform with a parallel dorsal scar pattern. The left lateral edge is utilized. Level 13 BN 1802.2: Flake with obverse backing along the right lateral edge. Flake blank has a parallel dorsal scar pattern. BN 1802 97: Flake with obverse backing along the right lateral edge. The most intensive backing retouch is from the proximal edge to the length of the flake. The distal edge of the flake blank has been snapped off and the distal, right lateral edge may be burinated. Level 14 BN 1805 77: Microlith fragment with limited obverse backing on the left lateral edge. The right lateral edge is utilized. The planform of the piece is remi niscent of an obliquely truncated geometric microlith but there is no modification along the distal edge. BN 1805 73 : Flake with natural backing on the left lateral edge. There is use wear and modification on the right lateral edge opposite the backing. Distal and proximal sections of flake blank are snapped off creating a sub triangular plan form. BN 1805 75 : Flake with obverse backing on the left lateral edge. Similar in overall morphology to BN 1805 73. BN 1805 78: Flake with obverse backing on th e right lateral edge. Right lateral ventral surface has step fracturing indicating possible use wear. Level 16 BN 1818.5: Medial flake fragment with semi collateral retouching. The cross section of this piece is almost equilaterally triangular
248 and the dorsal surface is bifacial. The retouching begins along the medial ridge of the dorsal surface and extends to the left lateral edge. BN 1818.6: Levallois flake with backing on the left lateral. The flake has a radial dorsal scar pattern and a facetted platform. The backing is predominantly inverse except along a limited distance near the proximal edge of the left lateral edge where there appears to be several obverse oriented flake scars. This would indicate that the retouching is not due to core trim ming activities and was intentionally retouched along the lateral edge. Level 18: BN 2571.1: Crescent. There is an alternating backing pattern on the left lateral edge of the flake blank. The backing angle is 88 and the retouching forms a convex left lateral planform. There is a large flake scar near the proximal dorsal left lateral edge and two scars on the proximal ventral right lateral edge that may be use wear damage. There is utilization along the right lateral edge. Level 23: BN 3105.15 : Utili zed flake with curved, possibly crescentic, obverse / natural backing along the left lateral edge. The utilization is located on the right lateral edge extending from the proximal edge to around 2/3 the length of the tool. Micro facets along the ventral right lateral edge and two larger flake facets on the ventral right lateral near the distal edge may be use wear.
249 APPENDIX D DETAILED DESCRIPTION S OF THE POINTS AT M OCHE BORAGO BN 4889.1 This is a small, unifacial point with a triangular planform that is made on an end struck flake. The point may be broken at the base but this seems unlikely. It is possible also that the proximal edge of the flake bank was snapped off prior to retouching of the dorsal surface in a crude method of bulbar thinning. Thi s is supported by flake scars along the break which were likely produced after the break occurred. The retouching is mainly on the distal surface producing a circular dorsal scar pattern. The retouching is distinctly finer along the distal left lateral ed ge of the dorsal surface. The tip is broken and retouched, but there is similar fine retouching here also. There are shallow bi lateral notches half way up from the base of the tool. These are interpreted to be mainly caused by, or for, hafting purposes. G10 Level 30 T Group DCC BN 4854.1 BN4854.1 is a small semi ovate point that has been bifacially retouched. Retouching is fully invasive and there is no indication which surface was originally ventral or dorsal. Surface 1 pr imarily has a unidirectional flake scar pattern though. The left lateral edge is retouched more intensively than the right lateral edge. On the obverse face (Surface 2) the flake scar pattern is more convergent. It is also opposed to the direction of th e flake scars on Surface 1. The base of the point on Surface 2 is retouched more heavily than the other faces and when viewed in profile, the point is distinctly lenticular. This suggests that the Surface 2 retouching was bifacial thinning and very likel y for hafting purposes. G10 L30 T Group Lower DCC BN 3574.1 The piece is made on an end struck flake blank and it has been retouched mainly unifacially on the dorsal surface there is a large flake scar on the ventral proximal surface. Retouch along the dorsal surface is mainly unidirectional (proximal dorsal) but there are several opposing flake scars along the distal edge, near the tip that may be due to impact fractures meaning that the tip of the piece is effectively miss ing. There is finer retouch along the left lateral edge near the striking platform and the right lateral is retouched concavely. If the opposing distal dorsal flake scars (distal to proximal) are due to impact fracturing then it is likely that the origin al planform of this piece was sub triangular with a semi convex base. G10 L25 T Group Lower CTT / HLB BN 3575.3 BN 3575.3 is a parti bifacial point made on an end struck flake. The planform is distinctly tear drop in shape an d the lateral edges widen at the base to produce a characteristic wide, and rounded, butt. The flake blank platform is not facetted. The bulbar thinning is located primarily on the right lateral ventral surface and extends half way up the right lateral edg e. At least 75% of the bulb is missing. The dorsal surface is retouched G10 L24 T Group Upper DCC
250 heavily. The dorsal scar pattern is technically circular but is more accurately described as bi lateral because the majority of flake scars were struck from the lateral edges and no t the proximal or distal edges. The dorsal left lateral surface is retouched more finely than the right lateral surface and is just slightly concave form the medial lateral to the tip. BN 3477.3 BN 3477.3 is a parti bifacial point made on an end struck flake. The point planform is teardrop shaped and the lateral edges expand at the base to produce a wide, rounded butt. The ventral right lateral surface has been retouched fairly intensively. Thi s appears to be bulbar thinning, but it is actually basal thinning because the point itself is actually made on the distal end of the flake blank. The point profile is lenticular. The dorsal surface retouch is semi circular with fine retouching along the butt and right lateral edge near the tip which is narrowly concave. G10 L24 T Group Upper DCC BN 3132.1 BN 3132.1 is a L evallois point. The flake blank platform surface is facetted. The dorsal scar pattern is convergent but th e flake scars themselves show that the striking platforms alternated end to end. There is moderate edge damage along the dorsal left lateral and also the ventral right lateral edges. The style of manufacture especially the dorsal scars oriented from the distal edge is reminiscent of Nubian type I technology (Vermeersch 2001). G10 T Group Upper DCC BN 3114.1 BN 3114.1 was recovered from the profile wall during the 2007 field season. It was plotted in situ but not excavated as such. It is a L evallois point with a convergent dorsal scar pattern. Unlike BN 3132.1, which has an alternating convergent dorsal scar pattern the flake scars on BN 3114.1 are unidirectional. Dorsal scar flaking is finer along the proximal dorsal surface and the platform is facetted. There is edge damage along the left lateral edge near the base and along both lateral edges equidistant from the butt this may be evidence of hafting. There is right lateral edge damage on the ventral surface. G9 T Group Upper BN 2116.2 BN 2116.2 is a small bifacial point. Retouching along both faces is intensive but there is a small patch of cortex remaining on the distal surface. The point has a tear drop shape plan form with a rounded but the lateral edges do not expand outwards very much creating almost an ovate shape. The ventral surface retouching is semi circular most bi lateral though and the flake scars themselves are broad and poorly refined. In contrast, the dorsal surface flaking is much more refined and circular. Retouching along the butt is especially fine, though it is along fairly non invasive. The tip end is broken off. G10 L19 T Group Upper DCC BN 1818.1 BN 1818.1 is a unifacial point made on a transverse flake. The flake blank platform is facetted. The dorsal scar pattern is semi circular but the G10
251 L16 flake scars are predominantly oriented bi laterally (in this case proximal dorsal). The dorsal surface flaking is fairly coarse but there is irre gular, shallow lateral retouching from the medial to the tip forming slightly serrated lateral edges. Technologically, this pieces appears to be Levallois but the retouching and plan form suggest it shares technological elements in common with other, smal ler tear drop shaped points within the assemblage. S Group VDBS BN 1802.90 BN 1802.90 is a small parti bifacial point made on the medial section of a flake. The dorsal scar pattern is bi lateral (lateral lateral). Th e base of the point appears to be the left lateral edge. The left lateral flake scars are slightly finer than those on the right lateral. There is also slightly more intensive flaking along the ventral surface of the left lateral edge to form a thinner profile. Furthermore, there is moderate right lateral ventral retouching to make a more convergent edge. The plan form of the point is reminiscent of other tear drop shaped points within the assemblage though BN 1802.90 is clearly more ovate. G10 L1 3 S Group VDBS BN 1799.12 BN 1799.12 is an end struck flake with a parallel / convergent dorsal scar pattern the two flake scars at the distal edge, and along the right lateral edge are opposed to the uniform direction of all other flake scars. The flake blank platform is facetted. This piece is reminiscent of Levallois technology. Whether this piece is a point, however, can be questioned. There is non invasive utilization along the ventral right lateral edge. This may indicat e that this edge was, in fact, the primary working edge of the tool and not the tip. G10 L12 S Group VDBS BN 1799.16 BN 1799.16 is a small bifacial point with intensive retouching on both faces. The retouching on Surface 1 is more i ntensive than on Surface 2. Surface 1 retouching itself is more intensive along the point butt and left lateral edges. Retouching on Surface 2 is more intensive along the right lateral edge. The point has a semi lenticular profile but the profile is ova te and the tip is also rounded. This may suggest the piece was retouched heavily during use and is thus the end product of a much different initial point shape, possibly being more elongate and tear drop shaped like bifacial point BN 2116.2. G10 L12 S Group VDBS BN 1799.13 BN 1799.13 is a small bifacial point with intensive retouching on both faces. Similar to 1799.16, Surface 1 is retouched more heavily than Surface 2. Retouching along Surface 1 is circular and there is a intens ive retouching along the medial right lateral edge that may be a shallow notch. The retouching on Surface 2 is largely bi directional and it is much coarser than on Surface 1, except at the tip / right lateral edge where there are finer flake scars. Ther e is no equivalent retouching on Surface 2 adjacent to the possible Surface 1 notch. Due to the medial location on the piece, the notch may relate to hafting. The point has a lenticular profile G10 L12 S Group VDBS
252 and the plan form is tear drop. There is similarity in desi gn with BN 1799.16 and this may be an example of a bifacial point that was not retouched as heavily (due to use?) as BN 1799.16. BN 1794.36 BN 1794.36 is made on the distal edge of an end struck flake It is possible that this piece is incomplete (being broken at the butt) but this seems unlikely because some dorsal scars emanate from the existing butt edge. The original dorsal scar pattern was likely parallel and opposed to the flake blank platform. The retouching pattern is semi circular. The point is retouched most heavily along the dorsal left lateral and butt edges. There is some retouching adjacent to a shallow divot on the medial right latera l, which is reminiscent of a notch for hafting purp oses (see BN 1799.13). There is no equivalent ventral retouching around the notch. The point has an irregular / semi lenticular profile and a triangular planform. G10 L11 S Group VDBS BN 4258.1 BN 4258.1 is a small parti bi facial made on an end struck flake. The dorsal scar pattern is circular and retouching intensity is similar along all edges. There are two large flake scars at the butt of the piece that terminate in step fractures medially along the axis of the tool. In profile, this creates a biconvex shape with basal thinning. There is retouching along the ventral left lateral edge near the proximal edge as well as retouching or possible edge damage / use wear along the ventral right la teral edge near the tip. The point planform is sub triangular almost tear drop with straight convergent edges at the tip that morph into parallel straight edges for the medial section of the tool to the butt. The butt is convex. H9 L26 S Group VDBS BN 2268.2 BN 2268.2 is a parti bifacial point made on the distal edge of an end struck flake. The flake blank was snapped medially (lateral lateral) similar in design to BN 17944.36. BN 2268.2 is retouched most heavily along the dorsal surface. The do rsal scar pattern is circular and there is equivalent retouching intensity along each edge. There is edge damage / utilization on the dorsal left lateral edge. The ventral surface is retouched most heavily at the tip where there are bi lateral (lateral lateral) flake scars. There is edge damage / utilization along the ventral right lateral edge. The point has a rhomboidal profile and the plan form is triangular. TU2S S Group VDBS YBS BN1693.3 BN 1693.3 is a part bifacia l point made on an end struck flake. There is intensive semi circular bulbar thinning on the ventral proximal surface which has completely removed the bulb of percussion. The dorsal scar pattern is circular and retouching is finest along both lateral edge s medially towards the point tip. The lateral edges expand towards the butt, which is rounded, forming a tear drop shape. G10 R Group RGCB
25 3 BN 1887.1 BN 1887.1 is an elongate unifacial point made on an end struck flake. There is no bulba r thinning. The point is retouched most heavily along the right lateral edge, but there is also fine retouching along the distal, left lateral edge at the tip. The original dorsal scar pattern of the flake blank may have been parallel. The proximal left lateral edge is naturally convergent but the proximal right lateral edge has been intentionally retouched similarly to produce a characteristic trapezoidal butt shape. G10 L5 R Group RGCA BN 2043.1 BN 2043.1 is a parti bifacial point made on an end struck (possibly transverse) whole flake. The flake blank dorsal scar pattern was parallel and opposed to the striking platform. The dorsal surface is retouched most heavily along the butt. The ventral surface is also retouched most heavily al ong the butt and extending along the ventral right lateral edge for bulbar thinning. The point has a lenticular profile. The lateral edges are slightly convex (and skewed to the left lateral) forming an ovate planform with a rounded base. H9 L15 R Group RCA BN 2072.1 BN 2072.1 is a parti bifacial point made on an end struck flake. The retouching is heaviest on the dorsal surface. The dorsal scar pattern is semi circular and the butt is retouched heaviest. There is intensive bi latera l (lateral lateral) / semi circular bulbar thinning on the ventral surface. The point has a lenticular profile. The plan form is elongate with characteristic slightly convex lateral edges that expand outwards towards a more steeply rounded semi trapezo idal butt. The semi trapezoidal butt shape is reminiscent of BN 1887.1. H9 L15 R Group RCA
254 REFERENCE LIST Allen, J.R.M., Brandt, U., Brauer, A., Hubberten, H.W., Huntley, B., Keller, J., Kraml, M., Mackensen, A. Mingram, J., Negendank, J.F.W., Nowaczyk, N.R., Oberhansli, H., Watts, W.A., Wulf, S., and Zolitschka, B., 1999. Rapid environmental changes in southern Europe during the last glacial period, Nature 400, 740 743. Ambrose, S., 2001. Middle and Later Stone Age settlement patterns in the Central Rift Valley, Kenya: comparisons and contrasts. in: Conard, N. (Ed.), Settlement Dynamics of the Middle Paleolithic and Middle Stone Age. Kerns Verlag, Tubingen, pp. 21 43. Ambrose, S.H., 1998a. Chronology of the Lat er Stone Age and food production in East Africa, Journal of Archaeological Science 25, 377 392. Ambrose, S.H., 1998b. Late Pleistocene human population bottlenecks, volcanic winter, and differentiation of modern humans, Journal of Human Evolution 34, 623 6 51. Ambrose, S.H., 2002. Small things remembered: origins of early microlithic industries in Sub Saharan Africa, Archeological Papers of the American Anthropological Association 12, 9 29. Ambrose, S.H., 2003. Did the super eruption of Toba cause a human po pulation bottleneck? Reply to Gathorne Hardy and Harcourt Smith, Journal of Human Evolution 45, 231 237. Ambrose, S.H., Elston, R.G., and Kuhn, S.L., 2002. Small things remembered: Origins of early microlithic industries in Sub Saharan Africa. Thinking Sm all: Global Perspectives on Microlithization, pp. 9 29. Andersen, K.K., Svensson, A., Johnsen, S.J., Rasmussen, S.O., Bigler, M., Rothlisberger, R., Ruth, U., Siggaard Andersen, M.L., Steffensen, J.P., Dahl Jensen, D., Vinther, B.M., and Clausen, H.B., 20 06. The Greenland Ice Core Chronology 2005, 15 42 ka. Part 1: constructing the time scale, Quaternary Science Reviews 25, 3246 3257. Anovitz, L.M., Elam, J.M., Riciputi, L.R., and Cole, D.R., 1999. The failure of obsidian hydration dating: sources, implica tions, and new directions, Journal of Archaeological Science 26, 735 752. Arnold, J.R. and Libby, W.F., 1949. Age determinations by radiocarbon content checks with samples of known age, Science 110, 678 680. Assefa, Z., 2006. Faunal remains from Porc Epi c: Paleoecological and zooarchaeological investigations from a Middle Stone Age site in southeastern Ethiopia, Journal of Human Evolution 51, 50 75. Bar Matthews, M., Ayalon, A., and Kaufman, A., 2000. Timing and hydrological conditions of Sapropel events in the Eastern Mediterranean, as evident from speleothems, Soreq cave, Israel, Chemical Geology 169, 145 156.
255 Bard, E., 2003. Indian Ocean sea surface temperature reconstruction. IGBP PAGES/World Data Center for Paleoclimatology, Data Contribution Series #2003 027. NOAA/NGDC Paleoclimatology Program, Boulder Colorado., Bard, E., Rosteck, F., and Sonzogni, C., 1997. Interhemispheric synchrony of the last deglaciation inferred from alkenone palaeothermometry, Nature 385, 707 710. Barham, L. and Mitchell, P ., 2008. The first Africans: African archaeology from the earliest toolmakers to the most recent foragers, Cambridge University Press, New York. Barton, R.N.E., Bouzouggar, A., Collcutt, S.N., Schwenninger, J.L., and Clark Balzan, L., 2009. OSL dating of t he Aterian levels at Dar es Soltan I (Rabat, Morocco) and implications for the dispersal of modern Homo sapiens, Quaternary Science Reviews 28, 1914 1931. Basell, L.S., 2008. Middle Stone Age (MSA) site distributions in eastern Africa and their relationshi p to Quaternary environmental change, refugia and the evolution of Homo sapiens, Quaternary Science Reviews 27, 2484 2498. Baumgartner, S., Beer, J., Masarik, J., Wagner, G., Meynadier, L., and Synal, H.A., 1998. Geomagnetic Modulation of the 36Cl Flux in the GRIP Ice Core, Greenland, Science 279, 1330 1332. Beckingham, C.F. and Huntingford, G.W.B., 1954. Some records of Ethiopia, 1593 1646. Hakluyt Society, London, p. lxxx. Benvenuti, M., Carnicelli, S., Belluomini, G., Dainelli, N., Di Grazia, S., Ferra ri, G.A., Iasio, C., Sagri, M., Ventra, D., Atnafu, B., and Kebede, S., 2002. The Ziway Shala lake basin (main Ethiopian rift, Ethiopia): a revision of basin evolution with special reference to the Late Quaternary, Journal of African Earth Sciences 35, 247 269. Berna, F., Matthews, A., and Weiner, S., 2004. Solubilities of bone mineral from archaeological sites: the recrystallization window, Journal of Archaeological Science 31, 867 882. Binford, L.R., 1962. Archaeology as anthropology, American Antiquity 2 8, 217 225. Binford, L.R., 1965. Archaeological systematics and the study of culture process, American Antiquity 31, 203 210. Binford, L.R., 1980. Willow's smoke and dogs' tails: Hunter gatherer settlement systems and archaeological site formation, America n Antiquity 45. Blanckenhorn, M., 1901. Das Pliocn und Quartrzeitalter in Aegypten ausschliesslich des Rothen Meergebietes. Neues zur Geologie und Palaeontology Aegyptens IV Zeitschrift der Deutschen geologischen Gesellschaft, Jahrgang, pp. 307 502. Bo nd, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L., Jouzel, J., and Bonani, G., 1993. Correlations between climate records from North Atlantic sediments and Greenland ice, Nature 365, 143 147.
256 Bond, G., Showers, W., Cheseby, M., Lotti, R., Almasi P., deMenocal, P., Priore, P., Cullen, H., Hajdas, I., and Bonani, G., 1997. A Pervasive millennial scale cycle in North Atlantic Holocene and glacial climates, Science 278, 1257 1266. Borena, A., 2008. Project proposal on natural resource conservation i n Mt. Damota area, International NGO Journal 3, 48 56. Bowler, J.M., Johnston, H., Olley, J.M., Prescott, J.R., Roberts, R.G., Shawcross, W., and Spooner, N.A., 2003. New ages for human occupation and climatic change at Lake Mungo, Australia, Nature 421, 8 37 840. Bradley, R.S., 1999. Paleoclimatology: Reconstructing climates of the Quaternary, Academic Press, San Diego. Brandt, S.A. and Brook, G.A., 1984. Archaeological and paleoenvironmental research in Northern Somalia, Current Anthropology 25, 119 121. B randt, S.A., 1982. A Late Quaternary cultural / environmental sequence from lake Besaka, southern Afar, Ethiopia. Ph.D. thesis, University of California, Berkeley, Ann Arbor: University microfilms Brandt, S.A., 1986. The Upper Pleistocene and early Holoce ne prehistory of the Horn of Africa, African Archaeological Review 4, 41 82. Brandt, S.A., 1988. Early Holocene mortuary practices and hunter gatherer adaptations in Southern Somalia, World Archaeology 20, 40 56. Brandt, S.A., Fisher, E., Hildebrand, E., a nd Wang, H., 2010. Archaeological and paleoenvironmental implications of >40 ka charcoal radiocarbon dates from Moche Borago, Ethiopia. 2010 Paleoanthropology Society Meeting, St. Louis, Missouri Brandt, S.A.A. P., Hildebrand, E.E. P., Negash, A., Marshal l, F., Lesure, J., Lamb, H., Mohammed, M.U., Ambrose, S., Abate, A., Demissew, S., Rots, V., Unbushe, D., Denu, D., Renne, P., and Fisher, E., 2006. SW Ethiopia as a Late Quaternary refugium: archaeology and paleoenvironment. National Science Foundation, Brandt, S., 2000. Emergency archaeological fieldwork and capacity building at the Gilgel Gibe hydroelectric project, Deneba, southern Ethiopia. Unpublished report to the Ethiopia tourism commission and Center for Research and Conservation of Cultural Herit age. Broecker, W., Bond, G., Klas, M., Clark, E., and McManus, J., 1992. Origin of the northern Atlantic's Heinrich events, Climate Dynamics 6, 265 273. Broecker, W.S., 1994. Massive iceberg discharges as triggers for global climate change, Nature 372, 42 1 424. Brooks, A.S., Nevell, L., Yellen, J.E., and Hartman, G., 2006. Projectile technologies of the African MSA. Transitions Before the Transition, pp. 233 255.
257 Brooks, C.E.P., 1914. The meteorological conditions of the ice sheet and their bearing on th e desiccation of the Globe., Quarterly Journal of the Royal Meteorological Society xi. Burns, S.J., Fleitmann, D., Matter, A., Kramers, J., and Al Subbary, A.A., 2003. Indian Ocean climate and an absolute chronology over Dansgaard/Oeschger Events 9 to 13, Science 301, 1365 1367. Burns, S.J., Fleitmann, D., Matter, A., Kramers, J., and Al Subbary, A.A., 2004. CORRECTIONS AND CLARIFICATIONS, Science 305, 1567a. Cai, Y., An, Z., Cheng, H., Edwards, R.L., Kelly, M.J., Liu, W., Wang, X., and Shen, C.C., 2006. Hi gh resolution absolute dated Indian Monsoon record between 53 and 36 ka from Xiaobailong Cave, southwestern China, Geology 34, 621 624. Capron, E., Landais, A., Lemieux Dudon, B., Schilt, A., Masson Delmotte, V., Buiron, D., Chappellaz, J., Dahl Jensen, D. Johnsen, S., Leuenberger, M., Loulergue, L., and Oerter, H., 2010. Synchronising EDML and NorthGRIP ice cores using d18O of atmospheric oxygen (d18Oatm) and CH4 measurements over MIS5 (80 123 kyr), Quaternary Science Reviews 29, 222 234. Cayre, O., Lance lot, Y., Vincent, E., and Hall, M., 1999. Paleoceanograpic reconstructions from planktonic formainifera of the Iberian margin: temperature, salinity, and Heinrich events, Paleoceanography 14, 384 396. Cerulli, E., 1956. Part III: North Eastern Africa: Peop les of Southwest Ethiopia and its borderland, International African Institute, London. Chalie, F. and Gasse, F., 2002. Late Glacial Holocene diatom record of water chemistry and lake level change from the tropical East African Rift Lake Abiyata (Ethiopia), Palaeogeography, Palaeoclimatology, Palaeoecology 187, 259 283. Chiu, T.C., Fairbanks, R.G., Cao, L., and Mortlock, R.A., 2007. Analysis of the atmospheric C 14 record spanning the past 50,000 years derived from high precision Th 230/U 234/U 238, Pa 231/U 235 and C 14 dates on fossil corals, Quaternary Science Reviews 26, 18 36. Clark, J.D., 1960. Human ecology during Pleistocene and later times in Africa South of the Sahara, Current Anthropology 1, 307 324. Clark, J.D., 1963. The evolution of culture in A frica, The American Naturalist 97, 15. Clark, J.D., 1964. The prehistoric origins of African culture, The Journal of African History 5, 161 183. Clark, J.D., 1965. The later Pleistocene cultures of Africa, Science 150, 833 847. Clark, J.D., 1970. The prehi story of Africa, Praeger, New York.
258 Clark, J.D., 1988. The Middle Stone Age of East Africa and the beginnings of regional identity, Journal of World Prehistory. 2, 235 305. Clark, J.D. and Williams, M.A.J., 1978. Recent archaeological research in Southeast ern Ethiopia, Annales d'Ethiopie 19 44. Clark, J.D., Williamson, K.D., Michels, J.W., and Marean, C.W., 1984. A Middle Stone Age occupation site at Porc Epic cave, Dire Dawa (east central Ethiopia), The African Archaeological Review 2, 37 71. Clark, J.D., 1954. The prehistoric cultures of the Horn of Africa: an analysis of the stone age cultural and climatic succession in the Somalilands and eastern parts of Abyssinia, Cambridge University Press, Cambridge. Clark, J.G.D., 1971. World prehistory, A New Outli ne, Cambridge University Press, Cambridge. Close, A.E., Straus, L.G., Eriksen, B.V., Erlandson, J.M., and Yesner, D.R., 1996. Plus Ca Change: The Pleistocene Holocene Transition in Northeast Africa. Humans at the End of the Ice Age: The Archaeology of the Pleistocene Holocene Transition Plenum Press, New York, pp. 43 60. Cohen, A.S., Stone, J.R., Beuning, K.R.M., Park, L.E., Reinthal, P.N., Dettman, D., Scholz, C.A., Johnson, T.C., King, J.W., Talbot, M.R., Brown, E.T., and Ivory, S.J., 2007. Ecological c onsequences of early Late Pleistocene megadroughts in tropical Africa, Proceedings of the National Academy of Sciences 104, 16422 16427. Cole, S., 1954. The prehistory of East Africa, American Anthropologist 56, 1026 1050. Cornelissen, E., 2002. Human resp onses to changing environments in Central Africa between 40,000 and 12,000 B.P, Journal of World Prehistory 16, 197 235. Cosford, J., Qing, H., Yuan, D., Zhang, M., Holmden, C., Patterson, W., and Hai, C., 2008. Millennial scale variability in the Asian mo nsoon: Evidence from oxygen isotope records from stalagmites in southeastern China, Palaeogeography, Palaeoclimatology, Palaeoecology 266, 3 12. Dansgaard, W., 1964. Stable isotopes in precipitation, Tellus 16, Medium. Dansgaard, W., Johnson, S.J., Clausen H.B., hl Jensen, D., Gudenstrup, N.S., Hammer, C.U., Hvidberg, C.S., Steffensen, J.P., Sveinbjornsdottir, A.E., Jouzel, J., and Bond, G., 1993. Evidence for general instability of past climate from a 250 kyr Ice core record, Nature 364, 218 220. del Mora l, R. and Lacher, I.L., 2005. Vegetation patterns 25 years after the eruption of Mount St. Helens, Washington, USA, American Journal of Botany 92, 1948 1956. Di Paola, G., 1972. The Ethiopian rift valley, Bulletin of Volcanology 36, 517 560.
259 Fisher, E.C., 2005. A complex systems theory of technological change: a case study involving a morphometrics analysis of stone age flake debitage from the Horn of Africa. University of Florida, Fisher, E.C., Bar Matthews, M., Jerardino, A., and Marean, C.W., 2010. Midd le and Late Pleistocene paleoscape modeling along the southern coast of South Africa, Quaternary Science Reviews 29, 1382 1398. Fletcher, W.J., Sanchez Gone, M.F., llen, J.R.M., Cheddadi, R., Combourieu Nebout, N., Huntley, B., Lawson, I., Londeix, L., Mag ri, D., Margari, V., Mnller, U.C., Naughton, F., Novenko, E., Roucoux, K., and Tzedakis, P.C., 2010. Millennial scale variability during the last glacial in vegetation records from Europe, Quaternary Science Reviews In Press, Corrected Proof. Flint, R.F., 1959. On the basis of Pleistocene correlation in East Africa, Geological Magazine XCVI, 265 284. Flores, J.A., Barcena, M.A., and Sierro, F.J., 2000. Ocean surface and wind dynamics in the Atlantic Ocean off Northwest Africa during the last 140 000 years, Palaeogeography, Palaeoclimatology, Palaeoecology 161, 459 478. Forster, P., Harding, R., Torroni, A., and Bandelt, H.J., 1996. Origin and evolution of native American mtDNA variation: a reappraisal, American Journal of Human Genetics 59, 935 945. Galbrait h, R.F. and Laslett, G.M., 1993. Statistical Models for Mixed Fission Track Ages, Nuclear tracks and radiation measurements 21, 459 470. Galbraith, R.F., Roberts, R.G., Laslett, G.M., Yoshida, H., and Olley, J.M., 1999. Optical dating of single and multipl e grains of quartz from Jinmium rock shelter, northern Australia, part 1: Experimental design and statistical models, Archaeometry 41, 339 364. Galbraith, R.F., Roberts, R.G., and Yoshida, H., 2005. Error variation in OSL palaeodose estimates from single a liquots of quartz: a factorial experiment, Radiation Measurements 39, 289 307. Garcea, E.A.A. and Giraudi, C., 2006. Late Quaternary human settlement patterning in the Jebel Gharbi, Journal of Human Evolution 51, 411 421. Gasse, F., 1977. Evolution of Lake Abhe (Ethiopia and TFAI), from 70,000 b.p., Nature 265, 42 45. Gasse, F., 2000. Hydrological changes in the African tropics since the Last Glacial Maximum, Quaternary Science Reviews 19, 189 211. Gasse, F. and Street, F.A., 1978. Late Quaternary lake leve l fluctuations and environments of the Northern Rift Valley and Afar Region ( Ethiopia and Djibouti), Palaeogeography, Palaeoclimatology, Palaeoecology 24, 279 325.
260 Genty, D., Blamart, D., Ouahdi, R., Gilmour, M., Baker, A., Jouzel, J., and Van Exter, S., 2003. Precise dating of Dansgaard Oeschger climate oscillations in western Europe from stalagmite data, Nature 421, 833 837. Gillespie, R., Street Perrott, F.A., and Switsur, R., 1983. Post glacial arid episodes in Ethiopia have implications for climate pr ediction, Nature 306, 680 683. Greenland Ice core Project, M., 1993. Climate instability during the last interglacial period recorded in the GRIP ice core, Nature 364, 203 207. Grootes, P.M., Stuiver, M., White, J.W.C., Johnsen, S., and Jouzel, J., 1993. C omparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores, Nature 366, 552 554. Grousset, F.E., Labeyrie, L., Sinko, J.A., Cremer, M., Bond, G., Duprat, J., Cortijo, E., and Huon, S., 1993. Patterns of ice rafted detritus in the glaci al North Atlantic (40 55 C), Paleoceanography 8, 175 192. Guiot, J., de Beaulieu, J.L., Cheddadi, R., David, F., Ponel, P., and Reille, M., 1993. The climate in Western Europe during the last Glacial/Interglacial cycle derived from pollen and insect rema ins, Palaeogeography, Palaeoclimatology, Palaeoecology 103, 73 93. Gutherz, X., Jallot, L., Lesur, J., Pouzolles, G., and Sordoillet, D., 2001. The Moche Borago (Soddo Wolayta) Rock shelter: December 2001 Excavation. Report on file, SNNPR Office of Cultur e and Information, Awassa, Ethiopia., Gutherz, X., Jallot, L., Poisblaud, B., Sordoillet, D., Pouzolles, G., Asfaw, A., Lakew, B., and Aba, G., 2000a. The Moche Borago Rock shelter: January February 2000 Excavation. Reports on file, SNNPR Office of Cultur e and Information, Awassa, Ethiopia., Gutherz, X., Said.H., and Lakew, B., 2000b. The Moche Borago Rock shelter: November 2000 Excavation. Report on file, SNNPR office of Culture and Information, Awassa, Ethiopia., Hedges, R.E.M. and Vanklinken, G.J., 199 2. A Review of Current Approaches in the Pretreatment of Bone for Radiocarbon Dating by Ams, Radiocarbon 34, 279 291. Hemming, S.R., 2004. Heinrich events: Massive late Pleistocene detritus layers of the North Atlantic and their global climate imprint, Rev iews of Geophysics 42. Henshilwood, C., D'Errico, F., Vanhaeren, M., van Niekerk, K., and Jacobs, Z., 2004. Middle Stone Age shell beads from South Africa, Science 304, 404. Henshilwood, C.S. and Marean, C.W., 2003. The origin of modern human behavior: a c ritique of the models and their test implications, Current Anthropology 44.
261 Herries, A.I.R. and Fisher, E.C., Multi dimensional modelling of magnetic mineralogy as a proxy for fire use and spatial patterning: evidence from the Middle Stone Age bearing sea cave of Pinnacle Point 13B (Western Cape, South Africa). Submitted, In Press. Hessler, I., Dupont, L., Bonnefille, R., Behling, H., Gonzales, C., Helmens, K.F., Hooghiemstra, H., Lebamba, J., Ledru, M.P., Lazine, A. M., Maley, J., Marret, F., and Vincens A., 2010. Millennial scale changes in vegetation records from tropical Africa and South America during the last glacial, Quaternary Science Reviews In Press, Corrected Proof. Hildebrand, E., 2003. Enset, yams, and honey: ethnoarchaeological approaches to the origins of horticulture in southwest Ethiopia. Washington University, St. Louis, Missouri Hildebrand, E., Brandt, S.A., Fisher, E.C., Mohammed, M.U., Wang, H., Ambrose, S., Roberts, R., Grun, R., Morgan, L., and Lesure, J., 2008. Late Pleistocene env ironments and technological change in a northeast African refugium: current research at Moche Borago rockshelter, southwestern Ethiopia. Paleoanthropology society meeting, 2008 Paleoanthropology Society, Huber, C., Leuenberger, M., Spahni, R., Fluckiger, J., Schwander, J., Stocker, T.F., Johnsen, S., Landais, A., and Jouzel, J., 2006. Isotope calibrated Greenland temperature record over Marine Isotope Stage 3 and its relation to CH4, Earth and Planetary Science Letters 243, 504 519. Hughen, K., Lehman, S., Southon, J., Overpeck, J., Marchal, O., Herring, C., and Turnbull, J., 2004. 14C activity and global carbon cycle changes over the past 50,000 Years, Science 303, 202 207. Hughen, K., Southon, J., Lehman, S., Bertrand, C., and Turnbull, J., 2006. Marine d erived C 14 calibration and activity record for the past 50,000 years updated from the Cariaco Basin, Quaternary Science Reviews 25, 3216 3227. Hull, K.L., 2001. Reasserting the utility of obsidian hydration dating: A temperature dependent empirical approa ch to practical temporal resolution with archaeological obsidians, Journal of Archaeological Science 28, 1025 1040. Jacobs, Z., Duller, G.A.T., Wintle, A.G., and Henshilwood, C.S., 2006. Extending the chronology of deposits at Blombos Cave, South Africa, b ack to 140 ka using optical dating of single and multiple grains of quartz, Journal of Human Evolution 51, 255 273. Jacobs, Z., Roberts, R.G., Galbraith, R.F., Deacon, H.J., Grun, R., Mackay, A., Mitchell, P., Vogelsang, R., and Wadley, L., 2008. Ages for the Middle Stone Age of southern Africa: implications for human behavior and dispersal, Science 322, 733 735. Jallot, L. and Pouzolles, G., 2000. Test pit 1 extension: excavation results. in: Gutherz, X., Jallot, L., Lesur, J., Pouzolles, G., Sordoillet, D ., Said, H., and Lakew, B. (Eds.), The Moche Borago Rockshelter: November 2000 Excavation.
262 Kamenov, G.D., Brenner, M., and Tucker, J.L., 2009. Anthropogenic versus natural control on trace element and Sr Nd Pb isotope stratigraphy in peat sediments of sout heast Florida (USA), similar to 1500 AD to present, Geochimica et Cosmochimica Acta 73, 3549 3567. Kebede, S., Travi, Y., Asrat, A., Alemayehu, T., Ayenew, T., and Tessema, Z., 2008. Groundwater origin and flow along selected transects in Ethiopian rift vo lcanic aquifers, Hydrogeology Journal 16, 55 73. Kebede, S., Travi, Y., and Rozanskic, K., 2009. The delta O 18 and delta H 2 enrichment of Ethiopian lakes, Journal of Hydrology 365, 173 182. Kebede, S., 2004. Approaches isotopique et geochemique pout l'et ude Nil Bleu et du rift Ethiopien. University of Avignon, France, Kebede, S., Travi, Y., Alemayehu, T., and Ayenew, T., 2005. Groundwater recharge, circulation and geochemical evolution in the source region of the Blue Nile River, Ethiopia, Applied Geoche mistry 20, 1658 1676. Kelly, R.L., 1995. The Foraging Spectrum, Smithsonian Institution Press, Washington D.C. Kiage, L.M. and Liu, K., 2006. Late Quaternary paleoenvironmental changes in East Africa: a review of multiproxy evidence from palynology, lake s ediments, and associated records, Progress in Physical Geography 30, 633 658. Kurashina, H., 1978. An Examination of Prehistoric Lithic Technology in East Central Ethiopia. (Ph.D. thesis, University of California, Berkeley)., University Microfilms, Ann Ar bor. Labeyrie, L.D., Juillet Leclerc, A., Kallel, N., and Blanc, P.L., 1991. Sea level and oceanic thermohaline circulation: changes over a glacial/interglacial cycle. in: Frenzel, B. (Ed.), Klimageschichtliche Probleme der letzten 130'000 Jahre Fischer Ve rlag, Mainz, Stuttgart, pp. 197 214. Lamb, A.L., Leng, M.J., Lamb, H.F., and Mohammed, M.U., 2000. A 9000 year oxygen and carbon isotope record of hydrological change in a small Ethiopian crater lake, The Holocene 167 177. Lamberti, M. and Sottile, R., 19 97. The Wolaytta Language, Rudiger Koppe Verlag, Cologne. Landais, A., Barnola, J.M., Masson Delmotte, V., Jouzel, J., Chappellaz, J., Caillon, N., Huber, C., Leuenberger, M., and Johnsen, S.J., 2004. A continuous record of temperature evolution over a seq uence of Dansgaard Oeschger events during Marine Isotopic Stage 4 (76 to 62 kyr BP), Geophysical Research Letters 31. Le Gal, E. and Molinier, N., 2006. Agricultural and economic analysis diagnosis of Obe Jage (Damot Gale, Wolayta). Institut National Agro nomique Paris Grignon, Leakey, L.N. and Solomon, J.D., 1929. East African Archaeology, Nature 124.
263 Leakey, L.S.B., 1931. The Stone Age cultures of Kenya colony, Cambridge University Press, Cambride. Leakey, L.S.B., 1936. Stone age Africa: An outline of pr ehistory in Africa, Oxford University Press, Oxford. Leakey, M.D., 1971. Olduvai Gorge Volume III: Excavations in beds I and II, 1960 1963, Cambridge University Press, Cambridge. Lee, R.B. and DeVore, I., 1968. Man the Hunter, Aldine, Chicago. Lesur, J., 2 000. Moche Borago's Fauna. in: Gutherz, X., Jallot, L., Lesur, J., Pouzolles, G., Sordoillet, D., Said, H., and Lakew, B. (Eds.), The Moche Borago rock shelter: November 2000 Excavation Lesur Gebremariam, J., 2008. Climatic changes in Holocene Horn of Afri ca and animals exploitations by human societies. Society of Africanist Archaeologists Leuschner, D.C. and Sirocko, F., 2000. The low latitude monsoon climate during Dansgaard Oeschger cycles and Heinrich Events, Quaternary Science Reviews 19, 243 254. Lez ine, A.M. and Casanova, J., 1991. Correlated oceanic and continental records demonstrate past climate and hydrology of North Africa (0 140 ka), Geology 19, 307 310. Libby, W.F., 1951. Radiocarbon dates 2, Science 114, 291 296. Libby, W.F. and Arnold, J.R., 1950. Radiocarbon Dates, Science 112, 453. Livingstone, D.A., 1975. Late Quaternary climatic change in Africa, Annual Review of Ecology and Systematics 6, 249 280. Lundblad, K. and Holmgren, K., 2005. Paleoclimatological survey of stalagmites from coastal areas in Tanzania., Geografiska Annaler: Series A, Physical Geography 87, 125 140. MacAyeal, D.R., 1993. Binge/purge oscillations of the Laurentide ice sheet as a cause of the North Atlantic's Heinrich events, Paleoceanography 8. Maitima, J.M., 1991. Vege tation response to climatic change in Central Rift Valley, Kenya, Quaternary Research 35, 234 245. Marean, C.W. and Assefa, Z., 2005. The middle and upper Pleistocene African record for the biological and behavioral origins of modern humans. in: Stahl, A. (Ed.), African Archaeology: A Critical Introduction Blackwell Publishing Ltd., pp. 93 129. Marean, C.W., Bar Matthews, M., Bernatchez, J., Fisher, E., Goldberg, P., Herries, A.I.R., Jacobs, Z., Jerardino, A., Karkanas, P., Minichillo, T., Nilssen, P.J., T hompson, E., Watts, I., and Williams, H.M., 2007. Early human use of marine resources and pigment in South Africa during the Middle Pleistocene, Nature 449, 905 908.
264 Marks, A. and Conard, N., 2006. Technology vs. typology: the case for and against a transi tion from the MSA to the LSA at Mumba Cave, Tanzania. in: Aubry, T., Almeida, F., Ara n jo, A.C., and Tiffagom, M. (Eds.), XV World Congress UISPP BAR Press, Lisbon Marlowe, F., 2010. The Hadza: Hunter Gatherers of Tanzania, University of California Press, B erkeley. Maslin, M.A., Seidov, D., and Lowe, J., 2001. Synthesis of the nature and causes of sudden climate transitions during the Quaternary. in: Seidov, Haupt, and Maslin, M.A. (Eds.), The Oceans and Rapid Climate Change: Past, Present and Future Mazer, J.J., Stevenson, C.M., Ebert, W.L., and Bates, J.K., 1991. The experimental hydration of obsidian as a function of relative humidity and temperature, American Antiquity 56, 504 513. McBrearty, S. and Brooks, A.S., 2000. The revolution that wasn't: a new in terpretation of the origin of modern human behavior, Journal of Human Evolution 39, 453 563. Meese, D.A., Gow, A.J., Grootes, P., Stuiver, M., Mayewski, P.A., Zielinski, G.A., Ram, M., Taylor, K.C., and Waddington, E.D., 1994. The accumulation record from the GISP2 Core as an indicator of climate change throughout the Holocene, Science 266, 1680 1682. Mehlman, M.J., 1989. Later Quaternary Archaeological Sequences in Northern Tanzania. University of Illinois at Urbana Champaign, Merrick, H., 1975. Change in Later Pleistocene Lithic Industries in Eastern Africa. Ph.D. Dissertation, University of California, Berkeley., University Microfilms Intl., Ann Arbor. Milankovitch, M., 1941. [New English Translation, 1998], Canon of Insolation and the Ice Age Problem. W ith introduction and biographical essay by Nikola Pantic, Alven Global. Mitchell, P., 2002. The Archaeology of Southern Africa, Cambridge University Press, Cambridge. Mohammed, U., Legesse, D., Gasse, F., Bonnefille, R., Lamb, H.F., and Leng, M.J., 2004. L ate Quaternary climate changes in the Horn of Africa. in: Battarbee, R.W., Gasse, F., and Stickley, C.E. (Eds.), Past Climate Variability through Europe and Africa Springer, Dordrecth, pp. 159 175. Muscheler, R., Beer, J., Kubik, P.W., and Synal, H.A., 20 05. Geomagnetic field intensity during the last 60,000 years based on 10Be and 36Cl from the Summit ice cores and 14C, Quaternary Science Reviews 24, 1849 1860. National Meteorlogical Agency, F.D.R.o.E., 2010. National Meteorlogical Agency City Index. Nee v, D. and Emery, K.O., 1967. The Dead Sea, depositional processes and environments of evaporites. Geolical Survey of Israel Bulletin, p. 147.
265 Negash, A., 2004. Unpublished report on the lithic analysis at Liben Bore, Ethiopia. Nilsson, E., 1929. Prelimi nary Report on the Wuaternary Geology of Mount Elgon and some Parts of the Rift Valley. Nilsson, E., 1949. The pluvials of East Africa. An attempt to correlate Pleistocene changes of climate, Geografiska Annaler 31, 204 211. North Greenland Ice Core Proje ct, 2004. High resolution record of Northern Hemisphere climate extending into the last interglacial period, Nature 431, 147 151. Olago, D.O., Street Perrott, F.A., Perrott, R.A., Ivanovich, M., and Harknessm D.D., 2001. EU/Th and 14C Isotope Dating of Lak e Sediments from Sacred Lake and Lake Nkunga, Kenya, African Journal of Science and Technology.Science and Engineering Series 2, 36 46. Oppenheimer, S., 2003. Out of Africa's Eden: The Peopling of the World, Jonathan Ball Publishers, Johannesburg. Pleurdea u, D., 2001. Gestion des Matieres Premieres et Comportements Techniques dans le Middle Stone Age Africain: Les Assemblages Lithiques La Grotte Du Porc Epic (Dire Dawa, Ethiopie). Ph.D. Dissertation, Museum National D'Histoire Naturelle., Museum National D'Histoire Naturelle. Pleurdeau, D., 2005. Human Technical Behavior in the African Middle Stone Age: The Lithic Assemblage of Porc Epic Cave (Dire Dawa, Ethiopia), African Archaeological Review 22, 177 197. Poisblaud, B., 2000. Preliminary study of lithic implements of the U.S. 1010 and comparisons with cuttings 7 and 9 of test pit 1. in: Gutherz, X., Jallot, G., Poisblaud, B., Sordoillet, D., Pouzolles, G., Asfaw, A., Lakew, B., and Aba, G. (Eds.), The Moche Borago Rock shelter: January February 2000 Exc avation. Report on File, SNNPR Office of Culture and Information, Awassa, Ethiopia., pp. 40 43. Prendergast, M.E., Luque, L., Dom+¡nguez Rodrigo, M., ez Mart+¡n, F., Mabulla, A.Z.P., and Barba, R., 2007. New excavations at Mumba rockshelter, Tanzania, Jou rnal of African Archaeology 5, 217 243. Reimer, P.J., Baillie, M.G.L., Bard, E., Bayliss, A., Beck, J.W., Blackwell, P.G., Ramsey, C.B., Buck, C.E., Burr, G.S., Edwards, R.L., Friedrich, M., Grootes, P.M., Guilderson, T.P., Hajdas, I., Heaton, T.J., Hogg, A.G., Hughen, K.A., Kaiser, K.F., Kromer, B., McCormac, F.G., Manning, S.W., Reimer, R.W., Richards, D.A., Southon, J.R., Talamo, S., Turney, C.S.M., van der Plicht, J., and Weyhenmeye, C.E., 2009. Intcal09 and Marine09 radiocarbon age calibration curves, 0 50,000 years cal BP, Radiocarbon 51, 1111 1150.
266 Revel, M., Ducassou, E., Grousset, F.E., Bernasconi, S.M., Migeon, S., Revillon, S., Mascle, J., Murat, A., Zaragosi, S., and Bosch, D., 2010. 100,000 Years of African monsoon variability recorded in sedim ents of the Nile margin, Quaternary Science Reviews In Press, Corrected Proof. Roberts, R.G., Galbraith, R.F., Laslett, G.M., Olley, J.M., and Yoshida, H., 1999. Optical dating of single and multiple grains of quartz from Jinmium Rock Shelter, northern Aus tralia: Part II, results and implications, Archaeometry 41, 365 395. Rogers, A.K., 2008. Obsidian hydration dating: accuracy and resolution limitations imposed by intrinsic water variability, Journal of Archaeological Science 35, 2009 2016. Rostek, F., Bar d, E., Beaufort, L., Sonzogni, C., and Ganssen, G., 1997. Sea surface temperature and productivity records for the past 240 kyr in the Arabian Sea, Deep Sea Research Part II: Topical Studies in Oceanography 44, 1461 1480. Ruddiman, W.F., 1977. Late Quatern ary deposition of ice rafted sand in the subpolar North Atlantic (lat 40° to 65°N), Geological Society of America Bulletin 88, 1813 1827. Sanchez Gone, M.F. and Harrison, S.P., 2010. Millennial scale climate variability and vegetation changes dur ing the Last Glacial: Concepts and terminology, Quaternary Science Reviews In Press, Corrected Proof. Sanchez Gone, M.F., Landais, A., Fletcher, W.J., Naughton, F., Desprat, S., and Duprat, J., 2008. Contrasting impacts of Dansgaard Oeschger events over a western European latitudinal transect modulated by orbital parameters, Quaternary Science Reviews 27, 1136 1151. Schulz, H., von, R., Erlenkeuser, H., and von, R., 1998. Correlation between Arabian Sea and Greenland climate oscillations of the past 110,000 years, Nature 393, 54 57. Shea, J.J., 2006. The origins of lithic projectile point technology: evidence from Africa, the Levant, and Europe, Journal of Archaeological Science 33, 823 846. Shinn, D.H. and Ofcansky, T.P., 2004. Historical Dictionary of Ethi opia, The Scarecrow Press, Lanham. Sordoillet, D. and Pouzolles, G., 2000. Stratigraphical and sedimentological survey of Moche Borago infilling. in: Gutherz, X., Jallot, L., Lesur J., Pouzolles, G., Sordoillet, D., Said, H., and Lakew, B. (Eds.), The Moch e Borago Rock shelter: November 2000 Excavation. Sowers, T., Bender, M., Labeyrie, L., Martinson, D., Jouzel, J., Raynaud, D., Pichon, J.J., and Korotkevich, Y.S., 1993. A 135,000 Year Vostok Specmap common temporal framework, Paleoceanography 8.
267 Stein, M. Goldstein, S.L., Schramm, M., Katz, A., and Starinsky, A., 1998. Lake Lisan sediments (paleo Dead Sea) as monitors of the last glacial history. PAGES Past Global Changes and their Significance for the Future University of London, p. 21. Stuiver, M. and Grootes, P.M., 2000. GISP2 Oxygen isotope ratios, Quaternary Research 53, 277 284. Stuiver, M. and Polach, H.A., 1977. Reporting of C 14 Data Discussion, Radiocarbon 19, 355 363. Svensson, A., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl Jensen, D., Davies, S.M., Johnsen, S.J., Muscheler, R., Parrenin, F., Rasmussen, S.O., Roethlisberger, R., Seierstad, I., Steffensen, J.P., and Vinther, B.M., 2008. A 60 000 year Greenland stratigraphic ice core chronology, Climate of the Past 4, 47 57. Svensson, A., Andersen, K.K., Bigler, M., Clausen, H.B., Dahl Jensen, D., Davies, S.M., Johnsen, S.J., Muscheler, R., Rasmussen, S.O., Rothlisberger, R., Steffensen, J.P., and Vinther, B.M., 2006. The Greenland ice core chronology 2005, 15 42 ka. Part 2: comparison to o ther records, Quaternary Science Reviews 25, 3258 3267. Thorne, A., Grun, R., Mortimer, G., Spooner, N.A., Simpson, J.J., McCulloch, M., Taylor, L., and Curnoe, D., 1999. Australia's oldest human remains: age of the Lake Mungo 3 skeleton, Journal of Human Evolution 36, 591 612. Tierney, J.E., Russell, J.E., Huang, Y., Damste, J.S.S., Hopmans, E.C., and Cohen, A.S., 2008. Northern hemisphere controls on tropical Southeast African climate during the past 60,000 Years, Science 1160485. Trigger, B., 1989. A His tory of Archaeological Thought, Cambridge University Press, New York. Van Campo, E., Duplessy, J.C., Prell, W.L., Barratt, N., and Sabatier, R., 1990. Comparison of terrestrial and marine Temperature estimates for the past 135 Kyr off Southeast Africa: A t est for Gcm simulations of paleoclimate, Nature (London) 348, 209 212. Van der Touw, J.W., Galbraith, R.F., and Laslett, G.M., 1997. A logistic truncated normal mixture model for overdispersed binomial data, Journal of Statistical Computation and Simulatio n 59, 349 373. Van Peer, P., 1998. The Nile corridor and the out of Africa model: an examination of the archaeological record, Current Anthropology 39, 115 140. Van Peer, P. and Vermeersch, P.M., 2000. Nubian complex and the dispersal of modern humans in N orth Africa. in: Krzyzaniak, L., Kroeper, K., and Kobusiewicz, M. (Eds.), Recent Research into the Stone Age of Northeastern Africa Poznan Archaeological Museum, Poznan, pp. 47 60.
268 Vermeersch, P.M., 2001. 'Out of Africa' from an Egyptian point of view, Qu aternary International 75, 103 112. Vidal, L. and Arz, H., 2004. Oceanic climate variability at millenial time scales: models of climate connections. in: Battarbee, R.W., Gasse, F., and Stickley, C.E. (Eds.), Past climate Variability through Europe and Afr ica Springer, Dordrecht, pp. 31 40. Vincent, A.S., 1984. Plant foods in savanna environments: a preliminary report of tubers eaten by the Hadza of northern Tanzania, World Archaeology 17, 131 147. Voelker, A.H.L., 2002. Global distribution of centennial s cale records for Marine Isotope Stage (MIS) 3: a database, Quaternary Science Reviews 21, 1185 1212. Wadley, L., 2001. What is cultural modernity? A general view and a South African perspective from Rose Cottage Cave, Cambridge Archaeological Journal 11, 2 01 221. Wang, Y.J., Cheng, H., Edwards, R.L., An, Z.S., Wu, J.Y., Shen, C.C., and Dorale, J.A., 2001. A High Resolution Absolute Dated Late Pleistocene Monsoon Record from Hulu Cave, China, Science 294, 2345 2348. Wayland, E.J., 1929. African Pluvial Perio ds and Prehistoric Man, Man 29, 118 121. Wayland, E.J., 1930. African Pluvial Periods and Prehistoric Man, Man 30, 236. Wayland, E.J., 1934. Rifts, rivers, rains and early Man in Uganda, The Journal of the Royal Anthropological Institute of Great Britain a nd Ireland 64, 333 352. Wayland, E.J. and Burkitt, M.C., 1932. The Magosian culture of Uganda, Journal of the Royal Anthropological Institute 1, 369 390. Weldeab, S., Lea, D.W., Schneider, R.R., and Andersen, N., 2007. 155,000 years of West African monsoon and ocean thermal evolution, Science 316, 1303 1307. Wendorf, F. and Schild, R., 1989. Summary and synthesis. in: Wendorf, F., Schild, R., and Close, A. (Eds.), The Prehistory of Wadi Kubbaniya, volume 3: Late Paleolithic Archaeology Southern Methodist Un iversity Press, Dallas Wendorf, F. and Schild, R., 1992. The Middle Paleolithic of North Africa: a status report. New light on the Northeast African past Heinrich Barth Institute, Cologne, pp. 42 78. Weninger, B. and Joris, O., 2008. A C 14 age calibrati on curve for the last 60 ka: the Greenland Hulu U/Th timescale and its impact on understanding the Middle to Upper Paleolithic transition in Western Eurasia, Journal of Human Evolution 55, 772 781. White, L., 1959. The Evolution of Culture, McGraw Hill, Ne w York. White, T.D., Asfaw, B., Degusta, D., Gilbert, H., Richards, G.D., Suwa, G., and Clark Howell, F., 2003. Pleistocene Homo sapiens from Middle Awash, Ethiopia, Nature 423, 742 747.
269 White, T.E., 1953a. A method of calculating the dietary percentage of various food animals utilized by aboriginal peoples, American Antiquity 18, 396 398. White, T.E., 1953b. Observations on the butchering technique of some aboriginal peoples, no. 2, American Antiquity 19, 160 164. Wiessner, P., 1982. Beyond willow smoke an d dogs' tails: a comment on Binford's analysis of hunter gatherer settlement systems, American Antiquity 47, 171 178. Wiessner, P., 1977. Hxaro: A regional system of reciprocity for reducing risk among the !Kung San. University of Michigan, Ann Arbor, Uni versity Microfilms Wiessner, P., 1980. Risk, reciprocity and social influence on !Kung San economies. in: Leacock, E.R. and Lee, R.B. (Eds.), Politics and History in Band Societies Cambridge University Press, London WoldeGabriel, G., Aronson, J.L., and Walter, R.C., 1990. Geology, geochronology, and rift basin development in the central sector of the Main Ethiopia Rift, Geological Society of America Bulletin 102, 439 458. Wolff, E.W., Chappellaz, J., Blunier, T., Rasm ussen, S.O., and Svensson, A., 2010. Millennial scale variability during the last glacial: The ice core record, Quaternary Science Reviews In Press, Corrected Proof. Woodburn, J., Lee, R.B., and DeVore, I., 1968. An introduction to Hadza ecology. Man the Hunter, pp. 49 55. Yellen, J., Brooks, A., Helgren, D., and Tappen, M., 2005. The Archaeology of Aduma Middle Stone Age Sites in the Awash Valley, Ethiopia, PaleoAnthropology 25 100. Yellen, J.E. and Harpending, H., 1972. Hunter gatherer populations and a rchaeological inference, World Archaeology 4, 244 253. Yellen, J.E., 1977. Long Term Hunter Gatherer Adaptation to Desert Environments: A Biogeographical Perspective, World Archaeology 8, 262 274. Zilho, J., 2006. Aurignacian, behavior, modern: issues of definition in the emergence of the European Upper Paleolithic. Towards a Definition of the Aurignacian, pp. 53 69. Zilho, J., Trinkaus, E., Constatin, S., Milota, S., Gherase, M., Sarcina, L., Danciu, A., Rougier, H., Quiles, J., and Rodrigo, R., 2007. The Pestera cu Oase people, Europe's earliest modern humans. in: P.Mellars, K.Boyle, O.Bar Yosef, and C.Stringer (Eds.), Rethinking the Human Revolution McDonald Institute, Cambridge
270 BIOGRAPHICAL SKETCH Erich Christopher Fisher was born Colorado. The oldest of two children, Erich and his family moved to Houston, Texas when he was in second grade. Although not a native, Erich considers Texas home. Erich received his B.A. in anthropology with a minor in cartography / GIS from Southwest Texas State Uni versity (known now as Texas State University, San Marcos) in 2002. That same year he began graduate school at the University of Florida and he earned his Master of Arts degree in 2005.