For the degree of Masters of Science THE LATER STONE AGE OCCUPATION AND SEQUENCE OF THE MAPUNGUBWE LANDSCAPE A dissertation submitted to the Faculty of Science, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Science Timothy Robin Forssman 0510124R University of the Witwatersrand Johannesburg 2010 ii DECLARATION I, Timothy Robin Forssman, declare that this is my own original work. It has been submitted for a Master of Science degree at the University of the Witwatersrand. It has not been submitted to any other academic institution. ............................................ Tim Forssman iii ?The question of questions for mankind ? the problem which underlies all others, and is more deeply interesting than any other ? is the ascertainment of the place which Man occupies in nature and of his relations to the universe of things.? Thomas Henry Huxley 1863 iv ABSTRACT Forager interactions with Bantu language-speaking farmers throughout southern Africa have yielded different outcomes. Attention has been paid to the way in which the foraging economy changed from the pre-contact into the contact period. On the Mapungubwe landscape this is particularly important as it is here where the first Iron Age state established itself. A series of excavations have been used to determine the forager sequence. However, it is shown here that this model excludes facets of foraging lifeways. Later Stone Age lithic scatters were identified during an archaeological survey. Sites were then selected for analysis from which a sample of artefacts was collected using a stratified unaligned sampling method and a timed collection. These data was then compared to the dated assemblages from excavations at Little Muck Shelter, Balerno Main Shelter, Balerno Shelter 2 and 3 and Tshisiku Shelter. In doing so, various discrepancies between shelter and open air assemblages are made evident. Namely, open air assemblages are generally dominated by quartz and lack the variety of formal tools found at shelters. In addition, shelter sites are dominated by crypto-crystalline materials. A comparison of two excavations echoes these patterns. Den Staat AB 32 is an open air site and compares well with open air assemblages, whereas a neat relationship between Mbere Shelter and other shelter excavations exists. Therefore, sites are grouped together based on similarities between their assemblages. They are also placed into date brackets established using typological cross-referencing with the dated assemblages. Using these dates, it has been shown that forager mobility was not inhibited by the Iron Age settlement of the area. It seems more likely that foragers were selecting sites in order to interact with farmers during certain periods and maintaining their autonomy during others. It is suggested that quartz dominated sites may represent a movement towards or into farmer homesteads as they are mostly located in the zone with the highest density of farmer settlements. Alternatively, these sites may be the result of variable activity patterns at special purpose sites. The findings presented here suggest that a reassessment of the forager record is needed. Open air sites need to be included in forager studies as our understanding of the forager occupation of the Mapungubwe landscape is at present incomplete. v ACKNOWLEDGEMENTS Many people have played a role in this project: Karim Sadr: Your patience and help have been much appreciated. Without your grounding advice this project would not be as resounding. Thomas Huffman: I have considered your advice along the way at each step. Thank you for it. Warwick Davies-Mostert and De Beers Consolidated Mines: Thank you for accommodation on the Venetia Limpopo Nature Reserve. To the following who allowed me permission to conduct my field work on their farms: Warwick Davies-Mostert and De Beers Consolidated Mines for Venetia Limpopo Nature Reserve and Little Muck, the Venter family for Hackthorne and Steve Wigg for Halcyon Game Lodge. Thanks also to Kathy Kuman and Thomas Huffman for letting me analyse the stone tools from Mbere Shelter and Den Staat AB 32 respectively. Some people gave me comments and assistance on some of the chapters: Kath Potgeiter, Matt Lotter, Ryan J. Gibbon and D?cio Jose Muianga; your insights are much appreciated. The Palaeontological Scientific Trust (PAST), the National Research Fund (NRF) and the University of Witwatersrand provided funding for this project. Some people have helped me along the way: John ?Danger? Power, thanks for driving me around at times and giving me all the data I asked for and to Wendy Collinson for doing the odd bit of driving. I hope I didn?t talk your ears off. Joel le Baron and Kath Potgeiter thanks for helping me in the field. Thank you to Thomas Huffman and Marilee Wood for analysing the ceramics and beads respectively. For all the fun, thanks Matt, Joel, Dom, Decio, Mike, Justin, John, Wendy, Kath, BIKE4BEASTS and GVI tourism. Special thanks go to Kath Potgeiter again, thank you for guidance, reading my work, helping me write it and pretending my yarns were interesting, you?re listening and advice helped me sound my arguments. Very importantly, I would like to thank my family for bearing my enthusiasm, worrying about my safety while walking in the bush and for your constant support. To all of you I am so indebted. vi TABLE OF CONTENTS Declaration ii Abstract iv Acknowledgements v Table of Contents vi List of Figures iv List of Table x CHAPTER one Introduction 1 CHAPTER two Literature Review 4 2.1. The Mapungubwe landscape 4 2.2. Foragers and their changing world 6 2.2.1. Changing views of southern African foragers 6 2.2.2. Pre-contact foragers 8 2.2.3. Early contact: the arrival of farmers 10 2.2.4. Late contact: the Middle Iron Age 13 2.3. The forager cultural record of the Mapungubwe landscape 16 2.3.1. Early pre-contact: 6000 BC to 1220 BC 16 2.3.2. Late pre-contact: 1220 BC to AD 100 17 2.3.3. Early contact: AD 100 to AD 900 18 2.3.4. Zhizo contact: AD 900 to AD 1010 19 2.3.5. Leopard Kopje contact: AD 1010 to AD 1300 20 2.3.6. A summary of the forager record 21 2.4. Open air archaeology 22 2.4.1. Survey method and surface collection 23 2.4.2. Chronological control of open air sites 25 2.4.3. Geographic information systems (GIS) 26 2.5. Summary 26 CHAPTER three Method 27 3.1. Research design 27 3.2. Research area 27 3.3. Survey method 29 3.4. Forager occupation 30 3.5. Surface analysis 30 3.2. Mbere Shelter 31 3.3. Den Staat AB 32 32 3.4. Data analysis and comparison with the dated sequence 32 CHAPTER four Analysis 34 4.1. The Holocene occupation of the Mapungubwe landscape 34 4.1.1. Survey results 34 4.1.2. Critical variables and site selection 37 4.2. Holocene open air site analysis 40 vii 4.2.1. Lithic artefacts from the grid analysis 42 4.2.1.1. Raw material 42 4.2.1.2. Formal tools 42 4.2.1.3. Formal tool's raw material 45 4.2.2. Lithic artefacts from the timed analysis 46 4.2.3. Non-lithic artefacts 47 4.2.3.1. Beads 47 4.2.3.2. Ceramics 48 4.2.3.3. Copper 49 4.2.3.4. Other artefacts 49 4.2.4. Comparison of open air assemblages with the dated sequence 50 4.2.4.1. The geographical setting of open air site clusters 50 4.2.4.2. Site clusters: raw materials, formal tools and formal tool?s raw material 53 4.2.5. Undated assemblages and the dated sequence 57 4.2.5.1. Raw material 57 4.2.5.2. Formal tools 58 4.2.5.3. Formal tool's raw material 59 4.2.5.4. Conceptual control 59 4.2.5.5. Forager and farmer spatial relationship 60 4.2.5.6. Beads and ceramics 63 4.3. Analysis of Mbere Shelter and Den Staat AB 32 65 4.3.1. Mbere Shelter 65 4.3.1.1. Lithic artefacts 65 4.3.1.2. Non-lithic artefacts 65 4.3.1.3. Sequence 65 4.3.2. Den Staat AB 32 66 4.3.2.1. Lithic artefacts 66 4.3.2.2. Non-lithic artefacts 66 4.3.2.3. Sequence 66 4.3.3. Mbere Shelter, Den Staat AB 32 and the dated sequence 72 CHAPTER five Discussion 74 5.1. Forager camps on the Mapungubwe landscape 74 5.2 The significance of Holocene surface scatters on the Mapungubwe landscape 76 5.3. Disparities in the excavated record 81 5.3.1. Mbere Shelter 81 5.3.1.1. Comparison with the dated sequence 82 5.3.1.2. Comparison with open air assemblages 82 5.3.2. Den Staat AB 32 83 5.3.2.1. Comparison with the dated sequence 84 5.3.2.2. Comparison with open air assemblages 85 CHAPTER six Conclusion and Recommendations 86 viii CHAPTER seven References 89 Appendix A 99 Appendix B 103 Appendix C 109 Appendix D 152 ix LIST OF FIGURES 2.1 The Mapungubwe landscape in northern Limpopo, South Africa 5 3.1 The survey zone 28 3.2 Environmental zones in the survey zone 29 4.1 Identified lithic scatters in the survey zone 35 4.2 Analysed sites and the vegetation map of De Beers Venetia Limpopo Nature Reserve 35 4.3 Surface lithic scatter distribution in the environmental zones 36 4.4 Distribution of analysed sites 37 4.5 Distribution of lithic scatter locations 38 4.6 Distribution of LSA site locations 38 4.7 Terrain ruggedness and non-perennial water courses 39 4.8 Site distribution 40 4.9 Artefacts from the MSA site (no. 144) 41 4.10 CCS and quartz dominated sites 51 4.11 CCS and quartz dominated sites and farmer settlements 52 4.12 Early pre-contact period LSA sites 61 4.13 Late pre-contact period LSA sites and Zhizo settlements 61 4.14 Early contact period LSA sites and Zhizo settlements 62 4.15 Zhizo contact period LSA sites and Zhizo settlements 62 4.16 Leopard Kopje contact period LSA sites and Leopard Kopje settlements 63 4.17 Bead details from Mbere Shelter 68 4.18 Ceramic details from Mbere Shelter 69 4.19 Lithic distribution at Mbere Shelter 70 x LIST OF TABLES 4.1 Site distribution 36 4.2 Sites selected intentionally for analysis 41 4.3 Site clusters based on raw material 43 4.4 Diagnostic formal tools 44 4.5 Diagnostic formal tools raw material 45 4.6 All raw material versus formal tool's raw material 46 4.7 Bead provenance 47 4.8 Earthenware provenance 49 4.9 Other artefacts at forager sites 50 4.10 Raw material site clusters 51 4.11 Location and cover of CCS and quartz dominated sites 53 4.12a-c Open air site clusters: a) CCS, b) quartz and c) multi-component sites 55 4.13 Chronological variable assessment 58 4.14 Formal tools from the timed analysis 59 4.15 Ceramic, bead and lithic time frame relationships 64 4.16 Mbere Shelter analysis 67 4.17 Den Staat AB 32 analysis 71 4.18 Critical variables from Mbere Shelter compared to the dated sequence 72 4.19 Critical variables from Den Staat AB 32 compared to the dated sequence 73