Archaeology of Mound-Clusters in West Africa 9781407313023, 9781407342672

Table of contents :
Front Cover
Title Page
Copyright
Table of Contents
List of Figures
List of Tables
Acknowledgements
Introduction
Chapter 1: West African Mound-Clusters: Review of the Evidence
Chapter 2: The Mouhoun Bend: Land, Humans, and Archaeological Survey
Chapter 3: Kebe-Sira-Kan-Tomo: Late Stone Age Lithic Scatters
Chapter 4: Diekono Settlement Complex
Chapter 5: Tora-Sira-Tomo Settlement Complex
Chapter 6: Kerebe-Sira-Tomo Settlement Complex
Chapter 7: Human Skeletal Remains: Taphonomy, Demography, and Palaeopathology
Chapter 8: Mound Dwelling Traditions of the Mouhoun Bend
References
Annex 1: Mouhoun Bend Skeleton Summaries
Annex 2: Radiocarbon Dates from the Mouhoun Bend Archaeological Project

Citation preview

Cambridge Monographs in African Archaeology 87

Series Editors: Laurence Smith, Brian Stewart and Stephanie Wynne-Jone

Archaeology of Mound-Clusters in West Africa Augustin F. C. Holl

BAR International Series 2660 2014

ISBN 9781407313023 paperback ISBN 9781407342672 e-format DOI https://doi.org/10.30861/9781407313023 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

Table of Contents

Acknowledgements Introduction Chapter 1 West African Mound-Clusters: Review of the Evidence Chapter 2 The Mouhoun Bend: Land, Humans, and Archaeological Survey Chapter 3 Kebe-Sira-Kan-Tomo: Late Stone Age Lithic Scatters Chapter 4 Diekono Settlement Complex Chapter 5 Tora-Sira-Tomo Settlement Complex Chapter 6 Kerebe-Sira-Tomo Settlement Complex Chapter 7 Human Skeletal Remains: Taphonomy, Demography, and Palaeopathology Chapter 8 Mound Dwelling Traditions of the Mouhoun Bend References Annex 1: Mouhoun Bend Skeleton Summaries K. C. Maes and P. L. Walker Annex 2: Radiocarbon Dates from the Mouhoun Bend Archaeological Project

i

x 1 3 11 17 22 37 112 169 178 188 191

195

List of Figures

Figure 1.1: Map of West Africa with the location of mound sites mentioned in the text Figure 2.1: The Study area Figure 2.2: Patterns of land-use around Douroula Figure 2.3: Network of fossil channels in the Mouhoun river meander Figure 2.4: The mining shafts from Tissi Figure 2.5: View of the galleries from the iron ore mine of Tissi Figure 4.1: Map of Diekono Figure 4.2: Stratigraphy of Diekono 1, west section Figure 4.3: Stratigraphy of Diekono 1, north section Figure 4.4: Occupation II floor Figure 4.5: Occupation III floor Figure 4.6: Occupation IV floor Figure 4.7: Occupation V floor Figure 4.8: Pottery from Occupation II and III Figure 4.9: Pottery from Occupation IV and V Figure 4.10: Coarse lithic artifacts Figure 4.11: Coarse lithic artifacts Figure 4.12: Iron artifacts Figure 4.13: Iron arm-ring fragment Figure 4.14: Occupation I floor Figure 4.15: Burial with large sherds shroud Figure 4.16: The deceased buried in tightly flexed position Figure 4.17: Occupation II floor Figure 4.18: Occupation III floor Figure 4.19: Occupation IV floor Figure 4.20: Pottery from Occupation I Figure 4.21: Pottery from Occupation II Figure 4.22: Pottery from Occupation III Figure 4.23: Lithic artifacts Figure 4.24: Iron artifacts Figure 4.25: Diekono 3, the coarse stone artifacts Figure 4.26: Distribution by weight of container 1 rocks Figure 4.27: Distribution by weight of container 2 rocks Figure 4.28: Diekono 3 iron artifacts Figure 5.1: Map of Tora-Sira-Tomo settlement complex Figure 5.2: Contour map of Tora-Sira-Tomo 1 Figure 5.3: Map of the excavation of Tora-Sira-Tomo 1 Figure 5.4: Stratigraphy of TST 1 probe A Figure 5.5: View of the vessels at TST 1 Figure 5.6: Vessels from TST 1 probe A Figure 5.7: Vessels from TST 1 probe B ii

3 11 12 13 13 14 22 22 23 23 23 24 24 26 26 26 26 27 27 28 28 28 28 29 29 29 31 31 31 32 32 34 35 36 37 37 38 38 38 39 39

Figure 5.8: Vessels from TST 1 probe B Figure 5.9: Stratigraphy of TST 1 probe C Figure 5.10: The top of TST 1 furnace in probe C Figure 5.11: The tuyeres arrangement in the furnace Figure 5.12: The furnace mouth on its east flank Figure 5.13: The underground tuyeres firing chamber at the bottom of the furnace Figure 5.14: Section of the base of TST 1 furnace below the tuyeres level Figure 5.15: Stratigraphy of the slag heap in probe D Figure 5.16: Miscellaneous finds from probe D Figure 5.17: TST 2, the quarry section Figure 5.18: Stratigraphy of TST 3-East Figure 5.19: Occupation I floor Figure 5.20: Sherds from Occupation I Figure 5.21: Occupation II floor Figure 5.22: Occupation II forge furnace Figure 5.23: Pottery from Occupation II Figure 5.24: Occupation III floor Figure 5.25: Concentration of vessels in Occupation III house floor Figure 5.26: Occupation III material culture Figure 5.27: Occupation IV floor Figure 5.28: Pottery from Occupation IV Figure 5.29: Occupation V floor Figure 5.30: Occupation V material culture Figure 5.31: The burial from Occupation IV Figure 5.32: Occupation VI material culture Figure 5.33: Stratigraphy of TST 3-West, East section Figure 5.34: Occupation I floor Figure 5.35: View of burnt features from Occupation I Figure 5.36: Unfired clay balls in their container Figure 5.37: Pottery from Occupation I Figure 5.38: Pottery from Occupation I Figure 5.39: Pottery from Occupation I Figure 5.40: Pottery from Occupation I Figure 5.41: Pottery from Occupation I Figure 5.42: Coarse lithic artifacts from Occupation I Figure 5.43: Occupation II floor Figure 5.44: Occupation III floor Figure 5.45: Bone and iron artifacts from Occupation III Figure 5.46: Occupation IV floor Figure 5.47: Occupation V floor Figure 5.48: View of the “cached” copper arm-rings Figure 5.49: Metal artifacts from Occupation V Figure 5.50: Occupation VI floor Figure 5.51: Burnt cowpeas (Vigna unguiculata) from Occupation II Figure 5.52: Burnt roof beams from occupation IV Figure 5.53: Occupation I floor Figure 5.54: View of the weaving/cloth dyeing workshop Figure 5.55: Pottery from Occupation I

iii

39 40 40 41 42 42 42 43 44 44 45 45 46 46 46 46 46 47 48 49 49 49 51 51 52 53 54 54 54 55 55 56 56 57 57 58 58 58 59 59 59 59 59 60 61 62 62 63

Figure 5.56: Pottery from Occupation I Figure 5.57: Occupation II floor Figure 5.58: Stratigraphy of TST 5 Figure 5.59: Occupation I floor Figure 5.60: The burial found in Occupation I deposit Figure 5.61: Occupation II surface Figure 5.62: Occupation III surface Figure 5.63: Plan and section of Occupation III storage pit Figure 5.64: Occupation IV floor Figure 5.65: Material culture from Occupation III Figure 5.66: Material culture from Occupation IV Figure 5.67: Stratigraphy of TST 6 Figure 5.68: Occupation I floor Figure 5.69: Occupation II floor Figure 5.70: Occupation III floor Figure 5.71: Occupation IV floor Figure 5.72: Material culture from TST 6 Figure 5.73: Stratigraphie of TST 7 Figure5.74: Occupation I floor Figure 5.75: View of the vessels arrangement Figure 5.76: Occupation I burials Figure 5.77: View of the burials Figure 5.78: Occupation II Floor Figure 5.79: Burial in occupation III deposit Figure 5.80: Pottery from occupation I Figure 5.81: Pottery from occupation II Figure 5.82: Coarse lithic artifacts Figure 5.83: Iron artifacts Figure 5.84: Stratigraphy of TST 8 Figure 5.85: Occupation I floor Figure 5.86: Occupation IIa floor Figure 5.87: Occupation IIb floor Figure 5.88: Occupation IIc floor Figure 5.89: Occupation III floor Figure 5.90: Material culture from TST 8 Figure 5.91: Stratigraphy of TST 9 Figure 5.92: Excavation of TST 9 Figure 5.93: Stratigraphy of TST 9 Figure 5.94: View of the secondary burial from the central cluster Figure 5.95: View of the southern cluster with burial 10, 11, and 12 Figure 5.96: Pottery from TST 9 Figure 5.97: Pottery from TST 9 Figure 5.98: Pottery from TST 9 Figure 5.99: The occupation surface of TST 10 Figure 5.100: View of TST 10 occupation floor I Figure 5.101: Occupation I at TST 12 Figure 5.102: Occupation II at TST 12 Figure 5.103: Material culture of TST 12

iv

63 64 65 65 66 66 66 67 67 69 69 70 71 71 71 72 72 74 74 74 75 75 77 77 77 78 79 79 80 80 80 81 81 81 82 82 84 84 85 85 86 87 88 89 90 90 90 91

Figure 5.104: Stratigraphy of TST 13 Figure 5.105: Occupation I floor Figure 5.106: Occupation II floor Figure 5.107: Material culture of TST 13 Figure 5.108: Stratigraphy of TST 14 Figure 5.109: TST 14 occupation floor Figure 5.110: Coarse stone artifacts Figure 5.111: Coarse stone and iron artifacts Figure 5.112: Stratigraphy of TST 15 Figure 5.113: Occupation I floor Figure 5.114: Occupation II floor Figure 5.115: Pottery from TST 15 Figure 5.116: Coarse stone artifacts Figure 5.117: Iron artifacts Figure 5.118: Additional iron artifacts Figure 5.119: Stratigraphy of TST 16 Figure 5.120: Occupation I floor Figure 5.121: Occupation II floor Figure 5.122: Material culture of TST 16 Figure 5.123: Stratigraphy of TST 17 Figure 5.124: Occupation I floor Figure 5.125: Tora-Sira-Tomo in phase I and II Figure 5.126: Tora-Sira-Tomo in phase III Figure 5.127: Tora-Sira-Tomo in phase IV Figure 5.128: Tora-Sira-Tomo mounds size in phase IV Figure 5.129: Map of Gnambakouon Sira Tomo Figure 5.130: Stratigraphy of GST 3 Figure 5.131: Occupation I floor Figure 5.132: Pottery from GST 3 Figure 5.133: GST 3 pottery: rims Figure 6.1: Map of Kerebe-Sira-Tomo settlement complex Figure 6.2: KST 1A: view of the circular installation Figure 6.3: Plan of the circular installation Figure 6.4: The pottery firing “kiln” at Tora Figure 6.5: The karité nuts grinding platform of Koussiri Figure 6.6: Pottery from KST 1A sample 1 Figure 6.7: Pottery from KST 1A sample 2 Figure 6.8: Stratigraphy of KST 1B Figure 6.9: Occupation I floor Figure 6.10: Occupation II floor Figure 6.11: Occupation II burial Figure 6.12: View of occupation III floor Figure 6.13: Occupation III floor Figure 6.14: Pottery from Occupation I Figure 6.15: Pottery from Occupation II Figure 6.16: Coarse stone artifacts from Occupation II Figure 6.17: Pottery from Occupation III Figure 6.18: Pottery from Occupation III

v

92 93 93 94 95 95 96 97 97 98 99 100 100 101 101 101 102 102 103 103 104 106 107 107 108 109 109 110 110 111 112 113 113 113 113 115 115 116 116 116 117 117 117 119 119 120 121 121

Figure 6.19: Coarse stone artifacts from Occupation III Figure 6.20: Pottery from the eroded Occupation IV Figure 6.21: View of KST 2, the quarry Figure 6.22: View of the laterite deposits Figure 6.23: Section of the quarry Figure 6.24: Stratigraphy of KST 3 Figure 6.25: Profile of Occupation I laterite floors Figure 6.26: Occupation I floor Figure 6.27: Occupation II floor Figure 6.28: Occupation III floor Figure 6.29: Occupation IV Figure 6.30: Pottery from KST 3 Figure 6.31: Base of a tri-legged bowl Figure 6.32: Iron artifacts Figure 6.33: Coarse stone artifacts from Occupation I Figure 6.34: Coarse stone artifacts from Occupation II and III Figure 6.35: Stratigraphy of KST 4 Figure 6.36: Plan of Occupation II features Figure 6.37: View from the north of the exposed occupation II features Figure 6.38: View of house H-1 and H-2 with a burial Figure 6.39: View of the eastern hut cluster Figure 6.40: View of the chicken egg-nest Figure 6.41: View of the food storage vessels Figure 6.42: View of the burnt cowpeas Figure 6.43: View of the burnt fonio Figure 6.44: Storage vessels Figure 6.45: Storage vessels Figure 6.46: View from the north of the northeastern hut cluster Figure 6.47: Concentration of storage vessels in house H-E9 Figure 6.48: Vessels from H-E5 Figure 6.49: Vessels from H-E3 Figure 6.50: Vessels from H-E10, H-E11, and the courtyard Figure 6.51: Vessels from H-E7 Figure 6.52: Coarse stone tools from occupation II Figure 6.53: Occupation III surface Figure 6.54: Pottery from Occupation III Figure 6.55: Additional occupation III pottery Figure 6.56: Occupation III iron artifacts Figure 6.57: Pot lids, basins and bowls bases Figure 6.58: Grinders and grindstones from H-E3 Figure 6.59: Grindstones from H-E5 Figure 6.60: Iron artifacts from KST 4 and KST 5 Figure 6.61: Stratigraphy of KST 5 Figure 6.62: Occupation I floor Figure 6.63: Occupation II floor Figure 6.64: Occupation III floor Figure 6.65: Occupation IV floor Figure 6.66: Occupation V floor

vi

123 123 124 124 124 125 125 126 126 126 129 129 130 130 130 131 131 132 132 133 134 134 134 135 135 135 136 136 137 137 138 138 139 140 140 141 142 143 146 147 147 147 148 148 148 149 149 150

Figure 6.67: Material culture from KST 5 Figure 6.68: Occupation I floor Figure 6.69: Occupation II floor Figure 6.70: Pottery from Occupation I Figure 6.71: Pottery from Occupation II Figure 6.72: Pot-lid and coarse stone artifacts Figure 6.73: Contour map of KST 7 Figure 6.74: Plan of KST 7 blacksmith workshops Figure 6.75: The furnaces sections Figure 6.76: Stratigraphy of KST 8 Figure 6.77: Plan of the excavation Figure 6.78: Plan and section of the furnace Figure 6.79: View of KST 8 furnace Figure 6.80: Pottery from KST 8 Figure 6.81: Plan of KST 9 Figure 6.82: View of the furnace fill top Figure 6.83: View of the accumulation of tuyeres Figure 6.84: View of the furnace bottom Figure 6.85: Map of KST 10 furnace Figure 6.86: View of the top of the furnace Figure 6.87: View of the furnace bottom Figure 6.88: Plan of the bottom of furnace 10 Figure 6.89: Pottery from KST-11 Figure 6.90: Kerebe-Sira-Tomo in phase I and II Figure 6.91: Kerebe-Sira-Tomo in phase III Figure 6.92: Kerebe-Sira-Tomo in phase IV

151 152 152 154 155 156 156 157 158 160 160 161 161 162 163 164 164 164 164 164 165 165 166 167 168 168

vii

List of Tables

Table 2.1: Mouhoun Bend Settlement survey Table 3.1: Composition of LSA lithic assemblages Table 3.2: Frequencyd istribution of raw materials Table 3.3: Patterns of use of lithic raw material Table 3.4: Cores measurements Table3.5: Distribution of formal tools Table 3.6: Mensurations of endscarpers Table 3.7: Mensurations of sidescrapers Table 3.8: Mensurations of microliths Table 3.9: Mensurations of points Table 3.10: Mensurations of sidescrapers Table 3.11: Mensurations of microliths Table 4.1: Diekono-I: general distribution of sherds Table 4.2: Faunal remains from Diekono-I Table 4.3: General characteristics of Diekono-II sherds Table 4.4: Faunal remains from Diekono-II Table 4.5: Distribution of stone artifacts from container 1 Table 4.6: Distribution of stone artifacts from container 2 Table 4.7: Distribution of sandstone artifacts into weight classes Table 5.1: TST-1: General distribution of sherds in Probe C Table 5.2: General distribution of potsherds from TST-3-East Table 5.3: General distribution of decoration techniques Table 5.4: General distribution of sherds thickness (mm) Table 5.5: TST-3-West: General distribution of sherds Table 5.6: TST-3-West: distribution of faunal remains Table 5.7: Mensurations of pottery from TST-4 Table 5.8: Distribution of TST-5 sherds Table 5.9: Characteristics of the reconstructed vessels of TST-5 Table 5.10: Faunal remains from TST-5 Table 5.11: General distribution of potsherds from TST-6 Table 5.12: General distribution of potsherds from TST - 7 Table 5.13: Characteristics of TST-7 vessels Table 5.14: General distribution of potsherds from TST-8 Table 5.15: General distribution of potsherds from TST-9 Table 5.16: General distribution of potsherds from TST-10 Table 5.17: General distribution of potsherds from TST-10 Table 5.18: General distribution of potsherds from TST-13 Table 5.19: General distribution of sherds from TST-14 Table 5.20: Tora-Sira-Tomo mound cluster at phase I (650 BC – AD 800) Table 5.21: Tora-Sira-Tomo mound cluster in phase II (AD 800 – 1100) Table 5.22: Tora-Sira-Tomo mound cluster in phase III (AD 1100 – 1400) viii

15 17 18 18 18 20 20 20 20 21 21 21 25 27 30 33 34 34 35 41 50 50 50 60 61 64 68 68 70 73 76 78 82 88 89 92 94 96 99 104 105

Table 5.23: Tora-Sira-Tomo mound cluster in phase IV (AD 1400 – 1650) Table 6.1: Characteristics of KST-1A sherds’ population Table 6.2: Characteristics of KST-1B sherds’ population Table 6.3: Measurements of vessels from KST-1B Table 6.4: Characteristics of the sherds’ population from KST-3 Table 6.5: Faunal remains from KST-3 Table 6.6: Kerebe-Sira-Tomo 4: Patterns of distribution of vessels from Occupation II and III Table 6.7: General characteristics of the sherds population from Kerebe-Sira-Tomo 4 Table 6.8: Kerebe-Sira-Tomo 4: Characteristics of the sherds population from Occupation II Table 6.9: Kerebe-Sira-Tomo 5: Characteristics of the sherds population Table 6.10: KST-6: Characteristics of the sherds population Table 6.11: KST-6: Vessels mensurations Table 6.12: KST-7: Samples of iron production by-products Table 6.13: KST-8: Samples of iron production by-products Table 6.14: KST-9: Samples of iron production by-products Table 6.15: KST-10: Samples of iron production by-products Table 6.16: KST-11: Samples of iron production by-products Table 7.1: Taphonomic condictions of the Mouhoun Bend skeletons sample Table 7.2: Demography of the Mouhoun Bend skeletons sample Table 7.3: Pathologies in the Mouhoun Bend skeletons sample

ix

105 114 118 120 122 127 128 144 145 150 153 157 159 162 165 166 167 171 171 172

Acknowledgements

The archaeological project carried out on the mound complexes of the Mouhoun bend in northwestern Burkina Faso started with funding from the Centre National de la Recherche Scientifique. At this initial step, this allowed for a preliminary survey and probe of two mounds. Support from the National Geographic Society committee for Research and Exploration (Grant # 6378-98) and the University of California, San Diego allowed for the organization of two additional field expeditions. The University of Ouagadougou and the Centre National de la Recherche Scientifique et Technique (CNRST) of the Burkina Faso Republic not only facilitated the operations of the project, delivered research permits, but were also strongly supportive of the new initiative. The project also benefited from the logistic support of the French Mission de Cooperation in 1997. I wish to express my sincere gratitude to all those who made this project possible, through direct or indirect support. A number of students participated to the field operations: Fabienne Scouflaire from the university of Paris X-Nanterre, Deanda Johnson, Stephen Dueppen, Solomon Kuah, Kelly Kirby, and Jenny Pao from the University of California, San Diego, Mariama and Alice Coulibaly from the University of Ouagadougou. Dr Lassina Kote from the archaeology program of the University of Ouagadougou played a crucial role in the organization of all the field operations and made life in Douroula as smooth and pleasant as possible for the field crews. Finally, I wish to thank the workmen from Douroula for their help during our three field seasons in the Mouhoun Bend

x

Introduction

The work reported in this monograph is the result of three field seasons and several years of laboratory research. The first field season, funded by a grant from the CNRSLaboratoire d’Ethnologie, took place in September – October 1997. The season was devoted to an extensive survey of the study area aimed at selecting the most promising localities for further in-depth investigations. The survey techniques used were partly judgmental, keyed around sites well known by local informants, but also aided by a reliance on low altitude air-photographs. The valley of the Mouhoun River which delineates the northern boundary of the study area was intensively surveyed in search for evidence of Late Stone Age foragers’ presence. The settlement complexes, at Gnambakuon-Sira-Tomo, Diekono, Tora-Sira-Tomo, and Kerebe-Sira-Tomo, were selected for further research. The field season ended abruptly because of a serious roll-over car accident. Fortunately, none of the research crew members had a life-threatening injury. I had several bruises on the head and arms, and suffered from abdominal pain for a number of days without any consequence. The following two field seasons in 1999 and 2000, organized from the University of California San Diego, took place in winter. They were almost entirely devoted to the excavation of the selected settlement complexes, particularly at Diekono, Tora-SiraTomo, and Kerebe-Sira-Tomo.

in this book, even if preliminary in nature, are particularly interesting. 1.2 A comprehensive approach A clear presentation of archaeological evidence is crucial for the possibility of an independent assessment of the theoretical perspectives and conclusions achieved in a particular research project. The comprehensive approach adopted in the Mouhoun Bend Archaeological Project is historical and anthropological. It is articulated at four nested levels: the mound, the settlement complex [moundsclusters], the region, and the sphere of long-distance interaction. The goal is to figure out how the ancient inhabitants of the Mouhoun Bend lived and made sense of their “daily” lives. At level 1, the individual mound, each occupation episode is considered as a specific moment in the ever shifting decision-making processes humans have to deal with on a daily basis. The construction of dwelling features, the selection, and procurement of material culture items like pottery, stone tools, and subsistence elements, are part of constant negotiation and re-negotiation of social relationships. The build-up of the site cultural deposit and the total length of its occupation are also assessed systematically.

1.1 The Mouhoun Bend Archaeological Project

At level 2, the focus shifts to the changing patterns of development of each settlement complex. Each of the investigated complex is made of a varying number of mounds, some used for habitation and others used for specific activities like iron-smelting, karité oil production, or weaving and cloth dyeing. Theoretically, a fine-grained chronology allows to trace the developmental trajectory of the settlement complex, from the initial phase of occupation to the time of abandonment. It is at this level that issues of occupational segregation can be discussed with sound empirical footing.

The Mouhoun Bend Archaeological Project (NW BurkinaFaso) goal is to investigate the settlement history, the development of food-producing economies, and the emergence and amplification of craft specialization in this small portion of West Africa. Prior to this research project, this area was literally an archaeological Terra Incognita. The Mouhoun River (formerly known as Black Volta) flows from the SW to NE, then winds its course in a U-shape bend to follow an almost N-S direction. Surveys operations and small-scale excavations were implemented in parallel. The area archeological record consists essentially of moundsclusters, a special settlement type which is still poorly understood. Such settlements are found in different parts of West Africa (Connah 1981; Mcintosh 1998; Filipowiak 1979; Holl 2002; Raimbault & Sanago 1991; Togola 1996). Analyses of human remains and palaeopatholgical assessments are at an initial stage and the results presented

Level 3 focuses on regional dynamics. It examines sitelocation and settlement patterns and investigates settlement complexes interaction. A more or less regular distribution of large settlement complexes suggests the existence of autonomous and self-sustaining villages. Their mutual relations certainly changed through time. They could have

1

Archaeology of West African Mound-Clusters

1.3 Organization of the book

ranged from alliance, rivalry, to hostility, shifting back and forth depending on circumstances.

The book is divided into eight chapters. Chapter 1 presents a discussion on the archaeology of West African mounds clusters. A number of cases studied in Cameroons/Nigeria, Guinea, Mali, and Niger are reviewed and summarized. As presented above, the comprehensive approach advocated in this book focuses on the systematic probing of each mound part of a larger complex under investigation. The aim is to work out the finest-grained chronology possible to allow for a better understanding of the history of the settlement under investigation. Chapter 2 examines the major characteristics of the Mouhoun Bend study area and presents the surveyed and excavated sites complexes. Chapter 3 discusses evidences from a series of Late Stone Age (LSA) lithic scatters found in the Mouhoun river valley. Chapter 4 presents the evidence from Diekono settlement complex which is located in the Mouhoun River flood-plain. Data from the 20 mounds complex of ToraSira-Tomo and its satellite Gnamba-Kuon-Sira-Tomo are presented in chapter 5. Chapter 6 discuses evidence from Kerebe-Sira-Tomo. Chapter 7 considers the material from the excavated burials and presents a systematic analysis of human skeletal remains. And finally, chapter 8 takes on a general comparative approach to mound clustering processes.

Finally, level 4 considers the evidence for long-distance interaction collected from the excavated sites. This material points to links with the Upper-Niger and the Inland Niger delta. The wider social and cultural implications of these circulated objects are however difficult to grasp. The variability of the recorded material culture is constantly scrutinized. Pottery shapes and decoration do not mate and cross-breed. It is the potters who, trained in one tradition or another, and faced with specific demands, make decisions about the kinds of wares that have better opportunity to satisfy the latent demands. The same holds for stone tools, and houses. The production and consumption of material culture are essentially socially mediated. Consequently, in this book, the data from each of the recorded occupation episode are presented as systematically as possible. The goal is to provide the reader with a solid and coherent data base that can be used in an independent assessment of the accuracy of the conclusions reached in this book. Despite the shifting fashions in anthropology and anthropological archaeology and the difficulties for its full implementation, the Popperian conjecture and refutation rationale to scientific inferences is still the most parsimonious methodological approach.

2

Chapter 1 West African Mound-Clusters: Review of the Evidence

Introduction Depending on their size and constituent materials, mounds result from the piling up on the same spot of human occupation by-products. They include habitation features, craft installations, collapsed building material as well as discarded and abandoned material culture. Mounds are therefore exclusively human-made and, depending on circumstances, can be either well preserved or significantly disturbed by erosion agencies. Habitation mounds dating from the Late Stone Age onwards have been recorded in different parts of West Africa, from the Chadian plain in the east to the modern states of Mali and Guinea in the west (Bedaux et al 2001; Connah 1981; Dueppen 2008; Filipowiak 1979; Gado 1993; 2004; Holl 2002; 2004; 2006; Polet 2004; Mcintosh & Mcintosh 1980; R. Mcintosh 2002; S. Mcintosh 1994; Raimbault & Sanogo 1991; Togola 1993; 1996).

Figure 1.1: Map of West Africa with the location of mound sites mentioned in the text

A small number of mounds were mapped and excavated in different areas of West Africa (fig. 1.1). The results of some of these works are summarized in this chapter. The issue being addressed revolves around the dynamics of mound-complexes. Why are some settlements made of single small or large mound and others with multiple mounds (mound-cluster). Number of suggestions, anchored on ethnicity, occupation, or various combinations of these have been offered as the rationale for mounds clustering and differentiation (Mcintosh & Mcintosh 1980; R. Mcintosh 1999). Most of the time however, the issue is not even raised and the reasons for the emergence of this kind of settlements are generally left unaddressed.

Nigerian specimens are confined to the northeast, along the southwestern shore of the Lake Chad and the Yobe river floodplain. A group of 8 mounds is reported at Bagade also called Bornu 18 - at 1 kilometer east of Damasak but is not described in detail (Connah 1981: 200). Yau (13o 33’ N/13o 15’ E), on the south bank of the Yobe river, is a complex of 9 or 11 mounds “situated in an area of loose sand .... part of an old flood plain or delta belonging to the Yobe” (Connah 1981: 201). The crescent-shaped, 150 m long and 65 m wide Yau mound 3, the largest of the complex, was selected for excavation. The probe measuring approximately 30 m in length and 2 m in width revealed a 7.77 m thick archaeological deposit dating from the 9-10th to the 13th century AD. Cultural remains scattered all over the surveyed mounds suggest that they may all have been settled at the same time. In this case, Yau may have been an important and impressive settlement in the Yobe river valley. It is however not possible to reject or confirm this possibility.

1.1 - The Chadian Plain Archaeological atlases cannot reveal precise information on the number of mounds present in each of the mapped site (Lebeuf 1969; 1980). It is therefore difficult to know if the phenomenon of mound-clustering is frequent or exceptional all over the Chadian plain. High resolution surveys in southern Bornu in Nigeria and the Houlouf region of northern Cameroon (Connah 1984; Holl 1988; 2002) point to the “exceptionality” of the phenomenon.

Ajere is located at 16 kms from Yau “as the crow flies” (Connah 1981: 212). It is a group of 4, possibly more mounds in the sand dunes area at 3 km from Lake Chad shore. “These mounds look like dirty sand dunes with a scatter of broken pottery on them and much of their surface lacked vegetation” (Connah 1981: 212). One of the mounds

A handful of mound-clusters have been investigated in the Chadian plain in northern Nigeria and Cameroon. The

3

Archaeology of West African Mound-Clusters

was tested but the excavation was halted at 3.20 m below the surface, exposing the upper part of the archaeological sequence. The site is not dated directly but the pottery points to a later date than Yau (Connah 1981: 213). As is the case for Yau, it is not known if all of Ajere mounds were inhabited at the same time, sequentially, or cyclically.

ca. 250-500 to 500-750 AD, during Krenak phase B and Mishiskwa phase A. Mound A and B, separated by 20 m are located in the west, while mound D is found at 280 m in the east. The size of the earliest and initial settlements cannot be known without the excavation of the whole mound. Mound A and B are oriented north-south, and mound D northeastsouthwest. In all three cases, the recorded cultural remains were part of habitation units represented by well-made and maintained round house floors with hearths (Mound A and B) and house floor truncated by a large clay pit (Mound D).

In both cases, the inhabitants of Yau and Ajere were mixed farmers, rearing cattle and sheep/goats, fishing catfish and Nile perch, hunting wild fowl and collecting edible freshwater mollusks like Aspartharia spp. and Pila wernei .

Four mounds, A, B, D, and E, were inhabited during the next step in the growth of the cluster, from 750 to 1000 AD, during Mishiskwa phase B. A new occupation took hold at Mound E, located at 50 m north of Mound D. They delineate two parallel mounds sets, Mound A and B in the west, Mound D and E in the east, on what were the banks of the main channel of the ancient Logone river delta. These four mounds developed from sand islands which were the results of the meandering and ensuing aggradation processes taking place between the active arms of the Logone river delta. Occupation evidence includes living floors (Mound A, B, D, and E), house floors with hearths (Mound D), courtyards with accumulated livestock dung and watering troughs (Mound A and B), a forge furnace (Mound B), cloth dyeing installations (Mound A), as well as a courtyard with fish-smoking pits (Mound E). The multiple facets of village life are represented. Specialized craft activities and the significant degree of maintenance of some of the house floors point to a relatively stable. settled mixed farming communities. Mound E appears to have been used for intensive fish-smoking activities.

A different field tactic was applied to the case of Ble mound complex in the Houlouf region of northern Cameroon. All five mounds were probed providing an interesting insight into the evolution of the complex as a whole. The Ble-Mounds cluster is located at 11o 59’ N and 15o W in an ancient delta of the Logone river (Holl 2002). It is a U-shaped cluster opened in the north and measuring approximately 750 m north-south and 300 m west-east. Ble Mound A, at the northern end of the western branch, measures 600 m north-south, 200 m east-west, and 5 m in height. It is a 12 ha site made of six distinct but fused eminences which seem to have been clusters of habitation features. Ble-Mound B, 0.5 ha in surface extent is found at 20 m southeast of the previous one. It is elliptically shaped, 100 m long, 50 m wide, and 3.5 m high. Ble-Mound C, at 20 m east of Mound B measures 2 ha and is also elliptical in shape, 200 m long west-east, 100 m wide north-south, and 3.5 m high. Ble-Mound D, is oriented northeast-southwest and located at 60 m east of Mound C. It measures 1 ha in surface extent, 120 m in length, 80 m in width, and 4 m in height. Finally, Ble-Mound E also measures 1 ha in surface extent, 185 m in length, 50 m in width, and 4 m in height. It is situated 200 m east of Mound A and 50 m north of Mound C and D.

All five mounds are inhabited during Ble phase A and B, from 1000 to 1400 AD. The main channel of the Logone River was silted and filled up. Mound C was founded on top of the dark gray silty clay sediment of the channel as the Logone River shifted its course further east where it is found today. Occupation evidence stretches over a broad range of activities which include ordinary domestic life with house floors, courtyards, refuse dumping areas. Mound E was devoted most of the time to fish-smoking activities with a special focus on fish oil production. Fish-smoking features were also exposed in Mound A, occupation X and XI, Mound C, occupation VIII. The latest occupation XIV and XV from Mound A appear to have witnessed the development of terracotta figurines and clay head-rests production. The figurines, effigy-jars, and head-rests were all broken intentionally and scattered all over mound A and B, in a destruction frenzy which may have been the result of warfare. The top deposit of Mound D contains numerous fish-smoking features. They may have been part of the economic activities which took place seasonally after the abandonment of the Ble settlement complex at the end of Ble phase B, around 1400 AD.

In summary, there are two different ways of assessing the size of the site under consideration. In one option, only the surface extent of each of the mound may be considered and added up to assess the extent of the settlement complex. In this case, the complex measures 14.5 ha in surface extent. This assessment is clearly unrealistic. The space between the mounds and the delineated territory were clearly an essential component of Ble inhabitants’ behavioral space and has to be considered as such. In the second option then, the settlement size, 22.5 ha, is assessed using the length and width of the mounds set, 750 m north-south and 300 m east-west. Each of the mounds was tested with one or two excavation units with one always set at the highest point to record the complete archaeological sequence. The Ble-Mounds cluster is the result of 1,200 years settlement history which started at the beginning of the first millennium AD, from 250 to 500 AD in the Krenak phase B (Holl 2002) and ended in the Ble phase B, ca. 1200 - 1400 AD.

Ble-Mounds cluster was settled by groups of mixed farmers, hunters, fishermen. The settlements grew incrementally from 3, 4, and finally 5 mounds over a period of some

Three mounds, A, B, and C, were settled initially from

4

West African Mound-Clusters

1000 years. The settlers took advantage of the sand islands frequent in the deltaic environments to create distinct settlements. Generalized domestic activities were the norm with aspects of specialized craft emerging from time to time in any of the mounds.

all the probes from Station 6, an extensive mounds cluster. Layer IV, the earlier, dated to the first half of the 10th century AD, measures 30 - 40 cm in thickness. Layer III on top of the previous one, dating from the second half of the 10th century AD, is 50 - 60 cm thick. Consequently, from the 10th to the 13th century, the core of the town shifted to the central part of the settlement complex, between the Farakole and Folonbadin rivers. The new complex was “made of many dwelling units scattered over an oval-shaped area measuring 700 by 350 m, extended over 15 ha. The habitation mounds [colonies] are set on the perimeter of the oval-shaped area with the center left free of constructions to be used as meeting and/or market place” (Filipowiak 1979: 192). Other areas, some already inhabited and other newly settled, were in use along with the main mounds cluster from Station 6. The royal residence may have been located on Mound 6M1, the largest of the three large mounds (Filipowiak 1979: 192). The 11th – 13th century settlement encompassed the mound complex, including the mounds from Station 6, Somonodougou (Station 5), Krekréloudou (Station 14), the extensive tell of Station 22, hamlets from Station 42 and 45 upstream along the Farakole, and finally, iron-smeltersblacksmiths quarters on the west shore of the Farakole at Station 17 (Filipowiak 1979: 193-194.).

In summary, the mounds clusters from the Chadian plain appear to have been a specific adjustment to the deltaic and flood plain environment. Yau was located in the flood plain of the Yobe River. Ajere was located in a delta, a few kilometers from the shore of lake Chad. And finally, Blemounds complex was also located in a deltaic environment. The intricate network of river channels was an impediment to the development of a single mound as is generally the case in the Chadian plain (LEBEUF 1969; 1981; CONNAH 1981; HOLL 1998; 2002). As a result, available sand islands were settled by different groups depending on the patterns of population growth and the availability of inhabitable. space. 1.2 - The Upper Niger: Niani Niani, the capital of the ancient Mali Empire is found in the upper Niger, today in the Guinea Republic. It is located at 11o 22’ north latitude and 8o 23’ west longitude in the Sudano-guinean climatic zone, at 370 m above sea level (fig. 1.1). Filipowiak’s excavation program dealt with an extensive mound complex at the confluence of the Sankarani, Folonbadin, and Farakole rivers, all tributaries of the Niger River in the Republic of Guinea (Filipowiak 1979). The settlement complex measures more than 2.5 km north-south and 4 km east-west and includes river shores, low hills, steep slopes. It is comprised of more than thirty clustered and scattered mounds, as well as caves, settled with highs and lows from the early 6th to the 17th century AD.

The Royal quarter excavated in the southwest of the Niani Mound complex and sitting on top of the pioneer small wattle and daub village was founded in the late 13th century and inhabited up to the 17th-18th century AD (Filipowiak 1979: 197-198). It included a mosque, 18 m long and 13 m wide, habitation units with series of connected circular houses, the king’s audience hall, a 20 by 20 m construction with four corners square towers and a dome-shaped roof, and finally, a large circular building exposed in an area measuring 10 by 5 m interpreted as the royal Palace (Filipowiak 1979: 235). The Royal complex of Niani-Kaba was inhabited up to the 17th century when it was destroyed by invading Bambara armies. The well was filled up with rubbles. The mosque was burnt down. The upper horizon of layer II dated to AD 1650 is made of carbonized remains of burnt features.

The earliest cultural deposit, confined to the southwest of the archaeological complex, is dated to the middle of the 6th century AD. In its initial phase during the 6th century AD, Niani was a small village on the left bank of the Sankarani River. It then measured some 80 m in diameter with wattle and daub houses, with in the surroundings inhabited caves, iron-smelting sites and burial mounds. From the 6th to the 9th century AD, the settlement grew considerably in size to include Larabou-So (the Arab quarter) in Station 6D, Krekreloudou (Station 14), Linkedibe “Red Gate” (Station 11), Station 45 - a gold mining area -, Station 43 (cave and open-air site), and Station 25 on the slope of the Farakole stream. In summary, Niani became an extensive settlement complex made of two classes of neighborhoods: “one, concentrated in the fertile area along the shore of the Sankarani with Station 1, 6D, 6M, and 14, and the other, with Station 9, 11, 25, 43, and 45, located preferentially on the valley slopes presenting a “montagnard’ character” (Filipowiak 1979: 187).

Niani was the capital city of the Mali Empire, from ca 1240 to 1400 AD and an important town after the collapse of Mali power. Iron-smelting and black-smithing installations were generally located away from habitation mounds. None of these special-purpose sites was excavated. It is therefore not known if these localities were either inhabited by ironworking people or simply used as workshops by commuters from other parts of the city. Burial evidence, the royal complex, and the mounds complex from Station 6 point to the existence of a political elite and a diaspora of North Africans and Arab merchants living at Larabou-So. It is not unlikely that some prestigious patronymic clans - like the Keita - part of the core of Manden politics may have had their own neighborhood in the “sprawling capital city” of the Mali Empire. Such important data are still beyond the grasp of archaeologist’s tools.

The settlement went through an accelerated growth in the 9th and 10th century AD. Layers III and IV were documented in

5

Archaeology of West African Mound-Clusters

1.3 - The Inland Niger Delta: Dia and Jenne-Jeno

immediately north of Jenne-jeno from which it is separated by a shallow water body, measures 8 hectares in surface extent. Kaniana, on the other hand, is located at some 2-3 km northwest of the modern town of Jenne. It is the largest of the sites tested in the mound complex, measuring 41 hectares (S.K. Mcintosh 1994; 1999).

Two important settlement complexes, Dia and Jenne-Jeno, have been investigated in the Inland Niger Delta. Dia is located on the shore of the Diaka, a tributary of the Niger River, in the west edge of the Inland Niger Delta flood plain. It is a large 100 ha mound complex divided into three components of different size: the 49 ha Shoma mound, the archaeological site, the 29 ha present-day town of Dia, and finally, the 28 ha Mara mound which includes the cemetery and a football field (Bedaux et al 2001).

An archaeological sequence outlining the development of Jenne-Jeno and its hinterland, based on a combination of careful excavation, stratigraphic correlation, radiocarbon dates, pottery seriation, as well as other material culture items, has been worked out. The sequence is divided into four phases (I to IV), ranging from 250 BC to AD 1400 (Mcintosh & Mcintosh 1980: 188-198).

The excavation program focused on Shoma mound in the western part of the settlement complex. The uncovered mound deposit measures 3.5 m in thickness, for a site settled from the first half of the first millennium BC to the 18th century AD. The settlement history is segmented into five period called horizons. The earliest Horizon I deposit belongs to a Late Stone Age site of remarkable size (Bedaux et al 2001: 841) with well preserved floors. Bone tools and material culture elements which link the Early Dia to the Pan-regional Late Stone Age are well represented in the recorded assemblages. Iron artifacts are present at the end of the horizon, from 200 BC onwards.

Phase I (250 BC – AD 50) witnessed the pioneer settlement of mixed farmers equipped with iron tools, who raised cattle and goats, cultivated African rice, fished, hunted, and foraged in the relatively rich in-land Niger deltaïc environment. Their dwellings were made of wattle and daub. The size of the initial site is not known. It is however very likely that distinct domestic units may have settled in this optimal zone during the dry spell of the end of the first millennium BC. The site developed progressively from the fusion of these built discrete spots.

Horizon II, between BC 300 and AD 300, marks a shift toward intermittent habitation during a particularly dry and arid period. The material culture repertoire is poor with however a continuation of the previous Horizon I elements. Horizon III, from AD 300 to 800 (?), suggests the existence of a permanent settlement with house floors and mud-brick architecture. Horizon IV was significantly impacted by erosion (Bedaux et al 2001: 844). It dates from AD 800 (?) to 1100, when the area was used as a cemetery. The tested portion of the site was abandoned from AD 1100 to 1400. Settlement resumes with Horizon V, from 1400 to 1700 AD. Dia was then a large walled city. The city wall, still partly visible along a 230 m stretch, consists of two parallel zig-zaging 0.50 m thick walls with the 1-3 m interim space filled with rubble. Elongated rectangular houses or rooms were visible from the surface. The excavation probes A and B show these houses to be situated on both sides of a narrow street, with each of the probed room attached to a courtyard (Bedaux et al 2001: 839).

Phase II, dated from AD 50 to AD 400, witnessed a relatively fast growth of the settlement and its population. The subsistence base remained the same but an intensification of resources exploitation and production certainly took place. It is inferred that Jenne-jeno size shifted from 12 ha in AD 100 to 25 ha in AD 400 (S.K. Mcintosh 1999: 70). There was no significant change in architecture either but iron, stone and a few glass beads from Mediterranean origins attest for the connection with local and inter-regional exchange networks. During Phase III (AD- 400 – 900), Jenne-jeno reached its maximum extent of 33 hectares around AD 800. Mud architecture is adopted and dwellings then consisted of round banco huts. The practice of urns burial emerged and a cemetery developed in the Jar-field 1 area. The deceased were nonetheless buried without grave goods. Beside iron, copper and gold are present in the area archaeological record, suggesting a connection to the long-distance trade network. The size of the local population is inferred to have peaked during the later part of Phase III; even if estimating population densities from mound size generally tends to inflate unverifiable population figures (S.K. Mcintosh 1999: 71-73). This does not diminish the relevance of Phase III settlement data, if and only if the tight correlation made between surface sherds’ scatters characteristics and site chronology holds.

The results from the Dia archaeological project do not allow for the investigation of the genesis of the settlement complex. It is indeed an important and complex site; both Mara and present-day Dia mound have to be probed to collect relevant data. Jenne-jeno, Kaniana, and Hambarketolo, excavated by S.K. McIntosh and R.J. McIntosh (1980; 1994), are some of the largest mounds of an extensive mound-settlement system to be found in the flood plain of the Inland Niger Delta. Jenne-jeno, at 13o 53’ 20” north latitude, and 4o 32’ 25” west longitude, is a teardrop-shaped 33 hectares mound, 760 m long, 550 m wide, rising to 8 m above the surrounding plain (Mcintosh & Mcintosh 1980: 63). Hambarketolo,

Phase IV, ranging from AD 900 to 1400, closes the developmental sequence of the Jenne-jeno town complex. It started with the construction of a 3-m thick and 2 kilometers long fortification wall around the 33 ha site. Rectilinear buildings appear at the turn of the millennium in AD 1000

6

West African Mound-Clusters

(Mcintosh & Mcintosh 1980: 191; S.K. McIntosh 1999: 71) with more frequent North African imports, brass, glass, and spindle whorls. In fact, building activities are shown to have been particularly intensive and impressive. In Jenne-jeno Probe M2 Phase IV deposit, “houses 7, 4, and 3 were built and destroyed within a short period of time” (Mcintosh & Mcintosh 1980: 191). However, the construction of three successive mud-brick houses on the same spot during a period of 150 years, does not seem particularly unusual as it amounts to a remarkable 50 years use-life per building. From AD 1200 on, terracotta statuettes representing warriors became part of the Jenne-jeno cultural repertoire up to the final demise and abandonment of the site around AD 1400 (R. Mcintosh 1999; S.K. Mcintosh 1995, 1999; Mcintosh & Mcintosh 1980, 1984, 1993).

homesteads, each inhabited by 1 or a small number of extended families. There were significant changes in the structure of settlements during the Middle Assemblage Period which ranges from AD 600 to 1200. The number of mounds per settlement decreases sharply but they tend to be larger on the average. The number of mounds per site ranges from 1 at Boundou Boubou North, Boundou Boubou South, and Bourgou Silatigui, to 6 at Akumbu. 12 mounds out of a total of 25 measure 6-10 to more than 20 ha in size. Two important regional centers situated at 25 kilometers from each other emerged during this period. They were in fact the two largest >20 ha mound complexes, one at the 6-mounds cluster of Akumbu and the other at the 5-mounds cluster of Toladie. Toladie, located in the central part of the Mema, measures approximately 80 ha in size. Akumbu, in the southwest, includes three habitation mounds: Akumbu A is 6-7 m high above the surrounding plain and measures 21 ha. Akumbu B is 7-8 m high and 8 ha in size, and Akumbu C, 7 m high and 3 ha in size. There are in addition two burial mounds and two small iron smelting sites at the eastern and northern edge of the complex (Togola 1993: 66-7).

The clustering of mound sites is partly induced by the geomorphologic dynamics of the Inland Niger delta floodplain with its dense network of permanent and seasonal river channels. The different mounds of the Jennejeno complex were suggested to have been the result of residential patterns based on distinct craft specialization and corporate identities (Mcintosh 1999; Mcintosh 1994; Mcintosh & Mcintosh 1980; Polet 2004). Differences in the distribution and relative density of distinct categories of cultural remains were seen as evidence of occupational and/or ethnic identities. These possibilities generally based on the ethnographic present have not yet been confirmed by archaeological excavations.

The number of settlements plummeted during the Late Assemblage Period, from AD 1200 to 1400. Goudourou, Niakare Dondi and Tiable Fame with 1 to 2 mounds each are the only settlements still inhabited. Goudourou and Niakare Dondi are single mound sites, the former with a 1120 ha mound and the latter with a 1-5 ha site. Both mounds from Tiable Fame measure 1-5 ha in size. This Later period also coincided with the rise to regional primacy of the Mali Empire. Trade routes connected to the Trans-Saharan trade system shifted east, circumventing the Mema even if Ibn Battuta asserted that there was a trade route in the region in the middle of the fourteenth century AD (Levtzion & Hopkins 1981: 279-284).

1.4 - The Mema region (Mali) The Mema is a vast ancient floodplain situated at approximately 100 km southwest of the Malian Lakes region and northwest of the Inland Niger Delta (Haaland 1980; Togola 1993, 1996). A number of sites was tested providing useful chronological indications. Late Stone Age sites like Kobadi date from 1600 to 350 BC. Mounds excavated at Site B-E, south of Boulel Ridge (Haaland 1980), Kolima, Akumbu A, and Akumbu B, date from AD 340-440 to AD 1275-1400 (Togola 1993, 1996). A regional survey allowed for the mapping of 137 archaeological sites distributed into 27 Late Stone Age sites, 94 habitation mounds, 15 iron-smelting sites, and two cemeteries with burial jars. Most of the habitation mounds are organized into mound-clusters. A combination of archaeological probes and survey data has allowed for the chronological ordering of the recorded settlement data (Togola 1996; Holl 2004: 142-3).

The genesis of the mound-clusters from the Mema region is not known. The deltaic context and the development of larger mounds during the Middle Assemblage Period point to a flexible adjustment to the local biomes depending on the circumstances. Some mounds are however exclusively devoted to habitation, or burial, or iron-smelting activities. 1.5 - The Oudalan Province in North Burkina Faso The Oudalan province is located in the north and Sahelian part of Burkina Faso. A sustained archaeological research program including intensive surveys and targeted excavation was conducted in a relatively large territory situated between 14o and 15o latitude north and 0o 20’ and 0o 40’ longitude west (Albert & Kahlheber 2001; Alabert et al 2000; Ballouche 2001; Hallier 2001, Kahlheber et al 2001; Neumann 2000, 2001; Neumann et al 2000; millogo & kote 2000; peltzer & magnavita-santos 2000). The area is characteristically made of extensive sand-dune systems oriented east-west separated by low lying plain and

Nine sites are dated from AD 300 to AD 500. Akumbu, Goudourou, Kolima, and Toule are single mound sites. The remaining ones are comprised of 4 (Niakare Dondi) to 18 (Boundou Boubou North) mounds. In general, the mounds of this Early Assemblage Period are small in size, measuring less than 1 to 5 ha. 27 out of 49 measured mounds are less than 1 ha in surface extent with the remaining 22 ranging from 1 to 5 ha. It is as if settlements were made of distinct

7

Archaeology of West African Mound-Clusters

inter-dune depressions. The latter are generally filled with semi-permanent lakes and lakelets.

magnavita-santos

2000). The Kissi settlement complex started as an iron-smelting site in the second half of the 1st millennium BC, 410-340 cal BC/300-200 cal BC. However, more sustained and consistent settlement activities took place in the first part of the 1st millennium AD. Mounds settled during this “early 1st - 4th century occupation phase were few and located very close to the lakelet shore (Magnavita 2009: 81-83). Settlement activities peaked during the “middle occupation phase from the 4th to the 9th century AD. Survey data suggest the site to have reached some 100 ha in surface extent. Kissi-3 cemetery dates from that period. It was in use in the 6th and 7th century AD and contained two particularly well furnished burials dated to “544 cal AD +/- 45 and 642 cal AD +/-33“ (pelzer & magnavita-santos 2000: 369). The latter includes iron weaponry - swords, daggers, bows and arrows - as well as elements of personal adornment like blue, green, yellow and translucent glass beads, and cowry shells. The population of the settlement complex was stable during the following late occupation phase, from the 9th to the 12th century. The communities of mixed farmers left the area at the beginning of the 13th century AD to be replaced by pastoral-nomadic peoples with large cattle herds (magnavita 2009). The archaeological record from Kissi reveals the existence of a relatively “wealthy” elite of “warriors” at the beginning of the second half of the first millennium AD in the sand dunes of northern Burkina Faso. These iron-using communities were connected to the international long distance trade network linking West Africa to the Maghrib, Ifriqiya, and the Machrek.

The survey operations have allowed the mapping of some 178 sites made of mounds. These mounds are generally arranged in clusters, with individual mounds located at distance varying from 100 to 200 m, with diameter ranging from a few tens of meters to 100 m maximum (albert et al 2000; hallier 2001). Six settlement complexes distributed along a south-north transect from Dori to Ti-n-Akof were selected for further investigation: Saouga, Gandefabou, and Ti-n-Akof are located on river shores while Dori, Kissi, and Oursi are on lake shores. The settlement history of the Oudalan province is particularly intriguing. Late Stone Age occupations dated from 2200 to 1000 BC were recorded at Ti-n-Akof in the north, Oursi in the central zone, and Dori in the south. There are no traces of human presence prior to ca 2200 BC and after 1000 BC. The area is re-settled at the beginning of the present era by iron-using communities up to AD 200 and from 500 to 1500 AD. The settlements from the second millennium BC were those of “highly mobile groups - most probably hunters and gatherers” (neumann et al 2000: 329). Agriculture, involving the cultivation of pearl millet - Pennisetum glaucum - was practiced at the end of the second millennium BC at Ti-n-Akof and Oursi. Iron-using communities are documented to have settled at Dori (BF 94/96), Saouga (BF 94/120), Kissi (BF 97/3), Oursi (BF 94/45, BF 97/13), and Ti-n-Akof (BF 94/133). All these settlements are made of varying number of mounds

Gandefabou settlement complex is located in the centrewest of the province. It is made of four sites arranged in pairs on the top of the dune cordon and the shore of the Kel Ewel intermittent river (millogo & kote 2000). The recorded sites are located at a little more than 1 km from each other. Gande I, a cemetery with large burial jars, and Gande IV, a shallow habitation mound, are set some 500 m apart on the dune top. Gande II, the habitation mound, and Gande III, the cemetery with large jar-burials, are 300 m apart on the north flank of the dune cordon. The relationship between a habitation mound and a cemetery is straightforward in this case which is dated to the 10 th - 13th century AD (millogo & kote 2000: 362).

Oursi settlement complex, in the north central part of the Oudalan province, is made of 10 mounds located on a dune cordon near a more or less permanent inter-dune lakelet. The largest of the recorded mounds was tested down to the sterile basal dune sand at 6 m below the mound top. The exposed archaeological deposit is 6 m thick. Late Stone Age occupations were found at 4.10 to 6 m below the mound top. They are overlain by 4 m of iron-using communities’ deposits. At Corcoba (BF 97/5), another mound of the Oursi complex, the LSA occupation is dated to 2175 - 1945 BC. 1 m deep LSA pits of unknown use were recorded. They contained fish remains, reedbuck, kob, large bovid bones, but no domestic animals (neumann et al 2001). The identified fishes belong to species of open, relatively deep and well oxygenated water. Those from the later Iron-Age deposits came from fish living in shallow poorly oxygenated water, thus pointing to a shift toward drier conditions during the first millennium AD (kahlheber et al 2001).

There are accordingly interesting variations between the explored mound-clusters to be investigated in the future. As far as Gandefabou is concerned, jar-burials were used for the internment of individuals of all ages and sex. Infants and children remains were found in both tested cemeteries. As shown by jar-burials 13 and 15 at Gande I, adult males were buried with their weapons, in this case, knives and arrowpoints, and adult female with their jewelry, necklaces, finger-rings, and bracelets. Interestingly, burial jars appear to have been re-used. Human remains found in piles suggest that the containers were emptied of their previous content and re-used for the burial of a newly deceased individual (millogo & kote 2000: 361).

Today, Kissi is a small hamlet inhabited by groups of Kel Tamesgayt, a fraction of the larger Kel Tamasheq tribe. The Kissi mound complex, which includes habitation mounds and cemeteries, is located on the dune cordon on the north flank of the inter-dunal lakelet. Excavations were carried out at a number of Kissi mounds, including three cemeteries and two habitation mounds (magnavita 2004; pelzer &

8

West African Mound-Clusters

A small portion of a wall made of stone and clay mortar was found resting on the dune sand at 0.30 m below the surface. The shape of the feature built is still unknown but it was probably part of a domestic installation. The sites excavated in the Gandefabou settlement complex are all surrounded by large quantities of iron working debris, like furnace bases and slag. Their concentration is particularly high at Gandefabou-Djelgobe and Saba Diaba 1 km further north (millogo & kote 2000: 362).

form a triangle around a relatively large 56 m long, 22 m wide, and 7 m high tumulus. The latter contains the remains of at least 28 individuals, with terracotta and alloyed copper figurines, cowry shells, pottery, and animal bones. El Oualadji excavated in 1904 provided important archaeological evidence which fit with Al-Bakri’s description of the burial customs of ancient Ghana kings. Lieutenant Desplagnes spectacular finds triggered the assimilation of the lakes region large earthen mounds with burial tumuli. R. Mauny sets out to excavate the mound of Kouga in 1954-1955 with the hope to replicate Desplagnes discoveries. He found a habitation mound dated to the end of the first millennium AD (Mauny 1961), but was nonetheless still convinced to have missed the burial chamber which, he thought, may have been located a few meters west (Sanogo 2004: 192). More recently, habitation mounds were excavated at different localities of the region, at Kawinza dated to AD 670-880, Mouyssan II, and Toubel.

Saouga, in the centre-south of the Oudalan province, is a cluster of 5 mounds dated to ca 900 - 1100 AD. It is located on the northern flank of an east-west dune cordon near the confluence of the Gorouol and the Goula rivers (albert et al 2000). This location, characteristic of almost all the archaeological sites of the province, was geared to take advantage of light and fertile soils for the cultivation of pearl millet - Pennisetum glaucum - and cowpeas - vigna unguiculata -. This crop system may have provided the staple food during the Iron Age.

Mouyssan II is a circular mound measuring 110 m in diameter and 9.6 m in height, located at approximately 14 kms south of the village of Soumpi. The exposed cultural deposit is 9.6 m thick made of a coherent succession of habitation and metallurgical activities debris. The remains of an arch-shaped mud brick wall were found at 7.59 - 7.80 m below the surface, and may have been part of a 6.70 m in diameter circular house. Mouyssan II was settled from 300 to 680 AD with the latest top three occupations devoted to metallurgical activities.

In summary, mounds clustering appears to have been an Iron Age phenomenon in the Oudalan province of northern Burkina Faso. There were some straightforward functional differences between the components of each of the studied mound complex. Some mounds were used exclusively for habitation, others were cemeteries, and others again were iron smelting workshops. The recorded archaeological data have not yet revealed significant recurrent variations among habitations mounds which may suggest that they could have been inhabited by different ethnic/occupational/ social groups. As shown by the settlements from the Gandefabou complex, it is very likely that each habitation mound had its own cemetery; a situation which may point to the existence of well delineated “corporate identities”. Such social entities - nobility, commoners, guilds of craft people, captives, etc. - are however extremely difficult to trace in archaeological contexts. The situation is even more challenging if one takes into consideration the relatively small size of the tested excavation probes, the constant re-distribution of the regional population, and the “palimpsestic“ nature of the occupational deposits.

Toubel is a two-mounds complex also located in the vicinity of Soumpi, at 6 kms in the west. The smaller mound, west of the complex, was selected for excavation. The exposed cultural deposit is 5.6 m thick (sanogo 2004: 193-4). The first occupation, dated to AD 340-560, started with metallurgical activities documented at 4.4 to 5.6 m below the surface. The uncovered cultural remains include a few potsherds, iron slag, burnt clay lumps, and series ash pockets. The next series of occupation are found at depth varying from 2.6 to 4.4 m. The recorded material culture consists of the remains of a round mud brick house, associated with charcoal, slag, and burnt clay lumps. The mound was used for refuse disposal with the ensuing accumulation of the top 2.6 m of the cultural deposit. With three primary burials found in the top part of the deposit, at 3.3, 2.8, and 0.5 m in a 4 by 4 m excavation probe, Toubel seems to have been turned into a burial ground during the first half of the 2nd millennium AD, from ca. 1215 to 1430 AD.

1.6 - The Lakes Region, Mali The Lakes region, south of Timbuktu in south central Mali, was explored from the beginning of the 20th century and surveyed more systematically in the 1980s (raimbault & sanogo 1991; sanogo 2004). Numerous sites have been recorded including “175 large mounds supposed to be burial tumuli” (sanogo 2004: 192). The mapped settlements are partitioned into three categories: (1) single-mound sites, (2) clusters of one large central mound surrounded by a few smaller mounds, and (3) multiple mounds clusters. A handful of sites were tested in the area starting with the pioneer excavation of Lieutenant Desplagnes in 1901 and 1904. Koi-Goureye - the earthen tumulus in Songhai language - is a three mounds complex on the shore of the Goudam pond, southwest of Timbuktu. All three mounds

In summary, the excavations conducted in the Malian Lakes region on a series of selected mounds clusters reveal an “Iron Age” occupation stretched over the 1st millennium and the first half of the 2nd millennium AD. The settlements are made of burial tumuli and habitation mounds. Dabi, Tondidaro, Koi-Goureye, and El-Oualadji are burial mounds ranging in date from 670 - 790 to 1030 - 1120 AD.

9

Archaeology of West African Mound-Clusters

Kouga, Kawinza, Mouyssan II, and Toubel, are habitation mounds dating from 300 - 440 to 1215 - 1430 AD.

Burkina Faso, and Guinea. The less than handful cases from the Chadian plain, in Northern Cameroon at Ble and Northeastern Nigeria at Yau and Ajere, are clearly exceptional. Read from another perspective, moundclustering was plainly an “Iron-using communities” phenomenon. This peculiar kind of settlements was created by communities of iron-using mixed-farmers and fishing-folks, with some practicing livestock husbandry of cattle, sheep and goats. As shown by the cases from Niani (Filipowiak 1979), Jenne-Jeno (McIntoch & McIntosh 1980), Kumbi-Saleh (Berthier 1997; Polet 2004), Awdaghost (Devisse 1983; Holl 2006), and probably Gao (Insoll1996, 2000), mound clustering was the very foundation for the emergence of West African early urban complexes.

1.7 - The Middle Niger: The Bura region The Bura region on the right bank of the river, is located in the Middle Niger around the capital city of Niamey, between the Gorouol and the Mekrou rivers (Gado 1993, 2004). A sophisticated and spectacular funerary art was brought to light in some of the sites excavated by B. Gado. Settlement systems vary from area to area. In two specific cases, the researchers have documented sites clusters. One, termed the “anthropomorphic funerary statuary area of Bura sites”, includes three sites variants which are suggested to characterize the local culture of Bura-Asinda-Sikka (Gado 2004: 164). Accordingly, each settlement complex includes cemeteries sites with anthropomorphic vessels and effigy terracotta statues, flat top mounds ritual sites for offerings, and finally, habitation sites. It ranges in date from the early 3rd century to the mid-11th century AD. The other, dated to 1270 -1430 AD and called the “Kareygooru tumuli culture” is found along the Niger River shore at KareygooruBirniwol where a series of six large mounds were recorded.

The research strategies implemented so far in most of the cases reviewed in this chapter had as their main goal the clarification of the sites chronologies, relying on the sampling of one mound among many of the complex under investigation. It is a perfectly legitimate research strategy, but it does not allow for the investigation of the internal dynamics of the mounds clusters. What are the social, economic, and political foundations for mound-clustering? The most coherent approach to this interesting question is derived from the work carried out at Jenne-Jeno by R. J. McIntosh and S.K. McIntosh. For R.J. McIntosh (1999), Jenne-Jeno town complex was the product of patterned interaction between complementary and equal social groups, for the mutual advantage of all in a heterarchical system. Hunters, fishermen, herders, warriors, farmers, and craft peoples lived in harmony without the need for a political elite in a “city without citadel”. According to this perspective, each mound may have been inhabited by a specific occupation/craft group part of the broader city-wide network. The suggestion, derived from the “ethnographic present” is particularly interesting but has not yet been tested with archaeological data. Addressing such a research problem requires a totally different excavation strategy. Accordingly, each mound of a settlement complex has to be tested to obtain an optimal sampling of the complex settlement history. Even if this research strategy is fully implemented, the average size of each of our usual archaeological probes is still too small to guarantee an accurate access to the whole settlement history. It is nonetheless better than generalizing from a single mound probe. It is this “comprehensive” research strategy that was implemented in the Mouhoun Bend Archaeological Project, the results of which are presented in this book.

In summary, the practice of mound clustering documented in the Middle Niger valley is clearly a geographic extension of the phenomenon documented in the Oudalan province in northern Burkina Faso. The river Gorouol, a tributary of the Niger River flows west-east, and may have been an axis of movement of people and ideas. Conclusion Mounds clusters are widespread in West Africa (fig. 1.1). They are a continental phenomenon generally found in floodplains and low lying lands in the Sahel and savanna belts. There are a number of cases in wetter southern latitudes, Niani in Guinea or Begho in modern Ghana. Beside the areas discussed so far in this chapter, moundsites are found elsewhere in West Africa, in the Chad republic part of the Chadian plain and the Borkou (Lebeuf 1969, 1980; Treinen-Claustre 1982), the Bandiagara and eastern flank of the Inland Niger Delta (Bedaux 1972; Bedaux et al 1978), the northern part of Ghana (Begho) and Cote d’Ivoire (Kong), and the northwestern Burkina Faso to be discussed in this book. Interestingly enough, mounds clusters have a narrower geographic extension. They are found along the Middle to the Upper Niger and surrounding lands, in Niger, Mali,

10

Chapter 2 The Mouhoun Bend: Land, Humans, and Archaeological Survey

Introduction

2.1 - Climate, Soils, and landscape

The Mouhoun bend is located in northwest Burkina Faso in the Mouhoun province. The Study area is delimited in the north and northeast by the meandering course of the Mouhoun River (fig. 2.1). The selected area measures 40 km East-West (3o 11’ / 3o 32’ longitude East) and 38 km North-South (12o 30’/ 12o 45’ latitude North). The land is flat in general, with elevation ranging from 294 to 249 m above sea level.

The study area is located in the south of the Gondo plain at elevation ranging from 200 to 300 m above sea level. It belongs to the Sudano-Sahelian climatic zone with annual rainfall varying from 600 to 900 mm. The wet season lasts 4-5 months. The soils are hydromorphic to pseudogley accumulated on the Bobo Dioulasso sandstone and schist. Most of the local rocks, sandstone, limestone, and dolomite, are sedimentary in nature. As part of the Sudanian domain,

Figure 2.1: The Study area

11

Archaeology of West African Mound-Clusters

the vegetation is a wooded and arboraceous savanna. It is predominantly made of acacias and thorn bushes on its northern fringe. Essentially however, the Sudanian domain includes such important trees as the Karité (Butyrospermum parkii), the néré (Parkia biglobosa), the cailcedra (Khaya senegalensis), and the baobab (Adansonia digitata). As a legacy from an ancient and remote tradition of agroforestry, “the karité and the néré, untouched by land clearing operations, and well protected by both the local population and the forestry regulations for their economic value, are still the predominant species in the agricultural fields all over the country ... The oleaginous karité kernel, known commercially as the “shea butter” seed, is a significant source of vegetable fat for the Burkina Faso population and contributes to the country’s exports (atlas of burkina faso 1998: 22). 2.2 – Present-day Human Geography Dedougou is the capital city of the Mouhoun province. Douroula, where the research base camp was located, is a sous-prefecture at approximately 50 kilometers north of Dedougou. The population density varies from 30 to 50 inhabitants per kilometer square throughout the region. It is settled principally by the Bwa and Marka, with scattered groups of Fulani pastoralists and an increasing number of Mossi. The Bwamu, - the language of the Bwa - is one of the Gur languages. That linguistic family also includes the Gurma-More, Gursi, Lobiri, Senufo, and Dogon (Atlas 1998). The Marka, speakers of a Mande language, are located along the eastern fringe of the study area. They are claimed to have spread from the Niger valley, following the Sourou and the Mouhoun rivers valleys. “The present extension of the Mandingue resulted essentially from the expansion of the Mali Empire through the conquests of Emperor Sunjata in the 13th century and the considerable role played by the Mande for centuries in West African long-distance trade networks” (Diallo 2000: 379). The findings of Diallo (2000: 383) comparative grammar analysis are particularly interesting. He found striking similarities on two grammatical aspects of Burkina Faso Marka and Gambian Mandinkan languages. These facts are explained by a common origin which is suggested to have been the Manden, the core area of the emergence and expansion of the Mali Empire in the 13th century AD. Today, the study area is inhabited by a mosaic of different communities, including the Bwa, Marka, and Mossi mixed-farming communities, growing sorghum and maize in an anthropic savanna parkland dominated by the karité (Butyrospermum parkii), small groups of mobile Fulani pastoralists, and colonies of Somono fishing communities settled along the Mouhoun River. The patterns of land use are quite straightforward, a dendritic network of small roads and footpaths radiating from Douroula in all directions (fig. 2.2) with rings of fields around the villages. Newly established fields located at more than 5 kilometers, as can be seen on figure 2.2 in the southwest, are connected to the village by footpaths.

Figure 2.2: Patterns of land-use around Douroula

In fact, the general density of settlements is particularly low in the study area. There are four villages with Douroula at the center surrounded at 4/5 km distance by Sa in the north, Souma, in the east, and Koussiri in the southeast (fig. 2.2). Sa is the residence of the “High priest” who presides over the installation ceremony of the chief of Douroula. Souma is the base of a relatively large colony of Somono fishermen, while Koussiri is a small “ordinary” Marka village. The local communities are subtly but strongly stratified, with the traditional ruling families at the top of the social hierarchy, free born individuals, cast-like occupational groups of blacksmiths, potters, leather-workers, wood-carvers, etc., and finally, the captives and their descendants. 2.3 - Archaeological Survey In the first stage of the field project in 1997, the surveyed area was narrowed in order to focus on the north and northeast. The west and southwest were surveyed later in 1999 and 2000. Contrary to the expectations, as expressed in the project proposal, no Late Stone Age site was found in the selected study area in 1997 despite persistent effort in the careful survey of eroding slopes. This absence was surprising but could have been explained by heavy aggradation in the flood plain during the last millennia. No palaeoclimatic research has ever been conducted in the Mouhoun River drainage basin. As such the above suggestion is highly speculative. This speculation is based on information collected from satellite images. These images show an intricate pattern of overlapping fossil-rivers channels (fig. 2.3). The Mouhoun River has changed its course and meanders several times during the last few

12

The Mouhoun Bend

Figure 2.3: Network of fossil channels in the Mouhoun river meander

millennia. If there were Late Stone Age sites along any of those fossil-rivers channels, there are two possibilities. On the one hand, if the Late Stone Age sites are still preserved, they are very likely concealed by thick sediment deposits. On the other hand, and depending on river’s current velocity, the shallow Late Stone Age settlement remains may have been eroded and carried away by the “wandering” river.

The iron-ore mining complex consists of clusters of deep shafts, 1.5 to 2 meters in diameter, 5 to 6 m deep, connected to each other by tunnels (fig. 2.4). In some cases mining operations resulted in the creation of large underground cavities with impressive roof-supporting pillars (fig. 2.5). The local iron ore is essentially from the lateritic crust which is ubiquitous in the area. This duricrust is constituted of two thick layers with significant difference in hardness and iron (FeO) content. The top layer is highly oxidized and altered by weathering. The deeper layer is on the average softer and richer in iron. The mining shafts were dug with spiraling stairways. But some of the interconnected series of shafts and subterranean galleries have collapsed. According to informants, some of these subterranean features were

The second and third surveys were however more successful. Late Stone Age Lithic scatters exposed by erosion agencies were found at Kebe-Kan-Sira-Tomo near Diekono on the right bank of the Mouhoun River, between the cliff and the water course. The presence of this material points to an earlier but undated occupation of the Mouhoun valley by mobile groups of Late Stone Age foragers. In winter 1999, an attempt to probe the sediment build up in one of the fossil channels of the left bank of the Mouhoun River in the U-shaped meander failed. It was in the middle of the dry season. The surface clay was too hard, dry, and dusty. In addition, the experience of crossing the river in a dug out with all the tools, students and workmen was extremely risky. The task was abandoned. A more focused survey operation was conducted on the left bank of the Mouhoun River, at the impressive iron-ore mine located next to the village of Tissi, some twenty five kilometers north of the study area. It is situated next to a complex of four shallow mounds with mining features spread over an area of approximately 500 m in diameter.

Figure 2.4: The mining shafts from Tissi

13

Archaeology of West African Mound-Clusters

the north-northeast, and Douroula in the centre. They are all multi-mounds complexes, some of them quite important as is the case for Kerebe and Souma. Those from Douroula are relatively small and low in general. The excavated settlements will be presented in their own chapters. Three mounds-clusters were found in the southeast, in the proximity of the Koyare intermittent river, a tributary of the Mouhoun River. Bawasso, (12o 31’ 28” N, and 3o 16’ 22” W), at 262 m asl, and located on the left bank of a small tributary of the Koyare is comprised of four mounds with a quarry (fig. 2.1). The recorded habitation mounds are elliptical, circular, and crescent-shaped. Two of the mounds are 146 to 148 m long, 83 to 135 m wide, and 3 to 4 m high. The remaining two are almost similar in size (tab. 2.1). The quarry which measures 25 m in diameter is made of two large pits and series of 4 to 9 mine shafts.

Figure 2.5: View of the galleries from the iron ore mine of Tissi

used as hiding places during the 18th and 19th centuries. No systematic scientific investigation has ever been conducted on these impressive mining sites. There is no clue on their probable chronological span. What is nonetheless certain at this stage, is the fact that at an unknown period of the past, there was an intensive exploitation of iron ore in the area. A topic which deserves further investigation.

Buo Kien (12o 31’ 48” N, and 3o 18’ 22” W), at 264 m asl is found at a few hundred meters of the head of another small tributary of the Koyare river (fig. 2.1). It is made of four habitation mounds and one quarry site. The quarry is circular in shape and rather shallow (tab. 2.1). The recorded habitation mounds are sub-circular to elliptical in shape, measuring 120 to 60 m in length, 56 to 38 m in width, and 1 to 3 m in height.

Beside the Late Stone Age site referred to above, seventeen settlements with two to seventeen mounds [Tomo in Marka language] each have been recorded so far in an area measuring 20 km east-west and 25 km north-south (fig. 2.1). Kirikongo and Tora-Sira-Tomo/Gnambakouon Sira Tomo are located at 280 to 284 m above sea level (asl). All the remaining settlements are found at elevation ranging from 280 to 260 m asl. There is an intriguing concentration of mounds clusters in the central part of the surveyed area, around Douroula. Two of these single mounds are “un-named”. The other two are called Bere-So-Tomo and Kienso-Tomo. They probably formed a single large settlement complex with Douroula and are therefore not considered as different settlements but part of an extensive more encompassing complex.

Buo Kien So Tomo (12o 30’ 25” N, and 3o 19’56” W) is located at 266 m asl at approximately one kilometer north on the left bank of the river Koyare (tab. 2.1). It is a seven mounds complex with five habitation mounds, one small cluster of clay vessels, and an iron-smelting workshop. The habitation mounds and the vessels cluster are circular to elliptical in shape and tend to be small in size. They vary in length from 15 to 53 m, in width from 15 to 43 m, and in height from 1 to 2.5 m. The complex quarry was not found. The iron-smelting workshop is circular in shape and measures 15 m in diameter (tab. 2.1). Kirikongo (12o 32’ 31 N, and 3o 23’ 13 W W), located at 282 m asl on the right bank of a seasonal tributary to the Koyare river, was surveyed in 1997 (Holl & Kote 2000) when 11 mounds were recorded (fig. 2.1). The data recorded were later given to S. A. Dueppen for his dissertation research. He carried additional work later at the complex and found that it was made of 13 mounds, 3 quarries, and 6 ironsmelting sites (Dueppen 2008). In fact, the core of Kirikongo is a relatively large sub-circular multi-mounds settlement measuring 550 m in diameter, approximately 23.75 ha in surface extent. The whole complex with the outlying mounds measures 700 m north-south and 600 m west-east, approximately 42 ha in total surface extent. Interestingly enough, all the iron-smelting sites are located along the west flank of the complex, allowing for the furnace smoke to be blown away from the village by the harmattan, the northeastern wind. The 1997 survey showed that the two largest mounds are oriented west-east and measure 150 to 160 m in length, 105 to 120 m in width, and >2 to 4 m in height (tab. 2.1). The next set in terms of size also includes

During the 1999 field season, survey was confined to the northern part of the study area, from Kerebe in the West to Sa in the east (fig. 2.1), along the river valley. Compared to the rest of the study area, the density of settlement is surprisingly low in the river valley if one considers that the region is prone to cyclical droughts. River blindness has plagued the whole Mouhoun River basin during the colonial period, and probably before. This may explain the avoidance of the river shores for direct settlement, as shown by the fact that there is no single permanent village on the river banks today in the study area. Instead, there are seasonal fishing camps. 2.4 - The surveyed settlements A number of settlements were visited but not surveyed and measured accurately (tab. 2.1). It is the case for Magnoumosso and Kerebe in the west, Sa and Souma in

14

The Mouhoun Bend

Table 2.1: Mouhoun Bend Settlement survey Site

Length ( m)

Width (m)

Surveyed Settlements 1 - Bawasso 12o 31’ 28” N, 3o 16’ 22” W, 262 m asl Bawasso -1 146 NN-SE 135 NE-SW Bawasso -2 73 73 Bawasso -3 70 E-W 65 N-S Bawasso -4 148 E-W 83 N-S Bawasso -5

25

25

2 - Buo Kien 12o 31’ 48” N, 3o 18’ 54” W, 264 m asl. BK-1 120 E-W 38 N-S BK-2 90 NW-SE 47 NE-SW BK-3 60 NE-SW 56 NW-SE BK-4 72 N-S 38 E-W BK-5 42 42 3 - Buo Kien So Tomo 12o 30’ 25” N, 3o 19’ 56” W, 266 m asl. BKST-1 43 43 BKST-2 53 N-S 35 E-W BKST-3 33 E-W 20 N-S BKST-4 44 N-S 35 E-W BKST-5 30 30 BKST-6 15 15 BKST-7 15 15 3 - Gnissa Ka Tomo 12o 35’ 38” N, 3o 18’ 39” W, 267 m asl. GKT-1 100 E-W 95 N-S GKT-2 75 E-W 58 N-S GKT-3 50 50 GKT-4 54 E-W 40 N-S 4 - Kirikongo 12o 32’ 31” N, 3o 23’ 13” W, 282 m asl. KRK-1 75 E-W 53 N-S KRK-2 150 E-W 120 N-S KRK-3 90 NE-SW 74 ESE-WNW KRK-4 85 E.W 81 N-S KRK-5 86 NE-SW 51 NW-SE KRK-6 160 E-W 105 N-S KRK-7 70 E-W 65 N-S KRK-8 40 E-W 35 N-S KRK-9 90 E-W 75 N-S KRK-10 104 N-S 86 E-W KRK-11 130 N-S 87 E-W

Height (m)

Shape

Nature

3 2.5 2.0 4.0

Crescent Circular Elliptical Elliptical

-

Circular

Habitation Habitation Habitation Habitation Quarry (4/9 mine shafts and 2 large pits)

2.0 2.5-3.0 3.0 1.0 0.6

Elliptical Elliptical Sub-circular Elliptical Circular

Habitation Habitation Habitation Habitation Quarry

2.5 2.0 1.5 2.5 1.5-2.0 1.0 ?

Circular Elliptical Elliptical Elliptical Circular Circular Circular

Habitation Habitation Habitation Habitation Habitation Cluster of vessels Iron Smelting site

3.0 2.5 2.5 0.0-0.80

Sub-circular Elliptical Circular Irregular

Habitation Habitation Habitation Quarry

3.5 3.5-4.0 2.5 3.5-4.0 3.5 >2.0 4.0 1.5 2.5 2.0 2.0

Sub-circular Elliptical Elliptical Sub-circular Elliptical Elliptical Sub-circular Sub-circular Elliptical Elliptical Elliptical

Habitation Habitation Habitation Habitation Habitation Habitation Habitation Habitation Habitation Habitation

Visited Settlements

Latitude

Longitude

Elevation

5 - Douroula 6 - Kerebe 7 - Magnimasso 8 - Sa 9 - Souma Tested and excavated settlements 12 - Kebe-Kan-Sira-Tomo 10 - Diekono 11 - Gnambakouon Sira Tomo 13 - Kerebe-Sira-Tomo 14 - Tora-Sira-Tomo

12o 37’ 26” N, 12o 38’ 33” N 12o 39’ 47” N 12o 38’ 45” N 12o 36’ 13” N

3o 18’ 23” W 3o 23’ 24” W 3o 26’ 06” W 3o 18’ 48” W 3o 16’ 12” W

271 m asl. 259 m asl. 263 m asl 258 m asl. 258 m asl

12o 37’ 31” N 12o 37’ 09” N 12o 35’ 36” N 12o 36’ 30” N 12o 38’ 07” N

3o 22’ 18” W 3o 21’ 39” W 3o 20’ 48” W 3o 20’ 20” W 3o 22’ 07” W

258 m asl 259 m asl 284 m asl. 267 m asl. 280 m asl.

15

Archaeology of West African Mound-Clusters

two mounds, 104 to 130 m long, 86 to 87 m wide, and 2 m high, both oriented north-south. The third group is made of 6 mounds with length ranging from 70 to 90 m, width from 51 to 81 m, and height from 2.5 to 4 m (tab. 2.1). Four of the mounds are oriented north-south, and the remaining two northwest-southeast. There is only one of the surveyed mounds to measure 40 m in length, 35 m in width, and 1.5 m in height. In summary, Kirikongo mounds are distributed in four size-classes. Four of the mounds were tested (Dueppen 2008), providing an interesting set of comparative material which will be discussed in the last chapter.

procedure should involve at least the probing of all the mounds found in a settlement complex. This will allow to work out a precise chronology of the settlement history backing detailed analyses of material culture, architecture, and subsistence remains. It is then and only then that variation – or lack of it – can be assigned to differences in social status. These are the principles guiding the field tactics implemented during the 1997-2000 field seasons in the Mouhoun Bend.

Gnissa Ka Tomo (12o 35’ 38” N, and 3o 18’ 39” W) at 267 m asl, located a few hundred meters northwest of the present day village of Douroula, is certainly part of the ancient Douroula settlement complex (fig. 2.1). It is comprised of three habitation mounds, circular to elliptical in shape, with a quarry. The quarry is of irregular shape and shallow, measuring 0.80 m in maximum depth (tab. 2.1). The recorded habitation mounds measure 50 to 100 m in length, 50 to 95 m in width, and 2.5 to 3 m in height.

The regional distribution of settlements is particularly interesting (fig. 2.1). Settlements are arranged in coherent patterns all over the study area. Buo-Kien-So-Tomo, BuoKien, and Bawasso are located at 3-4 kilometers from one to the next in the southeast. Douroula, Souma, Sa, and Kerebe-Sira-Tomo in the northeast are located at 5 kilometers from one to the next, and set in a quadrangular arrangement. With the exclusion of the Late Stone Age lithic scatters of Kebe-Kan-Sira-Tomo, the settlements from the central north-south axis, Kirikongo in the south, ToraSira-Tomo/Gnambakouon Sira Tomo, Kerebe-Sira-Tomo, and Diekono are more or less evenly spaced, located at 3 to 5 kilometers from one to the next. The same applies to the east-west axis, with Kerebe-Sira-Tomo on the cliff in the east, Diekono, Kerebe, and finally, Magnoumasso in the west (fig. 2.1). The settlement patterns described above point to the existence of autonomous and self-sustaining village communities, a hypothesis that will be evaluated with archaeological data in this book.

2.6 - Site location strategies: A Model

2.5 - The Excavation program The excavation program was designed to sample the available range of settlements in their different locations. Kebe-Kan-Sira-Tomo and Diekono are located in the Mouhoun River flood-plain. Kerebe-Sira-Tomo is located on the cliff dominating the river valley. And Tora-SiraTomo and its satellite Gnambakouon-Sira-Tomo are located on the topographic rise in the central part of the study area. Gnambakouon-Sira-Tomo located at approximately 600 m northeast of Tora-Sira-Tomo, was probably a small hamlet which may have been an out-growth of the latter mound complex. Kebe-Kan-Sira Tomo is the locality with scatters of Late Stone Age (LSA) lithics

Summary Elements of food production, evidence for technological innovation, and data on settlement dynamics have been recorded during the field part of the Mouhoun Bend Archaeological Project. The finds are well beyond the expectation of the investigators. A sophisticated ironsmelting technology appears to have been developed in the Mouhoun Bend between the seventh and the fifth century BC (Holl 2009a, b). The pyrotechnical performance of the uncovered furnace are not yet known. Future palaeo-metallurgical research will focus on these aspects through analyses of slag, blow-pipes, and furnace wall fragments. Faunal remains were systematically collected. The bio-archaeology component of the project is still at its preliminary stages and has just started with the detailed analysis of all the excavated human remains. Large samples of botanical macro-remains were found in their original if burnt contexts. These discoveries open unexpected perspectives on the archaeology of household in the context of mound sites. The Mouhoun valley appears to have been inhabited by groups of mobile LSA foragers. This early phase was followed by the installation of mixed-farming iron-using communities who left their intriguing mounds clusters in the landscape. Finally, a tentative social and political organization of the Mouhoun Bend settlement systems will be suggested in the last chapter of this book.

With the exception of the LSA lithic scatters site, each of the above mentioned site consists of a number of mounds of different size, with iron-working sites, and quarries. It has been suggested that settlements arranged as clusters of mounds may have been inhabited by different specialized groups such as blacksmiths, potters, hunters, bards, etc (R. McIntosh 1999; S.K. McIntosh 1999: 76). The suggestion is interesting but based on data from relatively recent social evolution of different communities of speakers of Mande languages, and ethnographic studies carried out during the last 100 years. It is however not known how far back in time evidence for such settlement system can be documented. But more important, it is not clear what can be the archaeological fingerprint of such a system. As currently used in West African archaeology, craft-specialization with its patterned social interaction-avoidance is considered to operate within a system of segregated residence; distinct mounds are thus considered to materialize such a spatial segregation, and as such are the signature of craftspecialization. The explanation is tautological. To date, no excavation program has ever been implemented to test the accuracy the hypothesis mentioned above. The testing

16

Chapter 3 Kebe-Sira-Kan-Tomo: Late Stone Age Lithic Scatters

Introduction

3.2 - Assemblages composition

This short chapter presents the material collected from Kebe-Sira-Kan-Tomo where a series of Late Stone Age (LSA) lithic scatters, was found at approximately one kilometer west of the Diekono mound complex. These stone artifacts scatters point to the presence of LSA foragers in the Mouhoun Bend sometime during the Holocene. Three of these lithic scatters were sampled with the results of the study presented in this chapter.

3.2.1. Assemblage 1 Assemblage 1 is comprised of 270 stone pieces (tab. 3.1). It includes 15 formal tools, 4 cores, with the debitage made of 230 flakes, 2 blades, and 19 chunks. A miniature ground axe in basalt was also found in the sample unit. Quartz and quartzite are by far the predominant raw materials represented in the sample (tab. 3.2), followed by chert, and finally jasper for two pieces. All the cores recorded in this assemblage are made of chert and are relatively small in size (tab.3.3), suggesting a rather intensive exploitation of the available raw material. They vary in size and shape, and measure 29 to 61 mm in length, 25 to 37 mm in width, and 15 to 25 mm in thickness. It is clear from the evidence at hand that stone knapping activities took place on the site but the accumulated by-products were later spread out by erosion agencies. The sample of recorded formal tools amounts to 17 artifacts (tab. 3.4). Points (4) and endscrapers (3) are the most frequent, followed by sidescrapers, notched pieces, used pieces, microliths, and others including a partially ground axe in granite.

3.1 - The site Kebe-Sira-Kan-Tomo, (12o 37’ 31” N and 3o 22’ 18” W) is not strictly speaking a mound site. It is located on the right bank of the Mouhoun River, at 258 m asl on a “badlands-like” surface with deep gullies. The encountered cultural remains consist of series of scatters of LSA lithic artifacts of unknown taphonomic history. Three key options are possible: 1) the artifacts can be the remains of stone knapping activities performed on the spots but slightly affected by post-depositional processes. 2) They could have resulted from the erosion of the hill slope, transported by water run-off and accumulated along the river shore. 3) Each scatter may have had a slightly different taphonomic trajectory with varying combinations of elements from option 1 and 2.

3.2.2 Assemblage 2 Assemblage 2 is much larger with 405 pieces (tab. 3.1). 348 pieces are un-retouched flakes, 28 are chunks, the rest consisting of 12 formal tools, 8 cores and 8 blades. Chert, recorded in 313 pieces, is the dominant raw material in this sample, followed by quartz/quartzite, others, and jasper (tab. 3.2). All the recorded cores are in chert. They fit in the measurements range of Assemblage 1 sample, their length varying from 28 to 50 mm, width from 22 to 44 mm, and thickness from 13 to 31 mm (tab. 3.3,). 12 formal tools have

The artifacts were found in concentrations of varying density on a “badlands-like” clayey silt sedimentary formation. The scatters, found along a narrow strip, are discrete, some larger and denser than others. Three of these scatters, termed assemblage 1, 2, and 3 were sampled with 1 by 2 m units, with all the available lithic pieces collected.

Table 3.1: Composition of LSA lithic assemblages Assemblage

Formal tools

Core

Flakes

Blades

Chunk

Others

Total

1 2 3 Total

15 11 13 39

4 8 6 17

230 348 627 1,205

2 8 15 25

20 28 26 72

2 1 2 4

270 405 689 1,364

17

Archaeology of West African Mound-Clusters

Table 3.3: Patterns of use of lithic raw material

Table 3.2: Frequencyd istribution of raw materials Assemblage

1

2

3

Total

Chert Quartz/Quartzite Jasper Others

34 228 2 6

313 85 2 3

242 442 4 1

589 755 8 10

Total

270

403

689

1,362

Chert

Quartz/ Quartzite

Others

Total

4 16 3 3 6

211 14 6

3

4 227 3 17 15

6 271 7 14 10

84 1

3 1 1

6 358 7 15 12

Cores Flakes Blades Chunks Tools

5 194 4 20 10

436 2

1 2 4 1

6 632 8 20 13

Total

573

754

18

1,345

ArtifactCategory Assemblage1 Cores Flakes Blades Chunks Tools

been recorded in the assemblage. They are divided into five artifacts categories with varying frequency. Sidescrapers and microliths with 4 specimens each are the most frequent (tab. 3.4); followed by points, endscrapers, and used pieces.

Assemblage2 Cores Flakes Blades Chunks Tools

3.2.3 Assemblage 3

Assemblage3

Finally, Assemblage 3, the largest sample with 689 pieces, includes 627 flakes, 15 blades and 26 chunks, as well as 13 formal tools and 6 cores (tab. 3.1). Quartz/quartzite was the preferred raw material followed by chert and a handful of pieces in jasper (tab. 3.2). All the cores but one are in chert. They vary in length from 31 to 48 mm, in width from 19 to 32 mm, and in thickness from 15 to 23 mm, the narrowest range in all the observed cases. On the average, cores are smaller suggesting an even more intensive exploitation of the lithic raw material in the sampling area of assemblage 3. With 6 specimens out of 13, sidescrapers are the most frequent tool represented among the formal tools from assemblage 3 (tab. 3.4). It is followed by endscrapers and microliths with 2 specimens each, and points, used pieces and others with 1 piece each. It is worth noting that assemblage 3 includes an additional interesting find, a quartz bead measuring 11 mm in length, 7 mm in width, and 6 mm in thickness, split in two during drilling.

Table 3.4: Cores measurements CoreNo

Length (mm)

Width (mm)

Thickness (mm)

35 41 29 61

28 25 20 37

25 15.5 16 15

Chert Chert Chert Chert

48 47 37 40 28 50

44 31 28 31 22 40

29 30 16 25 13 31

Chert Chert Chert Chert Chert Chert

35 40 48 33 32 31

19 21 32 23 28 21

18 21 21 23 22 15

Chert Chert Redjasper Chert Chert Chert

Rawmaterial

Assemblage1 437-1 437-2 439-1 439-3

3.3 - Stone tools production sequences

Assemblage2

All the steps involved in the manufacture of stone tools are represented in all the lithic scatters sampled at Kebe-SiraKan-Tomo (tab.3.1). As is generally the case, debitage byproducts are largely predominant. Beside a small number of cores, debitage is represented by three categories of lithic production waste, un-retouched flakes, blades, and chunks.

361-1 362-2 370-1 370-2 378-1 378-2 Assemblage3

3.3.1 Raw materials

393-1 393-2 393-3 402-1 402-2 411-3

Chert, quartz and quartzite are the predominant raw materials represented in all the studied assemblages with however significant variations in frequencies in each case (tab. 3.2). They are complemented by handful of pieces in jasper, sandstone, dolerite, granite, limestone, and laterite. The raw material procurement system was probably the result of a number of strategies which may have included chance finding of raw material dropped by the river, barter or exchange with neighbors along the river.

18

Kebe-Sira-Kan-Tomo

3.3.2 Debitage

18 mm (13 to 31 mm) in Assemblage 2, and finally, 8 mm (15 to 23 mm) in Assemblage 3.

3.3.2.1 Flakes

Cores were accordingly dropped when they were 29 to 61 mm long, 19 to 44 mm wide, and 15 to 30 mm thick, a stage at which they could hardly be grasped properly for further knapping. There are a few examples of prismatic uni-polar cores from which blades and bladelets were obtained. A few crescent-shape cores have also been recorded. In general however, partly because of the raw material in use, chert, quartz and quartzite nodules, the majority of the recorded cores are of irregular shape.

As can be expected from stone knapping stations, unretouched flakes represent the largest proportion of the lithic material collected in each of the sampled scatters (Tab.3.3). They range from 227 in Assemblage 1 to 632 in Assemblage 3. There are significant and interesting variations in the raw material represented. Each of the tested scatters seems to represent a special case of its own. Quartz/quartzite flakes are largely predominant in Assemblage 1. Chert is largely predominant in Assemblage 2. And finally, even if quartz/ quartzite is predominant in assemblage 3, the proportion of chert flakes, with 194 specimens, is significant (Tab.3.3).

3.3.3 Formal tools As can be expected from stone tools producing localities, the number of formal tools is in general very low, 41 artifacts out of 1,364 lithic specimens (Tab.3.4). Five categories of formal tools are represented in varying proportions in the sampled assemblages. Two additional categories, that of “used pieces” and “others” complement the repertoire. The former include the few cases of un-retouched flakes or blades with evidence of use and in this case essentially silica gloss. The latter is made of such artifacts as ground axes and burins.

3.3.2.2 .Blades There are very few blades with the total sample amounting to only 20 specimens. They are more or less evenly distributed among the tested assemblages with frequency varying from 5 to 8 (Tab.3.3). Chert is largely predominant with a total of 14 blades. The raw material of the remaining 6 pieces is green and red jasper, a fine-grained siliceous rock.

The number of formal tools per assemblage which ranges from 12 (Assemblage 2) to 17 (Assemblage 1), does not vary very significantly. The same tools categories are found in all the assemblages with minor variations. They tend to be small in size on the average, an important characteristic of West African Late Stone Age lithic traditions which are predominantly microlithic. Despite the small sample size, a detailed analysis of the most frequent tools, enscrapers, sidescrapers, microliths, and points, allow to outline the key characteristics of the stone-knapping tradition under investigation.

3.3.2.3 Chunks Chunks are relatively thick and bulky pieces resulting from accidental or intentional stone knapping operations. They amount to 61 specimens essentially in chert and quartz/quartzite, distributed almost evenly among the three assemblages with frequency ranging from 14 to 20 (Tab.3.3). 3.3.2.4 Cores The number of recorded cores varies from 4 to 8 (Tab.3.3). They represent the final stage of flake/blade removal when the tool-maker had decided to drop what was left of the original cobble or block of raw material. Theoretically, detailed analyses based on cores sizes can allow an insight into the decision process of the Mouhoun Bend Late Stone Age tool-makers, in this case the size threshold at which a core was abandoned. With the single exception of the red jasper core from Assemblage 3 (Tab.3.4), chert is the only raw material represented in the studied sample. Their length varies from 28 to 61 mm, a difference of 33 mm (Tab.3.4). This range is however smaller within each of the studied assemblages, 26 mm in Assemblage 1 (29 to 61 mm), 22 mm in Assemblage 2 (28 to 50 mm), and 17 mm in Assemblage 3 (31 to 48 mm).

3.3.3.1 Endscrapers There are 5 endscrapers in total (Tab.3.5). They are all in chert or quartz and measure 19 to 52 mm in length, 13 to 22 mm in width, and 3 to 11 mm in thickness (Tab.3.6). 3.3.3.2 Sidescrapers Sidescrapers are more numerous, amounting to 12 pieces (tab. 3.5) with all but two specimens in chert (tab. 3.7). They vary in length from 18 to 40 mm in Assemblage 3, 28 to 42 mm in Assemblage 2, and 23 to 45 mm in Assemblage 1. The ranges are narrower for width, 15 to 23 mm, and thickness, 3 to 10 mm. 3.3.3.3 Geometric microliths

The width presents a difference of 25 mm in all the sets, 17 mm (20 to 37 mm) in Assemblage 1), 22 mm (22 to 44 mm) in Assemblage 2, and finally, 13 mm (19 to 32 mm) in Assemblage 3. Thickness record follows a pattern of variation comparable to that of width with however narrower ranges, 10 mm (15 to 15 mm) in Assemblage 1,

The sample of geometric microliths amounts to 8 pieces with the largest set found in Assemblage 2 (tab. 3.8). Three main shapes, triangular, trapezoid, and crescent are represented with the last one largely predominant. The

19

Archaeology of West African Mound-Clusters

Table3.5: Distribution of formal tools Tool category

Assemblage 1

Assemblage 2

Assemblage 3

Total

Endscrapers Sidescrapers Microliths Notched pieces Points Used pieces Others Total *groundaxes

3 2 2 2 5 2 2* 18

1 4 4 2 1 1 13

1 6 2 1 1 1 12

5 12 8 2 8 4 4 43

Table 3.7: Mensurations of sidescrapers

Table 3.6: Mensurations of endscarpers No

Length (mm)

Width (mm)

Thickness (mm)

Raw material

No

19 52 24

14 22 13

3 11 8

Chert Chert Quartz

21

8

Chert

16

6

Chert

434-3 438-2

36

368 373-1 377-1 379

Assemblage 3 413-1

19

Thickness (mm)

Raw material

23 45

15 20

8 9

Quartz Chert

18 15 20 17

3 5 5 4

Chert Chert Chert Chert

18 20 17 16 23 15

8 8 5 4 10 4

Chert Chert Chert Chert Quartz Chert

Assemblage 2

Assemblage 2 373-2

Width (mm)

Assemblage 1

Assemblage 1 434-2 440-3 441-1

Length (mm)

30 32 28 42

Assemblage 3 389-1 389-2 403-1 403-3 405-1 410-1

recorded microliths vary in size from 13 to 32 mm for length, 11 to 16 mm for width, and 2 to 5 mm for thickness. Five of the specimens are in chert and the remaining three in quartz (tab. 3.8).

30 40 21 25 31 18

3.3.3.4 Points Eight points were recorded in the studied sample, with frequency per assemblage varying from 1 to 5 (tab. 3.9). Most of the specimens are partly broken. Their length varies from 24 to 48 mm, the width from 12 to 26 mm, and the thickness from 3 to 12 mm. Five of the specimens are in chert with the remaining 3 in quartz.

Table 3.8: Mensurations of microliths No

Length (mm)

Width (mm)

Thickness Shape (mm) material

(mm)

(mm)

(mm)

16 13

5 4

Crescent Triangle

Quartz Quartz

11 13 14 11

3 3 4 2

Trapeze Crescent Crescent Crescent

Chert Chert Chert Chert

15 15

5 3

Crescent Crescent

Quartz Chert

Raw

Assemblage 1

Summary

434-1 441-3

None of the lithic scatters sampled at Kebe-Sira-KanTomo could be dated. The range of finds which include cores, debitage, and formal tools suggests that each of the sampled area may have been a short term stone knapping workshop. But post-depositional processes have certainly altered the original configuration of the abandoned cultural remains and may even have created new patterns totally independent from intentional human behaviors. Chert, quartz and quartzite cobbles which may have been carried and dumped by the Mouhoun River appear to have been the favorite raw materials of Late Stone Age mobile foragers of the Mouhoun river valley. There are a few cases of jasper, sandstone, granite and dolerite. They too could also have been part of the river load, dropped on meanders and/or

16 20

Assemblage 2 365-2 367-1 380-1 380-2

13 32 29 22

Assemblage 3 397-1 412

20

19 25

Kebe-Sira-Kan-Tomo

Table 3.11: Mensurations of microliths

Table 3.9: Mensurations of points No

Length (mm)

Width (mm)

Thickness (mm)

Raw material

(mm)

(mm)

(mm)

material

No

12 15 23 26

6 3 12 5.5

Chert Chert Quartz Chert

15

5.5

Chert

19 16

8 3

Quartz Chert

434-1 441-1

24

13

5

Quartz

365-2 367-1 380-1 380-2

Length (mm)

Width (mm)

Thickness (mm)

Raw material

397-1 412

23 45

15 20

8 9

Quartz Chert

30 32 28 42

18 15 20 17

3 5 5 4

Chert Chert Chert Chert

18 20 17 16 23 15

8 8 5 4 10 4

Chert Chert Chert Chert Quartz Chert

Assemblage 3 389-1 389-2 403-1 403-3 405-1 410-1

30 40 21 25 31 18

16 20

16 13

5 4

Crescent Triangle

Quartz Quartz

13 32 29 22

11 13 14 11

3 3 4 2

Trapeze Crescent Crescent Crescent

Chert Chert Chert Chert

19 25

15 15

5 3

Crescent Crescent

Quartz Chert

The course of the Mouhoun River changed several times during the Holocene. It is highly probable that most of the traces left by Late Stone Age foragers were eroded or are now concealed under thick alluvium deposits. The use of quartz and quartzite for stone tools making is one of the characteristics of West African Late Stone Age traditions (Casey 1998, Neumann et al 2001). In the Oudalan province of northern Burkina Faso where LSA occupations are dated from 2100 to 1000 BC, “ most final Stone Age sites show quite simple structures with a concentration of stone artifacts mainly made of quartz and other siliceous rocks, as well as pottery.‘ (Neumann et al 2001: 328). The material recorded at Kebe-Sira-Kan-Tomo which may date from 2000 to 1000 BC fits in these LSA traditions with however the significant fact that not a single potsherd was found in the sampled lithics scatters.

Assemblage 2 368 373-1 377-1 379

Raw

following a drop in the velocity of the water current. The upper part of the Mouhoun river drainage in the south/ southwest is a dissected low plateau rich in diverse rock formations including granite, gneiss, sandstone, dolerite, etc.

Assemblage 1 434-3 438-2

Shape material

Assemblage 3

Table 3.10: Mensurations of sidescrapers No

Thickness (mm)

Assemblage 2

Assemblage 3 389-3

Width (mm)

Assemblage 1

Assemblage 1 427-1 28 440-1 30 441-4 48 442-2 25(?) 44211 2bis Assemblage 2 365-1 24 383-1 32

Length (mm)

21

Chapter 4 Diekono Settlement Complex

Introduction

4.1 - Diekono – 1

Diekono settlement complex (12o 37’ 09” N and 3o 21’ 39” W), at 259 m asl. Is located on the right-bank of the Mouhoun River at 100-150 meters from its immediate shore. The settlement consists of four distinct mounds set parallel to the river shore at distance varying from 200 to 300 meters from one to the next. Several potsherds scatters and fragmented heavy-duty stone artifacts were observed in the vicinity of the mounds cluster, more precisely in the “badlands-like” context created by deep gullying by an intermittent tributary stream. An iron-smelting site was found on the edge of one of the gullies, deeply eroded with its largest part already destroyed. The four largest mounds are all elongated in shape and oriented west - east. Three of the surveyed mounds located on the eastern side of the deep gully were extensively disturbed by an extraordinarily intensive activity of burrowing animals (hyenas, jackals, hares and rabbits). The westernmost mound located west of the gully was surprisingly not affected by burrowing animals. It is an oblong-shaped site measuring 220 m in length west - east, 110 m in maximum width north-south, and 3.00 m in maximum height above the surrounding flood-plain. Two probes were sunk at both ends of the mound, probe 1 in the east and probe 2 in the west (fig. 4.1). A third and smaller probe 3 was tested next to Probe 2.

Diekono -1 measures 12 m2, 4 m north-south and 3 m westeast. The exposed stratigraphic sequence is 1.00 to 1.20 m thick, with a storage pit dug down to 2.30 m (fig. 4.2 and 4.3). It is divided into five depositional units with, from bottom to top:

Figure 4.1: Map of Diekono

Figure 4.2: Stratigraphy of Diekono 1, west section

1. 2.30 - 1.00 m: the bottom natural deposit of brownyellow silty-clay. 2. 1.00 – 2.30 m: A light-gray ash fill of the storage pit with burnt animal bones. 3. 1.00 – 0.40 m: A series of superimposed dark-red laterite house floors. 4. 0.6/0.4 – 0.30 m: A light brown-gray silty clay, hard to compact with red gravel. 5. 0.30 – 0.00 m: A moderately hard dark brown-gray silty sand with gravel. Six occupation levels manifested by house floors and/or compact portion of courtyard with more or less scattered cultural remains were recorded. 4.1.1 Occupation I The storage pit exposed in the northeast corner of the excavated probe represents the earliest occupation of this part of the site. Its top sealed by a laterite house floor was

22

Diekono Settlement Complex

Figure 4.3: Stratigraphy of Diekono 1, north section

found at 1.30 m below the surface. It is 1.45 m deep and measures approximately 1.3 m in diameter (fig. 4.2 and 4.3). It was clearly part of an earlier occupation and may have been located in the courtyard. At the end of its use-life it was filled with refuse, probably from houses and hearths cleaning and contained a large amount of potsherds, animal bones, and burnt shells of fresh water oysters (Etheria elliptica). A charcoal sample collected from the upper part of the pit fill, at 1.30 m below the surface suggests the occupation episode to date to AD 440-759 (1430+/- 70 BP, ISGS-4296).

Figure 4.4: Occupation II floor

4.1.2 Occupation II Occupation II deposit, found at 0.80 to 1.00 m below the surface, is represented by a concentration of large potsherds, a stone door-socket, an iron slag, found on the periphery of a portion of a circular house floor made of crushed laterite (fig. 4.4). The documented house floor, dark red in color and cement-hard, sealed the storage pit in the northeastern corner of the excavated probe. The exposed portion of the round house floor measures 2.3 – 2.6 m in radius. A charcoal sample collected from the house floor dates the occupation episode to AD 541-777 (1380+/-70 BP, ISGS-4295). 4.1.3 Occupation III Occupation III was recorded at depth varying from 0.60 to 0.80 m. It is an impressive superimposition of more than ten house floors. Very few artifacts were collected from this sedimentary unit. An elongated but shallow pit containing a complete pot was found in the northwestern corner of the excavated probe (fig. 4.5). A post-hole was also found in the central part of the trial trench, with the rest of the exposure consisting of monotonous dark red superimposed house floors.

Figure 4.5: Occupation III floor

23

Archaeology of West African Mound-Clusters

4.1.4 Occupation IV Occupation IV is documented at 0.50 to 0.60 m. It consists of two sets of cultural remains: first, a portion of circular house floor, exposed on 2.70 m north-south along the eastern side of the probe, and 1.10 m west – east; and second, a series a large potsherds concentrated in the central part of the excavated probe in the western half (fig. 4.6). 4.1.5 Occupation V-VI Occupation V consists of two sets of large potsherds found at 0.30 to 0.40 m, in what may have been part of the courtyard (fig. 4.7). And finally, occupation VI associated with the top deposit appears to have been disturbed by trampling. A series of large potsherds was found stretched along the northern side of the probe. 4.1.6 Material culture A sample of 1,669 sherds weighing 66.59 kilograms was collected from the 12 m2 probe of Diekono - 1 (tab. 4.1). On the average, the general amount of sherds appears to increase from the earliest to the latest occupations, varying from 16 and 27 in occupation I, 74 to 187 in occupation II to IV, and finally to 415 to 867 in occupation V and VI at the top of the cultural sequence. As can be expected, body sherds are largely predominant, most of them decorated with twisted roulette impression ((tab. 4.1). There are a number of cases of sherds decorated with grooved and incised lines but twisted roulette impression sometime in association with grooved lines was by far the most popular decoration techniques. The technical characteristics of the recorded sherds assessed through their thickness point to a broad household use of clay vessels. Service vessels for food consumption and other beverage of the bowls and small hole-mouth pots categories tend to be small and thin walled in the range of 5 to 9 mm. They are represented in all occupation deposits with however significant variation (tab. .4.1). Cooking, food processing, and short-term storage vessels tend to have medium thick walls in the 9 to 12 mm range. They form the largest portion of the sherds sample approaching 1,000 pieces (tab. 4.1). And finally, the thick walled vessels in the range of 13 mm and above are generally used for the storage of beverage and food stuff. They are represented in all occupations pointing to the standard residential nature of this tested portion of Diekono mound.

Figure 4.6: Occupation IV floor

A number of vessel shapes was reconstructed. Two complete specimens were found in occupation II and III. The former is a low carinated pot with slightly restricted neck and everted rim decorated with twisted roulette impression and grooved lines (fig. 4.8-5). The latter is a medium size flat base bowl decorated with twisted roulette impression with entirely red-slipped inside (fig. 4.8-4). Larger vessels tend to be globular to elongated with restricted neck and more or less everted rim (fig. 4.8). Three special specimens were found in occupation IV deposit: a broken pot-lid (fig. 4.9-1),

Figure 4.7: Occupation V floor

24

Diekono Settlement Complex

Table 4.1: Diekono-I: general distribution of sherds Sample composition Stratigraphy (m)

Total (kg)

W

0.0-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40

867 415 83 74 187 27 16

Rim

Body

Base

D

ND

O

D

ND

O

25.2 17.7 6.32 3.87 10.8 1.8 0.9

50 37 14 5 17 5 3

15 13 1 2 11 -

16 2 1 1

515 289 69 57 146 18 10

25 8 5 3 -

246 59 4 2 23 3 2

7 6 1 1 1 1 -

1,669

66.59

131

42

20

1,104

41

339

18

D

ND

TR

GL

IN

TR-GL

834 393 78 66 186 24 15

40 21 14 8 15 3 1

395 240 71 54 123 17 10

11 1 20 -

1 3 -

164 82 8 11 19 6 3

1,596

102

910

32

4

292

0.0-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40

5-7 25 18 1 20 1 -

7.5-9 171 66 9 9 23 3 1

9.5-11 290 106 23 21 55 10 7

11.5-14 279 126 36 31 74 7 5

14.5 125 103 21 11 31 6 3

Total

65

282

512

558

300

Total Decoration techniques

0.0-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40 Total Thickness classes (mm)

Key: D = Decorated; ND = Non decorated; O = Others; TR = Twisted Roulette; GL = Grooved lines; IN = Incision

4.1.7 Faunal remains

the base of a footed bowl made of white clay (fig. 4.9-2) belonging to the Upper Niger pottery tradition, and finally, a decorated flat base of a bowl (fig. 4.9-3).

The recorded material culture is that of a small mixedfarming community settled near or on the shore of the Mouhoun River. The faunal remains point to an emphasis on the exploitation of aquatic resources (tab. 4.2). All the faunal remains from occupation I come from the ash fill of the storage pit. It is the only set containing sheep/goat bones as the only indicator of the presence of domestic animals. Catfish, of Synodontis sp. and Gymnarchus niloticus species, are represented in small frequencies. The eggshell fragments from Occupation III point to the presence of chicken while the 15 birds’ bones from Occupation I could not be determined to the species level. The bulk of the faunal remains is thus made of fresh water oysters (Etheria elliptica) shells and a handful of mussels (Unio sp.) with frequency varying from 2 to 47.

Sandstone is largely dominant among the recorded coarse lithic material. Most of the pieces are portable grindstones measuring 9.9 to 13 cm in length and 5.5 to 10.3 cm in width (fig. 4.10 and 4.11), a few hammer-stones (fig. 4.10-1 and 2, 4.11-2), a door-socket (fig. 4.10-4), and finally, an elongated grinder in dolerite (fig. 4.11-1) of the kind used in the grinding of karité nuts in the oil production process. A handful of iron artifacts were recorded throughout the tested cultural deposit (fig. 4.12). All but three are tangs or tangs fragments. One (fig. 4.12-4) is the distal segment of a rectangular blade. The second (fig. 4.12-10) is a small thin fan-shaped axe. And finally, the third 4.13), collected from the surface, is an arm-ring fragment.

25

Archaeology of West African Mound-Clusters

Figure 4.8: Pottery from Occupation II and III Figure 4.10: Coarse lithic artifacts

Figure 4.9: Pottery from Occupation IV and V

Figure 4.11: Coarse lithic artifacts

26

Diekono Settlement Complex

Figure 4.13: Iron arm-ring fragment

4.2 - Diekono - 2 Diekono - 2 is located at approximately 100 m west of Diekono - 1 (fig. 4.1). The tested trench is smaller, measuring originally 9 m2 with a 2 m by 1.20 m extension added later. The exposed stratigraphic sequence is 1.50 m thick, comprised of five layers with frequent superimposed dark red floors (fig. 4.14). The cultural sequence includes, from bottom to top:

Figure 4.12: Iron artifacts

Summary

1. 1.50 – 1.25 m: the bottom natural deposit of yellow silty clay overlain by a thick house floor. 2. 1.25 – 0.80/0.40 m: light brown silty clay with gravel. 3. 0.80 – 0.40 m: soft yellow sandy silt with gravel. 4. 0.40 – 0.20 m: compact and hard reddish clayey sand with large amount of gravel. 5. 0.20 – 0.00 m: loose brown gray silty sand top soil.

The absence of datable material from occupation III deposit upward is unfortunate. In summary, that part of the site appears to have been settled from the middle of the fifth century AD onwards. The pioneer settlers selected a spot already one to two meters above the surrounding flood plain. It is far from certain that the Mouhoun River was located where it is today during the first half of the first millennium AD. The accumulated cultural deposit measures only 1.3 m in thickness; it is nonetheless constituted of six distinct occupation episodes.

Five occupation episodes represented by house floors, ash dumps and other fragments of cultural remains were recorded.

Table 4.2: Faunal remains from Diekono-I Occupation Mammals Ovis/Capra sp. Fish Gymnarchus niloticus Birds Mollusks Etheria elliptica Unio sp. Undetermined Total

I

II

III

IV

V

VI

Total

23 4 15

-

5 6

-

6 -

-

23 10 5 21

7 -

2 -

5 -

22 1 -

47 2 3

-

83 3 3

49

2

16

23

60

-

150

27

Archaeology of West African Mound-Clusters

Figure 4.14: Occupation I floor Figure 4.16: The deceased buried in tightly flexed position

4.2.1 Occupation I Evidence from occupation I was found at 1.00 to 1.20 m below the surface. A series of superimposed floors, extended all over the tested probe, was exposed at 1.20 m (fig. 4.15). A hearth consisting of three large laterite blocks with a thick burnt surface in the middle was uncovered in the extension in the east side of the test trench. An ash dump was found in the northwestern corner. Finally, a burial was exposed at 1.50 m in the extension. The deceased was buried in a tightly flexed position, lying on the left side, oriented SE-NW, and facing southwest. The hands were found before the face and the body was protected with very large sherds from a purposely broken very large clay vessel (fig. 4.16 and 4.17). Occupation I is dated to AD 889-1185 (1020+/-70 BP, ISGS – 4294). 4.2.2 Occupation II Occupation II was uncovered at 1.00 m. The recorded remains consist of portions of circular house floors, one

Figure 4.17: Occupation II floor

in the NW measuring 1.20 m in maximum exposed radius and the other along the southern wall of the probe, 1.50 m in length and 0.40 m in exposed width. An upside-down large pot, a scatter of large sherds, and a concentration of fish bones were found on the latter house floor (fig. 4.18). The central feature is made of the same material as both house floors (ie: crushed laterite), but it has a northeastsouthwest elongated pear-shape, 2.30 m in length and 1.50 m in maximum width. It is very likely a portion of the “paved” courtyard delimited by both round houses yellow silt “pisé” walls. Occupation II is dated to AD 978-1243 (950+/-70, ISGS – 4293). Figure 4.15: Burial with large sherds shroud

28

Diekono Settlement Complex

Figure 4.18: Occupation III floor

4.2.3 Occupation III Occupation III was documented from 0.70 m to 0.93 m. It consists of four superimposed house floors found at 0.70 m, 0.80 m, 0.83 m, and finally, 0.93 m, all delineated by a 0.30 m wide curved yellow silt pisé wall oriented westnortheast (fig. 4.19). Two large storage vessels in relatively good state of preservation were uncovered in the extension. Occupation III is dated to AD 904-1222 (960+/-70 BP, ISGS – 4284).

Figure 4.19: Occupation IV floor

4.2.4 Occupation IV Occupation IV remains were exposed from 0.40 to 0.60 m below the surface. They consist of two distinct portions of circular house floors found in the southern half of the probe (fig. 4.20). A large grinder and a fragmented grindstone were retrieved from the surface of the largest house floor. Two large sherds were collected from the surface of the smaller one. Occupation IV ranges in date from AD 9041222 (960+/-70, ISGS – 4283) to AD 983-1260 (930+/-70, ISGS – 4282).

Figure 4.20: Pottery from Occupation I

4.2.5 Occupation V Finally, occupation V is attested at 0.20 m, at the bottom of the relatively reworked topsoil. The material is constituted of two partly damage clay vessels in upside-down position found along the southern section of the excavation probe. This final occupation of this part of the mound is dated to AD 981-1256 (940+/-70 BP, ISGS – 4281).

throughout the 1.50 m thick cultural deposit. The number of sherds per occupation ranges from 26 (occupation I) to 872 (occupation IV), pointing to a sustained increase through time (tab. 4.3). The recorded sherds are overwhelmingly decorated with twisted roulette impression, sometime associated with grooved lines. Read from the sherds thickness perspective, the bulk of the sampled pottery material belongs to the 9.5-11 and 11.5-14 mm classes, that of food processing, cooking, short and long term storage. Thicker storage vessels are well represented while thinner service vessels are marginal (tab. 4.3).

4.2.6 Material Culture Diekono-2 appears to have been inhabited from ca. AD 850 to 1250. All the installations - house floors, storage pits, and portions of courtyard - and the material culture elements recorded from this >12 m2 sampling unit points to standard domestic use. Architectural evidence indicates that the exposed houses [floors] were part of larger more integrated household units. 1,570 potsherds were collected

A number of complete or almost complete vessels were excavated throughout the deposit. The very large clay vessel used as a shroud for occupation I burial was refit in the laboratory. It was cut longitudinally into two halves and

29

Archaeology of West African Mound-Clusters

Table 4.3: General characteristics of Diekono-II sherds Stratigraphy (m)

Total (kg)

W

Rim D

Body

ND

O

D

ND

O

Base

Sample composition 0.0-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40 Total

183 648 224 150 233 106 26

3.58 14.32 5.50 20.22 10.00 3.15 0.97

12 36 19 21 27 16 4

1 9 14 3 5 2 -

1 8 5 4 2 -

118 416 134 65 145 46 8

2 61 13 5 11 13 5

51 113 34 48 31 3 7

5 1 3 1 2

1,570

59.74

135

34

20

932

111

297

12

Decoration techniques 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40

D 180 576 192 141 215 65 21

Total

1,390

ND 3 72 28 9 17 35 5

TR 104 458 120 46 92 52 9

169

5-7 28 12 1 2 3 46

GL 4 6 7 28 4 -

C 2 6 -

IN 7 -

TR-GL 18 95 25 7 31 6 1

881

49

8

7

183

7.5-9 16 118 41 11 41 15 1

9.5-11 56 197 61 38 59 28 -

11.5-14 80 218 73 51 88 47 4

>14.5 31 129 37 49 38 13 4

250

439

561

301

Thickness classes (mm) 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40 Total

Key: D = Decorated; ND = Non decorated; O = Others; TR = Twisted Roulette; GL = Grooved lines; IN = Incision

measures 81 cm in height, 65 cm in maximum diameter, 54 cm in mouth diameter, and 2 cm in wall thickness. It is an exceptionally large vessel that was diverted from its initial storage use. It is entirely decorated with twisted roulette impression combined with grooved lines. These grooved lines are designed as triangular “sail-boat” motifs on the upper half of the vessel underlined by two horizontal and parallel lines. Two additional large vessels were recorded in occupation I deposit (fig. 4.21-3, and 4.21-4). One, a globular red-slipped pot with restricted neck and everted rim decorated with panels of grooved lines (fig. 4.21-3) and the other, a spherical pot with twisted roulette impressions (fig. 4-21-4). Two bowls of slightly different shape, 10 to 15 cm in mouth diameter, 7 to 10 cm in total height and 0.7 to 0.9 cm in wall thickness, decorated on rim with comb stamping (fig. 4.21-1) and roulette impression (fig. 4.21-2), complement occupation I pottery sample. Two bowls and one globular pot were recorded from occupation II deposit (fig. 4.22). The globular pot, 23 cm in height

and 24 cm in mouth diameter has a restricted neck and everted rim. It is decorated with panels combining twisted roulette impression and grooved lines (fig. 4.22-3). One of the bowls represented by a rim-sherd is decorated with comb-dragging on rim and red-slipped horizontal and vertical bands (fig. 4.22-1). The remaining specimen, an open bowl, is decorated with comb-stamped horizontal lines on rim (fig. 4.22-2). A broader range of vessels shape and size was found in the five specimens occupation III sample (fig. 4.23). There are two small flat base bowls; one open and plain with finger impression around the base (fig. 4.23-4), and the other bell-shaped and elaborately decorated with red slip, comb stamping, twisted roulette impression and grooved lines (fig. 4.23-5). The small 15 cm in mouth diameter hole-mouth pot (fig. 4.23-1) is decorated with comb-stamping on rim underlined by a red-slipped horizontal band. The large storage vessel (fig. 4.23-6) measures 80 cm in height, 58 cm in mouth diameter, and 1.4 cm in wall thickness. It is a rounded base elongated vessel

30

Diekono Settlement Complex

Figure 4.21: Pottery from Occupation II

Figure 4.22: Pottery from Occupation III

Figure 4.23: Lithic artifacts

31

Archaeology of West African Mound-Clusters

Figure 4.25: Diekono 3, the coarse stone artifacts

occupation I (fig. 4.24-3), and two portable grindstones (fig. 4.24). Occupation III grindstone in dolerite, measures 23 cm long and weighs 1.8 kilograms. That from occupation V in sandstone is larger, measuring 32 cm long, 16 cm wide, and 8.4 cm thick, weighting 6.8 kilograms (fig. 4.24-1). Three iron artifacts including an arrow point (fig. 4.25-1) and two tang’s fragments (fig. 4.25-2 and 3) were found in occupation V, III, and I deposits. 4.2.7 Faunal remains The sample of faunal remains collected from Diekono-2 is rather small with only 158 pieces including 5 undetermined specimens (tab. 4.4). The amount of remains per occupation varies considerably, from 53 in occupation I to 2 in occupation V. The general impression which emerges from a quick glance at the data is that of a heavy reliance on aquatic resources, particularly fish and fresh water mollusks. Mammals remains, amounting to 3 bones of Ovis/Capra sp. and Bovidae sp., are present in occupation I sample which is not only the largest but also the most diversified. All the recorded fish remains, varying in frequency from 23 (Occupation II) to 1 (Occupation III), are those of bottom-feeders Siluridae, like Gymnarchus niloticus and Synodontis niloticus. Fresh-water oyster shells were recorded in all occupations deposits with frequency ranging from 53 (Occupation I) to 2 (Occupation V). In hinterland river contexts, colonies of freshwater oysters and mussels are generally more accessible when water level

Figure 4.24: Iron artifacts

with slightly restricted neck and everted rim, decorated with a combination of twisted roulette impression and grooved lines. The roulette was used to delineate a horizontal narrow band on the vessel neck. A “sail-boat” motif underlined by two horizontal and parallel lines was designed with grooved lines. Finally, there is a shouldered jug (fig. 4.23-3) with a particularly elaborate decoration combining twisted roulette impression and red-slipped horizontal and oblique bands. The other components of the material culture repertoire, metals and lithics, are poorly represented. A few pieces were found scattered throughout the deposit. The lithic sample includes three specimens, one sandstone grinder from

32

Diekono Settlement Complex

Table 4.4: Faunal remains from Diekono-II Occupation

I

II

III

IV

V

Total

(0.9-1)

(0.7-0.8)

(0.5-0.6)

(0.2-0.4)

(0.20)

1 2

-

-

-

-

1 2

21 13

23 -

1

-

-

21 23 14

1 2 1

-

-

-

-

1 2 1

Etheria elliptica Undetermined

53 2

13 -

6 -

14 3

2 -

88 5

Total

96

36

7

17

2

158

Mammals Ovis/Capra sp. Bovidae sp. Fish Gymnarchus niloticus Synodontis niloticus Others Reptiles Undetermined snake Varanus niloticus Birds Malacofauna

is low and this occurs during the dry season. The skewed distribution of faunal remains which are essentially made of fish bones and fresh water oysters’ shells may have resulted from a seasonal occupation of the tested Diekono -2 mound portion, and this for purpose of intensive dry season fishing and gathering of mussels and oysters.

The find is intriguing. There are two competing possibilities: 1) a trader-craftsman hoard or 2) a ritual deposit in a cultic place. The first possibility can be derived from a number of facts: 1) the diversity of raw materials represented. Some like the dolerite, are not available in the immediate region but can be found in the upper portion of the Mouhoun river drainage in the south-southwest. 2) The broad range of artifacts represented as if collected for recycling. And, 3) the presence of the river where the use of dug-outs may have facilitated the movements of people and goods along the Mouhoun river drainage.

Summary Diekono-2 area was inhabited from AD 850 to 1250, for approximately four centuries. The location of the site, in the floodplain and on the shore of the Mouhoun River, and the faunal remains made predominantly of Siluridae sp. bones and fresh water oysters shells, all point to a possible intensive dry season exploitation of the river’s resources by a small mixed farming and fishing community.

The second possibility is anchored on local traditional religious practices still alive despite centuries of Islam. According to these practices, allied lineages have a sacred tree grove where ancestors are represented and honored as regularly as possible. Cutting trees or collecting wood from such places is taboo. Each of the honored ancestors is said to be represented by one stone piece. A new rock piece is added to the sacred pile at the death of any lineage member who is considered to have achieved the status of ancestor. The ancestors are fed periodically, given offerings and libations at different times of the year, and consulted by their descendants when necessary. Such sacred tree groves can still be seen in Douroula and other villages of the study area today.

4.3 - Diekono-3 The third excavation unit, Diekono-3, is a small exposure measuring 1.10 m north-south and 1.20 m west-east geared to salvage two collections of stone artifacts concealed in two large clay containers, already partly exposed by erosion (fig. 4.26). The unit is located at some thirty meters northeast of Diekono-2, at the very edge of the mound. The bottom of both vessels was reached at 0.30 m below the surface. No archaeological deposit was found underneath. The larger container measures 60 cm in diameter and the smaller one, a relatively large size pot, 30 cm in diameter. The recorded stone pieces consist of a broad range of artifacts types, broken, worn-out, as well as complete specimens of grinders, pestles, hammer-stones, grindstones, made of different kinds of raw materials (tabs. 4.5 and 4.6).

It is not easy to find out which one of the option suggested above is the most plausible. Data can nonetheless be harnessed to assess the strongest model. The tradercraftsman model should display aspects of “economic” rationality. One would expect the collected rock pieces to be big enough to allow for effective recycling. In that case large rock pieces should be predominant in the sample

33

Archaeology of West African Mound-Clusters

under investigation. The length, width, thickness, and weight of each of the rock piece were recorded. Weight appears to be the best bulk/size indicator. Using sandstone, the most frequent raw material represented in the collected assemblages (tab. 4.7), one can observe that pieces weighting less than 500 grams are largely predominant in both containers. The distribution varies significantly in both cases (fig. 4.27 and 4.28). The predominance of small rock pieces and the relative insignificance of pieces larger than 500 grams do not support the trader-craftsman hypothesis.

In other words, the recorded stone pieces are generally too small to allow for an effective recycling into other types of coarse stone artifacts. The ritual hypothesis appears much more plausible. If ancestors are honored with a certain regularity based on a social and ritual calendar, one would expect a certain redundancy in the frequency of rock pieces in the samples under investigation. It is the case at two levels. Container 1 has 81 rock pieces, twice the amount found in container

Table 4.5: Distribution of stone artifacts from container 1 Raw material

D

S

Q

G

Sy

O

Undet.

Total

Grinder Grindstone Hammerstone Others Undetermined

1 1 1 -

9 19 4 35

1 -

1 1 1

1 2 -

2 1

1 1

13 24 2 5 38

Total

3

67

1

3

3

3

2

82

Key: D = Dolerite; S = Sandstone; Q = Quartzite; G = Granite; Sy = Syenite; Undet. = Undetermined.

Table 4.6: Distribution of stone artifacts from container 2 Raw material

D

S

Q

G

Sy

O

Undet.

Total

Grinder Grindstone Hammerstone Others Undetermined

1 -

12 8 1 13

1

1 -

-

1 1 -

1 1

2 13 10 1 15

Total

1

31

1

1

-

2

2

41

Key: D = Dolerite; S = Sandstone; Q = Quartzite; G = Granite; Sy = Syenite; Undet. = Undetermined.

Figure 4.26: Distribution by weight of container 1 rocks

34

Diekono Settlement Complex

Figure 4.27: Distribution by weight of container 2 rocks

2 (tabs. 4.5 and 4.6). The number of sandstone pieces varies from 33 in container 2 to almost the double (62) in container 1. There is an unmistakable proportionality which may have resulted from a sequential use of the containers. Container 1 may have been the first to be put into use. It was filled after a certain time. Container 2 was then put into use, resuming the same ritual practice at the same rhythm.

Two iron arrowheads, one tang fragment, and a short segment of an iron arm-ring were collected in the tested deposit. The arrowhead measure 9.5 to 8.2 cm in length with square to circular section. The arm-ring fragment is only 2.5 cm long and presents a circular 0.5 cm in diameter section. Two interesting and complete stone artifacts were found in both excavated containers, among the broken pieces. One is a kwe, a sandstone piece with a bored center probably used as weight for digging sticks. It measures 6 cm in diameter, 4.1 cm in thickness, and weighs 200 grams. The other is a very small 4.3 cm in diameter and 2.8 cm thick grinder made of quartzite with a shallow depression on one side.

In summary, the location of the uncovered rock artifacts sets on the edge of the site and in an area devoid of cultural deposit point more toward the possibility of an intentional ritual act. The option of trader-craft person recycling hoard is not totally irrelevant but is weaker than the ritual-cultic one outlined above.

4.3.2 Summary

4.3.1 Material culture

No datable material was found in Diekono-3, making it difficult to fit this unit within the time frame of Diekono mound complex occupation. However, the distribution of decoration techniques similar to that of Diekono-1 occupation III and IV and Diekono-2 occupation III point

The breakage of the uncovered containers, the erosion of the mound, and the ensuing aggradation around the clay vessels which contained the stone pieces collections triggered the accumulation of potsherds and other artifacts in the area of Diekono-3. The top 20 cm of the tested deposit was devoid of any piece of cultural remains. All the material, 178 sherds weighing 4.79 kilograms, a handful of iron artifacts, and 10 fragments of fresh water oysters (Etheria elliptica) shells, were collected at 0.20 -0.40 m below the surface.

Table 4.7: Distribution of sandstone artifacts into weight classes Class (grams)

Most of the recorded sherds, 159 out of 178, are decorated. Twisted roulette impression found on 127 sherds is overwhelmingly dominant, followed far behind by twisted roulette combined with grooved lines with 16 specimens, and 3 cases of comb-stamped sherds. Relatively thick sherds, in the range of 11.5-14 and >14.5 mm, amounting to 122, are the most frequent in the recorded sample.

35

Container1

Container2

Total

1000

9 11 14 17 5 3 5

7 14 6 4 1 1

16 25 20 21 6 3 6

Total

64

33

97

Archaeology of West African Mound-Clusters

to the end of the first millennium-beginning of the second millennium AD as the best estimate for the chronology of Diekono-3 ritual and cultic deposit. Conclusion The occupation sequence recorded at Diekono-2 ranges in time from ca AD 850 to 1250, approximately four centuries. The area of Diekono-1 was settled earlier from the middle of the first millennium AD. The lack of dating material from the major part of Probe 1 cultural sequence precludes any precise assessment of the contemporaneity between all the occupation episodes documented at both ends of the mound. Detailed analyses of the collected cultural remains provide some clues. In general, the western and largest mound of the Diekono settlement complex was settled from ca AD 500 to ca 1250. It is not known if the whole settlement was abandoned in the middle of the thirteenth century AD, or if human habitation simply shifted from the largest western mound to the smaller and distinct eastern components of the complex. Local conditions, including the possibility of wetter marshy environment could have led to the abandonment of the Mouhoun flood-plain in favor of higher location. As will be shown later, Tora-Sira-Tomo/ Gnambakouon-Sira-Tomo and Kerebe-Sira-Tomo mound complexes witnessed an important growth spurs in the first centuries of the second millennium AD. Some of the new inhabitants may have come from settlements like Diekono.

Figure 4.28: Diekono 3 iron artifacts

36

Chapter 5 Tora-Sira-Tomo Settlement Complex

Introduction Tora-Sira-Tomo (TST) settlement complex (12o 35’ 07” N and 3o 22’ 07” W) at 280 m asl., is located at approximately five kilometers southwest of Kerebe-Sira-Tomo. It is the largest settlement complex found in the study area, with 17 distinct mounds and special purpose-sites. It is spread over 900 m west-east and 500 m north-south, approximately 45 ha in total surface extent (fig. 5.1). Tora-Sira-Tomo 3, the largest single mound of the whole complex is stretched along the northern edge, and measures 260 m in length west-east, and 120 m in maximum width north-south. All 17 mounds were tested after three field seasons organized in 1997, 1999, and 2000. Five, Tora-Sira-Tomo 1, ToraSira-Tomo 2, Tora-Sira-Tomo 4, Tora-Sira-Tomo 9, and Tora-Sira-Tomo 17, are special purpose sites. The first is an iron-smelting site. The second a quarry, the third a cloth dyeing workshop, the fourth a special burial site, and the fifth, a karité oil production site. The remaining fourteen mounds appear to have been standard habitation mounds with varying intensity of occupation.

Figure 5.2: Contour map of Tora-Sira-Tomo 1

Figure 5.1: Map of Tora-Sira-Tomo settlement complex

37

Archaeology of West African Mound-Clusters

Figure 5.3: Map of the excavation of Tora-Sira-Tomo 1

5.1 - Tora-Sira-Tomo - 1 Tora Sira Tomo 1 is located at approximately two hundred meters southwest of the main and largest mound of the complex. It is sub-circular in shape, measures 18 m in length, 15 m in width, and 1.30 m in maximum height (fig. 5.2). The total of the excavated sample amounts to 65 m2 (fig. 5. 3) and covers the whole range of archaeological remains associated with iron-smelting. A large number of clay vessels set in upside down position were found disposed off along the northern part of the iron-smelting complex. The excavation of this portion of the site was divided into several sub-units, A in the northwest, E in the center-east, B and F in the east

Figure 5.4: Stratigraphy of TST 1 probe A

5.1.1 Probe A Probe A located in the northwest edge of the site measures 2 by 2 m. The archaeological deposit is 0.25 to 0.30 m thick and is comprised of three sub-horizontal layers (fig. 5.4): Layer 3 - 0.00 – 0.15 m: Dark brown loose silty clayey topsoil. Layer 2 - 0.15 – 0.20 m: Compact and hard brownyellow clay. Layer 1 - 0.20 – 0.30 m: Brown-reddish laterite gravels.

Figure 5.5: View of the vessels at TST 1

The uncovered archaeological remains were found in layer 3, laid on layer 2 surface. They consists of five clay vessels set in upside-down position at 0.50 to 0.30 m from one to the next (fig. 5.3). The lower part of the recorded vessels was not preserved (fig. 5.5). They all belong with minor variations to the same shape/size category. That of constricted neck pots with everted rim, decorated with twisted roulette impression (fig. 5.6). The mouth diameter ranges from 24 to 22 cm and wall thickness from 7.5 to 10 mm.

5.1.2 Probe B Probe B in the east measures 6 m2. Its stratigraphic sequence is similar to that of probe A and contains a concentration of 13 clay vessels, all found in upside down position. The curvilinear arrangement of the uncovered pottery (fig. 5.3 and 5.5) suggests the presence of a round hut/shelter which may have measured 2.20 to 2.50 m in diameter. Three

38

Tora-Sira-Tomo Settlement Complex

Figure 5.6: Vessels from TST 1 probe A

additional pots were found on the east flank of the main concentration. The pottery sample from Probe B is more diverse in shape and size (fig. 5.7 and 5.8). It includes a small pot with straight rim, globular pots with constricted neck and everted rim as well as a flat base specimen. Decoration patterns, which combine twisted roulette impression, horizontal, oblique, and vertical grooved lines, and herringbone motifs, are richer. Figure 5.7: Vessels from TST 1 probe B

Figure 5.8: Vessels from TST 1 probe B

39

Archaeology of West African Mound-Clusters

Figure 5.9: Stratigraphy of TST 1 probe C

5.1.3 Probe E

Layer 1 - 1.30 – 1.50 m: Light brown grey sediment with laterite gravels.

Probe E oriented NW-SE measures 5 m in length and 3 m in width and contains a total of 22 vessels more or less evenly distributed all over the tested 15 m2. The stratigraphy is similar to that of probes A and B, with however a thinner topsoil that was probably more eroded on this gently sloping part of the site. The recorded pottery sample is comprised of two main size/shape categories (fig. 5.9); one made of globular to slightly elongated vessels with more or less constricted neck and flared rim generally decorated with roulette impression and grooved lines. The other includes globular pots with inverted rim and decoration made of herringbone, incised wavy or straight lines, confined in most cases to the upper part of the vessel. There are very few significant variations in the characteristics of the sherds collected in the archaeological deposit from Probe E (tab. 5:1). The general distribution of vessels parts, decoration techniques, as well as sherds thickness present a similar pattern of variation in both layers suggesting little change in the material culture of the user of this iron-smelting site.

The base of a relatively large iron-smelting furnace was found at 0.45 m below the surface (fig. 5.10 and 5.11). A line of red bricks exposed along the trench section suggests that the uncovered furnace may have measured at least 3 m in height. A fire hardened surface was exposed on the eastern side of the furnace, the mouth of which was oriented northeast. The preserved furnace’s base measures 1.80 m in diameter. Its wall is made of two layers, an inner and an outer one, each built with superimposed irregularly shaped clay lumps. It is 0.20 m thick and was preserved up to a height of 0.25 m. A group of broken and complete tuyeres was found on the north flank of the furnace, and a set of three clay vessels exposed on the fire-hardened surface on the east flank.

5.1.4 Probe C Probe C – the furnace probe – is located at the center of the site and cuts through the thickest portion of the archaeological deposit (fig. 5.2). It measures approximately 40 m2 and consists of a 7 x 2 m trench oriented NW-SE abutting a larger SW-NE 7.5 x 4 m excavation unit. Probe C stratigraphic sequence measures 1.50 m in maximum thickness, at the highest point of the whole site (fig. 5.9). It is made of 4 layers: Layer 4 - 0.00 – 0.60 m: Brown-grey silty clay topsoil, partly a termite mound Layer 3 - 0.60 – 0.80/0.90 m: Light brown grey sediment from the collapsed furnace, including slag, blow pipes, and furnace bricks fragments. Layer 2 - 0.80/0.90 – 1.30 m: Brown-yellow hard and compact clay Figure 5.10: The top of TST 1 furnace in probe C

40

Tora-Sira-Tomo Settlement Complex

Table 5.1: TST-1: General distribution of sherds in Probe C Depth (cm)

n (kg)

W

Rim

Body

D

ND

U

D

ND

U

Sample Composition 0-20 20-40

210 584

5.2 5.9

9 5

12 18

4 4

64 215

35 56

86 286

Total

794

11.1

14

30

8

279

91

372

D

ND

TR

GL

Other

Decoration Techniques Parameters

Composite TR-GL

Other

0-20 20-40

120 294

163 510

47 74

57 175

1 17

1

10 25

5 2

Total

414

673

121

232

18

1

35

7

4.5-7

7.5-9

9.5-11

11.5-14

>14.5

0-20 20-40

17 27

75 237

81 224

31 83

8 14

Total

54

312

305

114

22

Sherds thickness Class (mm)

the proximal end and 5-6 cm at the distal one. The more or less balanced arrangement of the tuyeres coupled with the eastern orientation of the furnace mouth suggests that this installation was a natural draft furnace. The circular white chalky deposit at the center of the furnace is presumably made of the remains of the fluxing material used in the smelting process (fig. 5.12). It is not yet known if this was an intentional addition to the furnace, or the result of the types of wood used for fuel (HAALAND & SHINNIE 1985; SCHMIDT 1996). In any case, fluxing material optimizes the use of fuel by lowering the temperature at which the iron ore starts to melt. Surprisingly, the bottom slag is relatively small in size. With the exposure of the bottom slag, it was thought that the bottom of the excavated furnace was reached. But the structure of the smelting installation was much more complex and interesting than expected. It had many interesting hidden elements and the rest of the furnace was still to be found 1.20 m below ground (fig. 5.13), some 0.60 m below the level of horizontally laid tuyeres. A tronconically shaped underground chamber was dug in the natural brown-gray silty clay deposit. It measures 0.60 in diameter at bottom with the base and wall lined with a mixture of crushed laterite gravel and clay. Fourteen vertical but slightly tilted tuyeres arranged into two distinct sets were found in this part of the furnace installation. The western set has six tuyeres, and the eastern one eight (fig. 5.13 and 5.14), with the central space filled with brick fragments and termite nest material. The tuyeres found at the bottom of the furnace had no direct connection with the combustion chamber. The six specimens from the western set measure 26 to 14 cm in length and 10 to 13

Figure 5.11: The tuyeres arrangement in the furnace

Slag, bricks fragments, and broken tuyeres filled the inner part of the furnace. This deposit was accumulated on top of a tuyeres level, reached at 0.60 m below the surface, above a relatively thin bottom slag (fig. 5.10). These eight relatively well preserved tuyeres converge to the center of the furnace with in the middle a mass of burnt clay, bricks, and slag, capped with a whitish 5 to 10 cm thick circular chalky deposit (fig. 5.11). The preserved segments of the eight tuyeres are more or less organized into pairs. They measure 30 to 40 cm in length and 10 cm in diameter at

41

Archaeology of West African Mound-Clusters

cm in maximum thickness. The eight from the eastern set were longer on the average, 26 to 40 cm with a narrower thickness range of 11-12 cm. A further examination of these tuyeres has shown that they were not used in any smelting process yet and were filled with easily removed sediment. A review of the literature on African iron metallurgy (bocoum 2001; coulibaly 2006, haaland & shinnie 1985; holl 2009, kiethega 2006, 2009; martinelli 2008; schmidt 1996) has failed to provide comparable evidence. There are two possible options to understand and explain the structure of TST-1 furnace.

In option 2, the underground chamber filled with termite nest material and hollow blow pipes may have been used as a “heat-trap” to prevent sudden temperature drops in this natural draft furnace in case of sharp reduction of wind velocity. Whatever the case however, the furnace structure uncovered at TST - 1 is a sophisticated piece of craft engineering. There is no known case in African literature on metallurgy, in archaeology as well as ethnography (Bisson et al 2000; Bocoum 2001; Coulibaly 2006; Haaland & Shinnie 1985; Holl 2009; Kiethega 2006; Martinelli 2008). The pyrotechnological performance of the whole installation has yet to be studied in detail. The furnace was a natural draft one, operating without below blowers on the natural strength and persistence of the dry season northeastern wind (the harmattan). That is why the furnace mouth is oriented NE.

In option 1, it is possible to suggest that the iron-smelters had developed a system allowing for a constant supply of ready to use blow-pipes, an important economy of scale. A new supply of dry tuyeres was set to be fired below the combustion chamber, taking advantage of the high temperature generated by the furnace during the iron ore smelting process (fig. 5.14). Once the process is completed, the furnace is left to cool, the bloom is collected, and the new load of fired tuyeres is recovered to run the next shift. This scenario faces a serious hurdle; most of the time, tuyeres are not fired prior to their use in the smelting process (Bocoum 2010: personal communication). They are dried and set in the furnace when needed.

However, the greatest surprise of the excavation of TST-1 smelting site is the radiocarbon date obtained from a large charcoal sample collected on the fire-hardened surface at less than one meter from the furnace mouth. The reading shows the smelting site to date to 501 – 386 BC (1 sigma)

Figure 5.12: The furnace mouth on its east flank

Figure 5.13: The underground tuyeres firing chamber at the bottom of the furnace

Figure 5.14: Section of the base of TST 1 furnace below the tuyeres level

42

Tora-Sira-Tomo Settlement Complex

or 761 – 212 BC (2 sigma), (ISGS – 4349, 2360+/-70 BP). This is clearly an early case of iron-smelting furnace in this part of West Africa, west of the Niger-Nigeria where earlier iron-smelting sites have already been recorded (Grebenart 1985; Okafor 1993; Holl 1993, 1997, 2009a, b; Holl & Kote 2000). The technical expertise involved in the conception, construction, and operation of the TST-1 furnace is very impressive indeed. 5.1.5 Probe D Probe D is located in the southwest of the smelting site and measures 2 by 1 m. The test has shown all the southern half of the site to be a slag heap where all the by-products of the smelting process were discarded. The exposed stratigraphic sequence is 1.00 m thick made of three layers (fig. 5.15 a and b): Layer 3 (0 – 0.50 m): light brown grey sediment made of slag, blow-pipes fragments, and furnace wall’s bricks Layer 2 (0.50-0.80 m): compact brown-yellow clay Layer 1 (0.80-1.00 m): Brown-grey clay with gravels. A number of miscellaneous finds were recovered from layer 3 in Probe D, two hammer-stones, one in dolerite and the other in sandstone, a sandstone grinder, tuyeres fragments, and a 8 cm long fragment of an undetermined iron object (fig. 5.16). The site seems to be partitioned into two parts, with the furnace in the middle dividing east-west line. The set of complete or near complete vessels were found exclusively in the northern half and iron ore smelting by-products were discarded in the south half. This site layout can be explained by the intermittent, very likely seasonal, use of the smelting site. At the peak of the production season, natural draft furnace can be in operation for days. Working crews take shifts in order to keep the momentum and as such may have camped nearby in lightly built shelters. The recorded pottery is predominantly made of relatively large liquid containers in which the beverages necessary for the working crews were stored. The variation in decoration patterns can have many probable explanations: One the hand, working crew members may have been recruited from different surrounding settlements, serviced by different potters. They brought with them the material they used in their respective villages. In this case, the recorded diversity reflects the elasticity of the actual pottery offer. On the other hand, the smelting site may have been in use for a relatively long period which may have witnessed some variation in the patterns of pottery decoration. A varying combination of both suggestions outlined above is also a genuine possibility.

Figure 5.15: Stratigraphy of the slag heap in probe D

excavation measuring 100 m north-south and 25 m in maximum width, located at approximately 500 m south of Tora-Sira-Tomo 1, the iron-smelting site. The depth of the quarried area varies from 0.35 to 2.00 m with the excavated part colonized by a dense vegetation of grass, shrubs and trees. On the average and relying on 1.75 m mean depth, some 1,468.75 cubic meters of laterite were extracted from TST-2 quarry. The exposed laterite section present two layers with the top one measuring 1 m in thickness on the

5.2 - Tora-Sira-Tomo-2 Tora-Sira-Tomo-2 is the quarry from which iron ore and the laterite used in construction, particularly for flooring, was collected. It is a large and impressive open V-shaped

43

Archaeology of West African Mound-Clusters

average (fig. 5.17). Two exploratory 1 by 1 m probes were sunk in shallow sediment deposits at the foot of the laterite wall. They both failed to produce any piece of cultural remains beside two worn out potsherds. 5.3 - Tora-Sira-Tomo-3 Tora-Sira-Tomo-3, the largest mound of the whole complex is stretched along the northern side of the cluster. It is oriented east-west and measures 240 m in length and 155 m in maximum width. Two excavation probes were sunk at both east and west ends of the site in order to achieve a more balanced assessment of the mound’s occupation history. 5.3.1 Tora-Sira-Tomo - 3 -East Figure 5.16: Miscellaneous finds from probe D

Tora-Sira-Tomo-3-East unit covers 3.5 x 3 m and has a 2.20 m thick cultural deposit with the following layers (fig. 5.18): 16 - 0 – 0.30 m: Dark-brown grey loose silty topsoil with gravels 15 - 0.30 – 0.50/0.80 m: Yellow brown grey clayey sand with brown gravels. 14 - 0.50 – 0.80 m: Brown grey silty sand interspersed with thin ash layers 13 - 0.50 – 0.80 m: Reddish brown grey silty sand with laterite blocks 12 - 0.80 – 0.90 m: Superimposed dark red laterite house floors 11 - 0.90 – 1:00 m: Light grey ashy fill 10 - 1.00 – 1.15 m: Light brown clayey sand 9 - 1.15 – 1.20 m: Brown grey ashy house floor 8 - 1.20 – 1.70 m: Light brown silty clay with gravels 7 - 0.90 – 1.75 m: Brown-yellow compact clay wall collapse 6 - 1.75 – 1.80 m: Dark reddish laterite house floor 5 - 1.80 -1.90 m: Compact brown-yellow clay 4 - 1.90 – 1.95 m: Dark reddish laterite house floor 3 - 1.95 – 2.05 m: Brown-yellow compact clay 2 - 2.05 – 2.10 m: Dark reddish laterite house floor with a light grey ash lense 1 - 2.10 – 2.25 m: Bottom dark brown clay.

Figure 5.17: TST 2, the quarry section

foundation’s rituals. They consist of a 0.50 m in diameter laterite slab (fig. 5:19) in the center, another laterite block and a sandstone grinder along the south section, and two children skulls in the northeast. The age of the children range from 4 to 7 years and one of the skulls was found in a hole-mouth pot in the northeast angle of the tested unit. Three body sherds, one plain and the other decorated with roulette impression and grooved line, were found in the deposit (fig. 5-20). The dwelling features associated with these ritual deposits were not exposed in the probe and are very likely to be found on any side of the excavation unit.

The lower two/third of the cultural deposit from level 11 to 2 resulted from successive episodes of house construction. The upper third points to a possible change in the use of this part of the mound prior to its abandonment as will be shown below. The occupation sequence will be discussed from the earliest to the latest occupation episodes. The excavation probe was reduced to a 1.5 x 1.5 m test unit located in the northwest from 2.05 to 2.25 m below the surface.

5.2.3 Occupation II Occupation II is documented at 1.80 m. It includes three features and a portion of the courtyard of a domestic unit (fig. 5:21). Feature 25 and 26 are pink-red round house floors while feature 27 is a collapsed forge furnace made of irregularly shaped clay-bricks (fig. 5:22). The furnace measures 0.70 m in diameter at base and abuts the northwest side of feature 25. The chimney had collapsed westward creating a jumble of pink-yellow fire-hardened clay-bricks. A large piece of iron slag was found on the floor of feature 26. The total amount of cultural remains is rather low

5.3.2 Occupation I Occupation I deposit sits on top of the natural bottom dark brown clay. The uncovered cultural remains found at 2.00 to 2.20 m below the surface point to the practice of

44

Tora-Sira-Tomo Settlement Complex

Figure 5.18: Stratigraphy of TST 3-East

suggesting a high level of maintenance of the domestic unit. The uncovered pottery includes one plain sherd, a small straight side bowl, a slightly constricted neck pot, and a necked pot with everted rim (fig. 5.23). A sandstone grinder complements the artifacts’ sample from occupation II. 5.2.4 Occupation III Occupation III is documented at 1.50 m and consists of two partially exposed dark reddish round house floors. The relative positioning of the houses, with the preserved floors arranged along a NW-SE axis, is similar to that documented in the previous occupation II. The exposed portion of the NW floor, feature 15, measures approximately 2 m in radius (fig. 5.24). It is made of two distinct and superimposed crushed laterite floors pointing to a sustained and careful maintenance of the dwelling features. It was clearly a storage house as indicated by the large number of crushed and well preserved clay vessels, sandstone grinders, and laterite blocks used as pots’ stands (fig. 5.25). A small portion of the second floor measuring less than 1 m in diameter was exposed in the southeastern corner of the probe. It is partly impacted by an intrusive burial from a later occupation IV.

Figure 5.19: Occupation I floor

The diversity and amount of cultural remains uncovered in feature 15 points to a sudden, unplanned abandonment of the compound. The recorded domestic kit is particularly rich and contains a broad range of pottery vessels, as well as stone and iron tools (fig. 5.26). The pottery repertoire consists of five distinct shapes of bowls decorated with

45

Archaeology of West African Mound-Clusters

Figure 5.20: Sherds from Occupation I

Figure 5.23: Pottery from Occupation II

Figure 5.21: Occupation II floor

Figure 5.24: Occupation III floor

lines of roulette impressions, grooved and/or incised lines (fig. 5.26 a – e). Medium to large size cooking vessels, globular to slightly elongated with everted rim with mouth diameter varying from 20 to 37 cm, are all decorated with twisted roulette impression associated with grooved lines in a number of specimens (fig. 5.26 f – k). One of the vessels (k) was used for steam-cooking. The remaining specimens (fig. 5.26 l – q) are globular to elongated vessels with their lids, used for the storage of grain and liquids. 5.2.5 Occupation IV Occupation IV surface was exposed at 1.30 m. It consists of a small portion of a round house floor in the southeast with a larger part of the courtyard all over the tested unit (fig. 5.27). The recorded remains are arranged into sherds clusters, one found on the house floor (feature 5) and the remaining four stretched along a west-east axis across the north half of the excavated probe. The exposed house floor measures 1.5 m in length and 1.0 m in width. It sealed an

Figure 5.22: Occupation II forge furnace

46

Tora-Sira-Tomo Settlement Complex

Figure 5.25: Concentration of vessels in Occupation III house floor

underneath burial that had partly disturbed a house floor from the underlying occupation III level. The exposed portion of the burial contains the lower legs with the rest of the skeleton stuck in the section. The deceased is an adult individual buried in a slightly flexed position, oriented south-north, and facing east. Occupation IV is dated to AD 1260 – 1380 (ISGS 3928) (Annex B).

globular, one with everted rim and the other with a flat base, are represented in the sample (fig. 5.30 a – b). Cooking and serving pots are represented by three specimens (fig. 5.30 c – e), as well as six variants of storage vessels (fig. 5.30 f – k), necked or with strongly constricted neck. With the notable exception of vessel 5.30 I, the pottery decoration is made of horizontal, oblique, and /or vertical grooved lines and twisted roulette impression. The coarse lithic sandstone material is divided into two grindstone fragments and one grinder (fig. 5.30 l – m). Occupation V is dated to AD 1210 – 1290 (ISGS 3931) (Annex B).

The recorded cultural remains consist exclusively of pottery, as sherds and complete or nearly complete vessels (fig. 5.28). Service vessels of the bowl category are represented by seven shape classes (fig. 5.28 a – g) with decoration confined to linear motifs around the base, on rim, and as oblique bands. Small to larger pots (fig. 5.28 h – p), probably used for cooking and storage, are all decorated with twisted roulette impression with or without horizontal and oblique grooved lines. They are globular to elongated with more or less constricted neck and everted rim, with two flat base specimens (fig. 5.28 o - p).

5.2.7 Occupation VI Occupation VI remains are found at 0.20 to 0.60 below the mound surface, overlain by a slightly disturbed topsoil deposit. The recorded material includes a relatively well preserved primary burial, a circle of large laterite blocks, and a number of sherds and large laterite blocks in the center of the circular feature (fig. 5.31). The burial at 0.60 m contains the remains of an adult individual oriented SENW, facing southwest, with legs slightly flexed, and the hands under the skull (fig. 5.31). The sediment all around the skeletal remains is more or less ashy, suggesting that the grave may have been dug under a house. If this was the case, the circular arrangement of large laterite blocks exposed above the skeleton at 0.40 to 0.20 m may have been the foundation of a small 2:00-2:50 m diameter round hut. The cultural remains found within the hut include a grindstone, - a large rectangular laterite slab with well smoothed surface -, three fragments of smallsize grindstones, two small sandstone hammer-stones, two iron arrow tips, as well as numerous sherds (fig. 5.32).The recorded sample of pottery includes six vessels’ shapes with one case of clay pot-lid. There are bowls with straight rim

5.2.6 Occupation V Occupation V remains, exposed at 0.80-0.85 m, consist of a small portion of the courtyard in the south and a relatively large segment of a round house floor measuring approximately 3.00 m in maximum radius. Two small size laterite blocks were found on the house floor which is overlain by a pile of brown-yellow silty sediment. The latter sedimentary material probably resulted from the collapse of the house walls. A small hole, 20 cm in depth and 30 cm in diameter with fire-hardened sides, probably a posthole, was uncovered in the east of the floor (fig. 5.29). The recorded cultural remains consist of numerous sherds, a number of complete or almost complete vessels, as well as a handful of coarse stone pieces (fig. 5.30). Two kinds of bowls, both

47

Archaeology of West African Mound-Clusters

Figure 5.26: Occupation III material culture

(fig. 5.32 a – b), a handled cooking-serving pot with slightly constricted neck and everted rim (fig. 5:32c), a hemispheric pot (fig. 5:32d), an elongated pot with everted red-slipped rim (fig. 5.32 e), and finally, a globular pot with everted rim (fig. 5.32 f). They are decorated predominantly with twisted roulette impression combined with horizontal and parallel grooved lines on two specimens. Occupation VI is dated to AD 1412 – 1477 (ISGS 3927) (Annex B).

sample weight by the number of sherds [6.3 kgs/283], points to a high degree of trampling. With the exception of the pioneer deposits at the bottom of the archaeological sequence, occupation floors at 1.50 m, 1.30 m, 0.85 m, and 0.40 m tend to have larger sherds samples (tab. 5.2). As is the case for the samples of complete or almost complete vessels discussed previously, most of the collected sherds are decorated with twisted roulette impression as the predominant technique (tab. 5.3). The combination of twisted roulette impression with grooved lines is second among the most popular vessels decoration with significant variation in frequency between occupations. The modal thickness of sherds ranges from 7.5-9 mm to 11.5-14 mm (tab. 5.4) with however a subtle trend toward increasing vessels wall thickness.

5.2.8 The topsoil The topsoil from the surface to 0.25/0.30 m is partly disturbed by burrowing animals. It is nonetheless relatively rich in sherds which can unfortunately not be assigned to any precise occupation (tab. 5:2). The relatively high fragmentation index of 0.02, obtained by dividing the total

48

Tora-Sira-Tomo Settlement Complex

Figure 5.29: Occupation V floor Figure 5.27: Occupation IV floor

Figure 5.28: Pottery from Occupation IV

49

Archaeology of West African Mound-Clusters Table 5.2: General distribution of potsherds from TST-3-East

Topsoil VI V IV III II I

Rim

Body

Base

Depth (cm)

Level

n

w (kg)

D

N

U

D

N

U

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-225

1 2 3-4 5 6 7 8 9 10-11 12-13 14-15

283 338 160 132 611 162 180 322 53 1 3

6.3 12.24 4.9 7.69 13.30 5.51 4.54 20.45 2.27 0.025 0.025

9 14 5 5 48 20 17 18 7 -

9 12 4 3 14 3 8 2 -

9 2 2 4 3 2 1 -

116 158 66 80 356 112 105 170 38 1 2

16 36 20 3 64 5 13 5 2 1

121 113 63 35 103 17 33 50 4 -

1 4 1 6 2 1 1 -

225

-

2,245

77.25

143

55

23

1.204

165

539

16

Occupation

Total

Key: D = Decorated; N = Non decorated; U = Unknown

Table 5.3: General distribution of decoration techniques Depth (cm)

Level

D

ND

TR

GL

C

Other

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-225

1 2 3-4 5 6 7 8 9 10-11 12-13 14-15

255 287 134 122 511 152 157 238 50 1 2

26 48 25 10 100 10 23 84 3 1

86 126 61 69 249 92 98 127 35 1 -

9 10 2 2 6 2 -

2 3 5 4 3 2 1 -

1,909

330

944

31

20

Occupation Topsoil VI V IV III II I Total

-

Composite TR/GL

Other

7 1 -

21 25 10 9 116 23 20 85 8 2

5 2 1 1 11 3 1 3 -

8

319

27

Key: D = Decorated; ND = Non Decorated; TR = Twisted Roulette impression; GL = Grooved Line; C = Comb impression.

Table 5.4: General distribution of sherds thickness (mm) Occupation Topsoil VI V IV III II I Total

Depth (cm)

Level

5-7

7.5-9

9.5-11

11.5-14

>14.5

0-20 20-40 40-60 60-80 80-100 100-120 120-140 140-160 160-180 180-200 200-225

1 2 3-4 5 6 7 8 9 10-11 12-13 14-15

13 24 10 1 31 4 7 14 1 -

53 71 31 20 125 27 39 108 8 2

71 90 34 35 205 49 57 73 10 1

75 79 47 49 157 44 56 84 17 1 -

62 75 40 27 43 26 20 34 16 -

-

-

105

484

625

609

343

50

Tora-Sira-Tomo Settlement Complex

Figure 5.30: Occupation V material culture

5.2.9 Faunal remains

in the previous floor, with two circular houses, one used for extensive storage of household’s gear. The abandonment of such a rich repertoire of cultural remains may have been triggered by sudden and traumatic events, raiding and/or illness. Occupation IV floor is comprised of a small portion of a round house and part of the courtyard with a number of in-situ broken vessels. Occupation V remains consist essentially of a large part of a round house floor with a small portion of courtyard, the former overlain by a thick deposit of collapse mud wall material. And finally, occupation VI

Direct subsistence data were strongly localized in certain deposits. All of the faunal remains were collected from occupation IV and V deposits, at depth varying from 0.85 to 1.40m. One sheep/goat (ovis/capra sp.) bone and one shell of fresh water oyster (Etheria elliptica) were found in occupation VI deposit at 0.60 m. The bulk of the lumped animal bones sample is made of 12 sheep/goat, 3 cattle, and 22 undetermined mammals bones, 25 catfish (Gymnarchus niloticus) and 24 undetermined fish bones, 8 monitor lizard (Varanus niloticus) bones, 3 undetermined bird bones, and finally, 15 undetermined bones fragments and splinters. Summary The lower half of the exposed archaeological sequence, from occupation I to III, between 1.40 and 2.25 m, is not dated directly because of the lack of adequate dating material. The upper half, from 1.40 to the surface, with occupation IV to VI deposits ranges in time from AD 1260-1380 to AD 14101480, approximately two centuries of occupation. TST-3East probe presents an interesting succession of activities that took place at this eastern end of the elongated ToraSira-Tomo 3 mound. The earliest settlers performed a rite of foundation by burying two children skulls in the bottom natural clay layer. Their dwelling features may have been set beyond the limit of the excavation probe. Occupation II at 1.80 m presents evidence of two circular house floors which were very likely part of an articulated compound. The presence of a forge furnace in the courtyard is a strong indication of a blacksmith habitation unit. Occupation III, at 1.50 m, re-iterates the spatial arrangement documented

Figure 5.31: The burial from Occupation IV

51

Archaeology of West African Mound-Clusters

Figure 5.32: Occupation VI material culture

52

Tora-Sira-Tomo Settlement Complex

Figure 5.33: Stratigraphy of TST 3-West, East section

contains the remains of a complete primary burial overlain by the foundation line of a small round hut. The topsoil caps the archaeological sequence which may have started at the end of the first millennium BC or beginning of the first millennium AD.

cause of the fire is unknown. The richness and the diversity of the uncovered remains are remarkable. The excavated 25 m2 includes 5 round huts floors with well preserved pise walls. Three round huts “plastered” floor are set along the NW-SE diagonal line. The northwestern-most floor exposed portion measures 2.5 m in maximum radius. The central one, slightly elongated in shape is 3 m long and 2.5 m wide; and the third southeastern one, an extension of the central hut, has an exposed portion of 2.5 by 1.5 m. The eastern hut which appears to be the largest of the unit was exposed on 4 m north-south and 1.5 m east-west. It was devoted to the storage of domestic gear and supplies which may have been set on wooden racks against the hut’s wall, as suggested by the presence of a post-hole (fig. 5:35). A small circular installation made of the same yellow silty clay is found in the NNE of the probe. A lateral niche on the east flank of the central hut and a sunken space between the NW, the central hut, and the circular installation was used for the storage of large, possibly water clay containers. There are two distinct large concentrations of clay vessels, one in the central hut associated with a fire place with burnt wood and the other in the courtyard (fig. 5.34 and 5.35). In all the cases, laterite blocks were used to stabilize the pile up clay vessels. A small flat base bowl capped with a lid and containing 25 unfired clay balls (fig. 5.36) was found on the floor of the central hut at the north end of the vessels concentration.

5.4 - Tora-Sira-Tomo – 3 -West Tora-Sira-Tomo-3-West is located at the west end of the mound at approximately 200 m of TST-3-East. The test unit measures 25 m2 (5 by 5 m). The cultural deposit is comparatively thinner, measuring only 1.30 m in thickness. The excavation was carried down to 1.60 m to eliminate the possibility of a hidden cultural level underneath. The exposed stratigraphic sequence is made of 8 layers (fig. 5.33) 8 – 0 -15 cm: Loose dark brown grey topsoil 7 – 15 – 30 cm: Yellow brown silty clay with small gravels 6 – 30 – 50 cm: Dark reddish laterite floor 5 – 50 – 100 cm: light brown grey ash fill 4 – 50 - 1/120 cm: Yellow brown silty clay 3 – 100 – 120 cm: Light brown-yellow silty clay with gravels 2 – 110 – 125 cm: Burnt wooden pole 1 – 125 – 160 cm: Dark brown grey bottom natural clay.

The range of shape and size recorded in the vessels sample collected from occupation I floor cover most of the functional requirements of household’s life. Service vessels of the bowl category are represented by footed and flat base specimens (fig. 5.37 a – d, and 5.40 b) with elaborate decoration involving bands of twisted roulette impression, rocker stamping, and red slip. Medium and large cooking vessels including steamers are represented by globular to elongated pots (fig. 5.37 e – j) with slightly constricted neck and straight to everted rim decorated with a combination of twisted roulette impression, horizontal/oblique grooved lines, and red slipped bands. Storage vessels are on the average larger in size with thicker walls. The represented 12 specimens are distributed into, narrow necked and elongated (fig. 5.38 c, d, e; 5.39 c); slightly globular and elongated (fig. 5.38 f, g; and 5.39 c, d) with everted rim; and finally, globular to elongated (fig. 5.39 a – b and 5.40

The cultural sequence is comprised of six successive occupations with floors exposed at 0.15, 0.30, 0.60, 0.80, 1.00, and 1.20 m. The regularity in the thickness of occupations deposits, which range from 15 to 30 cm, with four out of six deposits measuring 20 cm in thickness, is remarkable and may point to a fast build-up of the cultural deposit. 5.4.1 Occupation I Occupation I deposit was exposed from 1.00 to 1.40 m below the surface. It contains the remains of a domestic complex that was destroyed by fire with parts of the features wall preserved to a height of 20 cm. The floor itself was exposed at 1.20 m with the accumulation of cultural remains sloping down to 1.40 m (fig. 5.34 and 5.35). The

53

Archaeology of West African Mound-Clusters

a – b). Twisted roulette impression, grooved lines, red slipped bands are used for the decoration of these large size vessels, with in some cases (fig. 5.39 a, 5.40 a, and 5.41) a remarkable aesthetic effect. Liquid storage vessels are represented by five specimens of large size jugs, globular to slightly elongated in shape (fig. 5.40), decorated with comb impression, twisted roulette impression and grooved lines. Finally, one flat sandstone grindstone and a triangular shaped grinder were found on the floor of feature 33, the eastern circular hut (fig. 5.42). Occupation I is dated to AD 997-1158 (ISGS-4590) (Annex B). 5.4.2 Occupation II Occupation II floor was exposed at 1:00 m below the surface sealed by a 20 cm thick deposit of collapsed building material. Three circular hut floors were exposed along the eastern and northern sides of the probe delineating a courtyard which extends all over the remaining space (fig. 5.43). The excavated portions of the huts measure 2.5 m in diameter on the average. Feature 17, the eastern hut contains two additional installations; a 0.50 m in diameter circular fire-pit with three laterite blocks and a series of four relatively large laterite blocks which may have been used to support and stabilize large storage jars. Burnt wooden logs were also found in the courtyard, on the southwest flank of hut floor 17. The sherds recorded from occupation II floor point to a potting tradition very similar to that of the previous occupation I. Grinding equipment, a circular portable. grindstone and an almost complete spherical grinder were part of the collected cultural remains. In fact, occupation II is dated to AD 1000 – 1162 (ISGS-4585), almost the same time range as occupation I.

Figure 5.34: Occupation I floor

5.4.3 Occupation III Occupation III deposit was exposed at 0.60 to 0.80 m. The remains found at 0.80 m include a large portion of a round hut floor and part of the courtyard. The house [feature 10] appears to have been a relatively large construction measuring slightly more than 4 m in radius (fig. 5.44). Large pieces of sun dried clay were found on and next to the floor. The set next to the floor edge may have been part of the wall. The sherds sample point to a high degree of fragmentation. The recorded material culture includes a series of four bone awls made on catfish dorsal spine, one barbed iron spearhead, as well as two possible iron spear fragments (fig. 5.45). Occupation III is dated to AD 1040-1240 (ISGS-4588) (Annex B).

Figure 5.35: View of burnt features from Occupation I

5.4.4 Occupation IV Occupation IV floor was found at 0.60 m overlain by a 20 cm thick fill of collapsed building material. The uncovered remains consists of feature 8, a small portion of a round hut measuring 2.5 m and 1.5 m in width in the east; feature 9, a loose installation made of laterite blocks in the northwest; and the courtyard (fig. 5.46). The laterite blocks may have

Figure 5.36: Unfired clay balls in their container

54

Tora-Sira-Tomo Settlement Complex

Figure 5.37: Pottery from Occupation I

Figure 5.38: Pottery from Occupation I

55

Archaeology of West African Mound-Clusters

Figure 5.39: Pottery from Occupation I

Figure 5.40: Pottery from Occupation I

56

Tora-Sira-Tomo Settlement Complex

Figure 5.41: Pottery from Occupation I

Figure 5.42: Coarse lithic artifacts from Occupation I

57

Archaeology of West African Mound-Clusters

Figure 5.43: Occupation II floor

Figure 5.45: Bone and iron artifacts from Occupation III

been used as stands for a storage bin. Occupation IV is dated to AD 1220-1300 (ISGS-4587) (Annex B). 5.4.5 Occupation V Occupation V floor was exposed at 0.30 m below the surface. It is comprised of three round huts dark reddish laterite floors (fig. 5.47). Feature 3, the NW floor measures approximately 3 m in maximum exposed radius. Feature 4, the SE floor, has an exposed portion measuring 2.00 m in radius. And finally, feature 7, the SW floor under the probe’s stair, has a narrow 1.00 m long and 0.25 m wide portion. Laterite blocks, potsherds as well as a number of installations are found scattered all over the courtyard. A fire hardened surface with white ash (feature 6), probably a small outdoor fire place, is found on the east flank of the largest floor. Another partially exposed installation (feature 5), made of three relatively large laterite blocks, is found in the northeast; it may have been part of a larger storage bin foundation. An iron spearhead and two alloyed copper arm-rings were found in the deposit (fig. 5.48 and 5.49). The arm-rings were found concealed in the ground 10 cm below the floor. Occupation V is dated to AD 1190-1280 (ISGS 4586) (Annex B).

Figure 5.44: Occupation III floor

5.4.6 Occupation VI Occupation VI floor was exposed at 0.15 m, overlain by loose brown grey topsoil. The sediment contained a very

58

Tora-Sira-Tomo Settlement Complex

Figure 5.48: View of the “cached” copper arm-rings

Figure 5.46: Occupation IV floor

Figure 5.49: Metal artifacts from Occupation V

Figure 5.47: Occupation V floor

Figure 5.50: Occupation VI floor

59

Archaeology of West African Mound-Clusters

Table 5.5: TST-3-West: General distribution of sherds Occupation

Depth

Level (cm)

n

W (kg)

Rim

Body

Base

Sample composition VI V IV III II I

0-20 20-40 40-60 60-80 80-100 100-120 120-140

1 2 3 4 5 6 7

78 436 1,036 608 154 575 405

3.95 20.25 44.80 43.30 9 44.65 50.95

9 48 117 100 15 92 52

66 379 907 503 133 276 339

4 6 2 1 2 5

Total

-

-

3,292

216.9

433

2,603

20

D

ND

TR

GL

IN

RS

TR/GL

Other

69 338 908 539 145 605 382

9 98 128 69 9 52 23

61 258 738 408 122 408 246

2 -

10 11 17 11 7

3 20 115 94 16 86 40

4 62 160 110 4 175 122

13 14 3 9 -

Decoration techniques VI V IV III II I

0-20 20-40 40-60 60-80 80-100 100-120 120-140

1 2 3 4 5 6 7

Distribution of thickness classes Occupation

Depth

Level

4.5-6

6.5-8

8.5-10

10.5-12

12.5-14

>14.5

VI V IV III II I

0-20 20-40 40-60 60-80 80-100 100-120 120-140

1 2 3 4 5 6 7

2 2 1 -

1 4 22 6 3 15 6

7 68 164 61 18 69 24

8 42 159 64 15 62 28

8 77 165 125 20 235 128

54 245 406 340 98 257 219

Total

-

-

5

57

411

378

758

1,619

Key: D = Decorated; ND = Non decorated; TR = Twisted roulette; GL = Grooved line; IN = Incision; RS = Red slip.

small amount of remains, some 78 highly fragmented potsherds. The dwelling evidence consists of two round huts dark reddish laterite floors set along a NW-SE axis (fig. 5.50). Feature 1, the NW specimen, measures 3-3.5 m in exposed radius. Feature 2, the SE floor, has a small 1.8 m by 0.75 m exposed portion. A number of laterite blocks are scattered all over the excavated probe. 5.4.7 The sherds population The total sherds population amounts to 3,292 pieces weighting 216.9 kilograms (tab. 5.4). Occupation I has the highest proportion of large sherds. Twisted roulette impression, red slipping, and combined twisted roulette impression/grooved lines are the most frequent decoration techniques used all along the occupation sequence of ToraSira-Tomo-3-West. Thick sherds from the >14.5 mm class of storage vessels are predominant in all the cases (Tab.5.5).

Figure 5.51: Burnt cowpeas (Vigna unguiculata) from Occupation II

5.4.8 Faunal Remains Faunal remains were collected from occupation III, IV, and V deposits while burnt cowpeas (Vigna unguiculata) were

60

Tora-Sira-Tomo Settlement Complex

Table 5.6: TST-3-West: distribution of faunal remains Taxa

VI

V

IV

III

II

I

Total

-

66 8 47 -

79 14 40 -

141 39 57 21

-

-

286 61 144 21

-

4

11

30 7

-

-

30 22

-

-

1 -

7 9 6 1

-

-

8 9 6 1

Undetermined

-

2

-

10

-

-

12

Total

-

125

210

333

-

-

668

Mammals Ovis/Capra sp. Bos Taurus Bovidae sp Undetermined Fish Gymnarchus niloticus Synodontis niloticus Tilapia sp. Others Reptiles Varanus niloticus Shells Etheria elliptica Aspartharia sp. Limicolaria sp. *Cyprea monetaria Birds

* Non subsistence item

space. Evidence of destruction by fire is prominent in occupation I and II but their causes are difficult to decipher. Accidental fires are very likely but one cannot rule out the possibility of hostile raids, an issue that will be addressed later in the discussion of regional settlement dynamics.

found in a pot from house floor 17 in occupation II (fig. 5.51). Domestic animal bones, from sheep/goat and cattle, are largely predominant in all the samples (tab. 5.6) with the proportion of the former significantly higher. A number of wild mammals may be represented in the Bovidae sp. category. Fish from the nearby Mouhoun River, catfish (Gymnarchus niloticus, Synodontis niloticus) and Tilapia sp. are well represented in occupation III and IV. Land-snails (Limicolaria sp.), fresh water oysters (Etheria elliptica), and fresh water mussels (Aspartharia sp.) are more frequent in occupation III but their status in the subsistence system is far from clear. There is in addition, one case of cowry shell (Cyprea monetaria) in occupation III dated to AD 1037-1241 (ISGS 4588). Most of the birds’ remains are also clustered in occupation III deposit and consist of shaft bones with 4 undetermined eggshell fragments. Despite the lack of precise information and based on the characteristics of present-day local wild-life, these eggshells can probably be attributed to chicken (Gallus gallus) or guinea fowl (Numida maleagris). All in all, the uncovered subsistence remains point to a mixed farming community which took advantage of aquatic resources available from the nearby river.

In general, the occupation of Tora-Sira-Tomo-3, as seen from both the East and the West probes, is stretched at least over half a millennium, from ca. AD 1000 to AD 1500, with high probability for earlier occupations in the eastern part of the mound. The lower and undated half of TST-3-East sequence, from 1.40 10 2.25 m, is at least contemporaneous with TST-3-West whole occupation, if not earlier if judged by the presence of a blacksmith workshop that may have been contemporaneous with Tora-Sira-Tomo-1 ironsmelting site.

Summary The occupation sequence documented at Tora-Sira-Tomo-3West spans some three centuries, from ca. AD 1000 to AD 1300. All the features recorded in all six occupations point to domestic activities in part of the delineated household

Figure 5.52: Burnt roof beams from occupation IV

61

Archaeology of West African Mound-Clusters

5.5 - Tora-Sira-Tomo – 4 Tora-Sira-Tomo-4 is a small elongated mound located at approximately 250 m southeast of the main mound ToraSira-Tomo-3. It measures 23 m in length N-S and 15 m in width E-W, with an archeological deposit some 0.60 m thick. The excavation was carried down to a depth of 0.80 m to eliminate any possibility of another cultural level underneath. An initial excavation unit of 5 by 3 m was expanded to reach 12 m by 6 m, approximately 50% of the mound. The exposed stratigraphic section is made of the following layers (fig. 5.52): 2 - 00 – 0.30 m: Light brown-grey silty sand topsoil 1 - 0.30 – 0.80 m: Dark brown-grey silty clay. 5.5.1 Occupation I Two episodes of occupation have been recorded. Occupation I material was found at 0.60 m below the surface and exposed on 72 m2, 12 m long and 6 m wide (fig. 5.53 and 5.54). A circular hut floor (feature 10) measuring 3 m in diameter was found at the south end of the probe. Many medium to large size clay vessels arranged into 11 clusters, all crushed to pieces, are scattered all over the remaining portion of the excavated probe, interspersed with groups of laterite blocks. The latter were pot-stand used to support and stabilize the clay containers during their use. The delineated activity does not seem to revolve around ordinary subsistence matters. All the recorded clay containers (fig. 5.55 and 5.56), with the exception of one jug-like specimen (fig. 5.56-4), measure 30 to 48 cm in mouth and/or maximum diameter (Tab.5.7). The original height of all the specimens is unknown. They appear to have been left in upside-down position resulting in the preferential destruction of the vessels’ base. The presence of dark purple dye powder and one spindle whorl suggest this place to have been used for weaving and cloth dyeing.

Figure 5.53: Occupation I floor

The shape, size, and decoration of the recorded vessels point to three main classes of pots. Globular vessel with slightly constricted neck and everted rim (fig. 5.55: 1 – 9, and fig. 5.56: 1-3) is the most represented class. It is followed by that of large tronconical vessels with slightly everted to straight rim (fig. 5.56: 5-7); and finally, one specimen of large globular jug (fig. 5.56: 4). Twisted roulette impression associated with grooved lines and/or red slip bands is attested on all but one specimen. During its first occupation, the excavated area of ToraSira-Tomo-4 was used as a specific activity area devoted to cloth weaving and dyeing. Approximately 50% of the mound was excavated. Dwelling features and other daily

Figure 5.54: View of the weaving/cloth dyeing workshop

62

Tora-Sira-Tomo Settlement Complex

Figure 5.55: Pottery from Occupation I

Figure 5.56: Pottery from Occupation I

63

Archaeology of West African Mound-Clusters

Table 5.7: Mensurations of pottery from TST-4 Feature number

MD (cm)

MAD (cm)

TH (cm)

Th (cm)

38 24 34 30 34 32 40 38 22 42 30

48 30 48 32 40 36 47 44 24 47 36

22 10 24 11 17 13 20 12 15 17 19

1.5 1.2 1.4 1.3 1.5 1.4 1.7 1.7 1.2 2.0 1.0

Twisted roulette and red slip Twisted roulette and oblique grooved line Twisted roulette and red slip Twisted roulette, horizontal grooved line, red slip Twisted roulette, red slip inside rim Twisted roulette, horizontal and oblique groove lines Twisted roulette, diagonal grooved lines, and red slip Twisted roulette, diagonal grooved lines, indents, and red slip Twisted roulette, horizontal grooved lines Twisted roulette, horizontal grooved lines, red slip, buttons Twisted roulette

34 36 32

36 36 33

12.5 15 12

1.5 1.2 1.5

Twisted roulette and red slip Twisted roulette Twisted roulette

14

25?

7

1.1

Red slip inside and outside

Decoration

Globular vessels 10A 11B 13 13D 14A 17A 18 18A 18B 18K 19A Tronconic vessels 5 12A 16A Jug-like vessel 19B

life installations may be located in the un-tested portion of the site. Despite this possibility, Tora-Sira-Tomo-4 can be considered as a special purpose site during its first occupation.

a number of scattered laterite blocks. Three grinders, one in granite and two in sandstone, two sandstone grindstone fragments, and one amorphous sandstone fragment were found on the occupation surface. The post-hole at one of the corner of the laterite floor suggests that this feature may have had a light roof. Occupation II is dated to AD 12861445 (ISGS-4591) (Annex B) and points to a shift toward standard residential use of Tora-Sira-Tomo-4.

5.5.2 Occupation II Occupation II deposit was exposed at 0.30 m below the surface and excavated over 15 m2. The recorded cultural remains can be partitioned into three categories: 1) a series of laterite blocks delineating a L-shaped space in the north of the probe (fig. 5.57); 2) a relatively large but collapsed and broken globular storage vessel in the north and within the laterite block delimited space; and finally, 3) an elongated rectangular shaped laterite floor, with a post-hole, measuring 2.5 by 1 m in the west/southwest, associated with

A small sample of faunal remains was collected from Occupation II deposit. They consists of 1 land-snail shell (Limicolaria sp.), 3 molars of sheep/goat, and 5 undetermined fragments of mammals bones. The presence of the land-snail shell may have simply resulted from its natural death among cultural remains. Summary Occupation I is not dated directly because of the lack of adequate material. The recorded pottery presents general characteristics found in almost all the excavated sites precluding any more precise assignment to any specific time segment. It is however very likely that the site may have been used for cloth weaving and dyeing between AD 1000 and 1200. A shift to residential use occurred during the ensuing occupation II in AD 1285-1445. Grinding equipment and faunal remains point to ordinary daily life activities. 5.6 - Tora-Sira-Tomo – 5 Tora-Sira-Tomo-5 is located at approximately 100 m south of Tora-Sira-Tomo-3, the main mound of the settlement

Figure 5.57: Occupation II floor

64

Tora-Sira-Tomo Settlement Complex

Figure 5.58: Stratigraphy of TST 5

complex. It is a sub-circular site measuring 50 m in maximum diameter. The tested unit set on the highest point of the mound measures 9 m2. The exposed stratigraphic sequence accumulated on the weathered laterite crust measures 1.25 m in thickness and is comprised of three layers (fig. 5.58 a and b): 3 – 0.25 – 0.00 m: Dark brown-gray sandy silt partly reworked topsoil. 2 – 1.00 – 0.25 m: Light brown silty clay with small amount of gravels 1 – 1.25 – 1.00 m: Brown clay with small amount of gravel. 5.6.1 Occupation I Four occupation episodes have been recorded in TST-5 cultural deposit. The earliest, occupation I, is documented at depth varying from 0.80 to 1.25 m (fig. 5.59). It is essentially a burial located in the northern side of the probe. It was dug down in the laterite crust. The upper part of the

Figure 5.59: Occupation I floor

65

Archaeology of West African Mound-Clusters

body was exposed with the rest stuck in the probe’s section. The deceased seems to have been forced into a too small burial pit (fig. 5.60). The exposed burial portion measures 0.60 m in width and 1.00 m in length. The body is oriented south-north, laid on the left side and facing west. The right arm is folded at right angle. The right leg is equally flexed at right angle. The left arm and left leg are extended. The geometry of the deceased position is stunning. The grave pit

is more or less surrounded by series of large laterite blocks, four in the west, and two in the east, with one found in the pit under the deceased pelvis (fig. 5.59). There are four additional laterite blocks in a northeast-southwest linear arrangement. It is however not clear if they were part of the burial structure. A charcoal sample collected from the burial fill at 1.00 m below the surface dates the occupation to AD 1289 – 1449 (ISGS – 4357, 560+/-70 BP) (Annex B). The deceased is a 35+ years old adult female. She had calculus flecks on her teeth, carious lesions on LM1 and RM1, moderate to heavy attrition, possible ante-mortem tooth loss with alveolar absorption on LM2, as well as crowding of the mandibular dentition (Maes & Walker, Annex 1). 5.6.2 Occupation II Evidence for occupation II was found at 0.50 to 0.60 m. It consists of a portion of a circular house floor measuring 2.00 m in length north-south and 1.30 m west-east, exposed in the northwestern part of the excavated probe (fig. 5.61). Two large laterite blocks, 1.00 m apart and inserted in the perimeter of the circular floor on the southeastern side, appear to have been part of the house substructures. Finally, a series of very large sherds belonging to a large storage vessel were found on the house floor along the western side of the tested probe. Occupation II is not dated directly but a charcoal sample collected 0.20 m below provides an AD 1331 – 1673 date (ISGS – 4356, 440+/-70 BP) (Annex B). 5.6.3 Occupation III

Figure 5.60: The burial found in Occupation I deposit

Remains from occupation III were exposed at 0.40 m. They consists of few dispersed large laterite blocks, few large potsherds, and mostly a storage pit. This pit was later used for refuse disposal and filled with ash, sherds, fragmented

Figure 5.61: Occupation II surface

Figure 5.62: Occupation III surface

66

Tora-Sira-Tomo Settlement Complex

Figure 5.64: Occupation IV floor

was exposed on 1.10 m west-east, and 0.85 m north-south. Second, a series of large laterite blocks arranged in a linear northeast-southwest pattern, practically along the diagonal of the probe. And third, a relatively large storage vessel found crushed in the northwestern corner of the tested unit. Occupation IV was not dated directly but certainly belonged to the later phase of occupation of the whole settlement complex. Tora-Sira-Tomo-5 appears to have been settled for one and half to two centuries, from AD 1289/1408 to AD 1149/1649, resulting in four occupation episodes found in a 1.20 m thick cultural sequence. The tested archaeological deposit is not particularly rich in cultural remains. A sample of 719 potsherds and 53 pieces of animal bones was recorded.

Figure 5.63: Plan and section of Occupation III storage pit

5.6.5 Material Culture stones, and animal bones (fig. 5.62). The storage pit is located in the southwestern corner of the tested unit. It is an elongated oval shape feature measuring 1.50 m in its long west-east axis, 1.00 m in maximum width, and 1.00 m in depth (fig. 5.63). The pit’s section is of an inverted bell shape. Occupation III is dated to AD 1408 – 1649 (ISGS – 4350, 400+/-70 BP) (Annex B).

Occupation I sherds population amounts to 6 pieces, 1 rim, 4 body, and 1 base sherds. Twisted roulette is the predominant decoration technique (tab. 5.8). It is very likely that the sherds found in Occupation I deposit may in fact belong to the overlying occupation II level. In this perspective, the area may initially have been used for the burial of individuals from another mound prior to the effective occupation of the spot. Occupation II sherds sample is larger. Twisted roulette impression sometime in association with grooved lines is the exclusive decoration technique in use during that period (tab. 5.8). Sherds tend to be thicker, with 125 specimens belonging to the >14.5 mm thickness class. Occupation III has the largest sherds sample. The largest proportion of sherds measures 9.5 to 14 mm in thickness with twisted roulette impression and grooved lines as exclusive decoration techniques. The same general characteristics hold for occupation IV sherds sample.

5.6.4 Occupation IV Finally, evidence from occupation IV, the latest of ToraSira-Tomo-5 cultural sequence, was found at 0.30 m, at the bottom of the partially reworked topsoil. The uncovered remains consist of three distinct sets of material evidence (fig. 5.64). First, a fire-place, comprised of a slightly disturbed series of average size laterite blocks arranged in a horse-shoe shape, found on a fire-hardened surface in the southeastern corner of the excavated unit. The burnt surface

67

Archaeology of West African Mound-Clusters

Table 5.8: Distribution of TST-5 sherds Sample composition Depth (m)

Occupation

n

W (kg)

Rim

Body

D

ND

U

D

ND

U

Base

Sample Composition IV III II

0.00-0.30 0.30-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.25

I Total Decoration techniques

177 315 188 33 1 5

23.6 23.85 18.95 1.5 0.075 0.600

25 38 6 2 1

5 2 1 -

6 3 5 1 -

106 190 116 26 2

9 16 11 1 -

26 62 49 4 1

2 1 1 1

719

68.57

72

8

15

440

37

142

5

D

ND

TR

GL

C

TR-GL

163 293 176 33 5

14 22 12 1 -

98 122 86 22 3

1 8 -

1 -

25 60 31 4 -

Total

670

49

331

9

1

120

Distribution of thickness classes (mm)

5-7

7.5-9

9.5-11

11.5-14

>14.5

2 8 -

6 28 2 5 1

33 92 14 7 2

47 118 61 7 1 1

89 69 111 14 1

10

42

148

235

284

IV III II

0.00-0.30 0.30-040 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.25

I

IV III II

0.00-0.30 0.30-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.25

I Total

Key: D = Decorated; ND = Non decorated; TR = Twisted roulette; GL = Grooved line; C = Cordon; U = Unknown.

Table 5.9: Characteristics of the reconstructed vessels of TST-5 Vessel No

MD (cm)

MAD (cm)

H (cm)

Th (cm)

20 40 10

40? 50? 40?

10 18 7

1.2 1.8 0.9

Twisted roulette impression and grooved lines Twisted roulette impression and grooved lines Twisted roulette impression and red slip

10 20 20 48

12 40? 30? 48

10 10 9 22

1.1 1.2 1.2 2.1

Twisted roulette impression and grooved lines Twisted roulette impression and appliqué cordon Twisted roulette impression and grooved lines twisted roulette impression

Decoration

Occupation III 1 2 3 Occupation IV 1 2 3 4

68

Tora-Sira-Tomo Settlement Complex

Figure 5.65: Material culture from Occupation III

Vessel shapes and size were reconstructed for Occupation III and IV specimens, three from the former and four from the latter. Occupation III vessels (fig. 5.65, tab. 5.9) include a globular pot with slightly constricted neck and everted rim decorated with twisted roulette impression and grooved lines (fig. 5.65-1); a slightly elongated pot with inward oriented rim also decorated with a combination of twisted roulette impression twisted and grooved lines (fig. 5.65-2); and finally, a high-necked globular jug decorated with a red slip band and roulette impression (fig. 5.65-3). Occupation IV has four specimens distributed as follows: a hemispheric small pot with slightly constricted neck and everted rim (fig. 5.66-1); a large globular pot with constricted neck and slightly everted rim with an appliqué cordon on shoulder, decorated with twisted roulette impression (fig. 5.66-2); a globular pot with inwardoriented and thickened rim decorated with twisted roulette impression and a horizontal grooved line (fig. 5.66-3); and finally, a hole-mouth basin with flat base decorated with twisted roulette impression (fig. 5.66-4). Iron implements complement the repertoire of material culture from Tora-Sira-Tomo-5. Four specimens have been recorded, two each from occupation III and IV. A barbed spear head measuring 9.7 cm in length and a fragment of an unknown object (fig. 5.65-4 and 5) were found in occupation III deposit. A short dagger and an iron blade measuring respectively 6.2 and 6.8 cm were collected from Occupation IV layer. Approximately half of the collected animal bones are undetermined splinters (tab. 5.10). Sheep/goat bones are predominant, with a single bone, a distal portion of the

Figure 5.66: Material culture from Occupation IV

69

Archaeology of West African Mound-Clusters

Table 5.10: Faunal remains from TST-5 Taxa

IV

III

II

I

Total

(0-30)

(30-40)

(60-80)

(80-100)

-

6 5

5 1 -

-

11 1 5

-

-

7

-

7

1

-

-

-

1

2

1

-

-

3

Numida meleagris Undetermined

4

12

1 6

2

1 24

Total

7

24

20

2

53

Mammals Ovis/Capra sp. Bos Taurus Undetermined Fish Gymnarchus niloticus Reptiles Varanus niloticus Shells Etheria elliptica Birds

humerus, belonging to cattle. The remaining pieces belong to catfish, monitor lizard, freshwater oyster, and guinea fowl (tab. 5.10). Summary The occupation of Tora-Sira-Tomo-5 started with a burial which may have been dug by people inhabiting another mound. Genuine habitation started during occupation II in AD 1330-1640 and lasted up to occupation IV in AD 1410-1650. The size of the excavation probe did not allow for the exposure of more than a portion of a house floor and a small part of its associated courtyard. 5.7: Tora-Sira-Tomo – 6 Tora-Sira-Tomo-6 is located at approximately one hundred meters west of Tora-Sira-Tomo-5. It is an elongated oblongshaped mound oriented roughly north-south and measuring 60 m in length and 40 m in width. The excavated probe, measuring 9 m2, was set on the highest point of the northern half of the mound. The probe was tested down to a depth of 2.00 m and the recorded stratigraphic sequence is comprised of four main layers (fig. 5.66a and b): 4 - 0.30 – 0.00 m: soft brown sandy topsoil. 3 - 1.25 – 0.30 m: hard and compact brown-gray silty sand with a moderate amount of gravel. 2 - 1.60 – 1.25 m: yellow brown-gray silty clay 1 - 2.00 – 1.60 m: bottom yellow silty clay deposit with gravels.

Figure 5.67: Stratigraphy of TST 6

5.7.1 Occupation I The uncovered cultural sequence consists of four occupation episodes. The earliest, occupation I, was exposed at 1.50

70

Tora-Sira-Tomo Settlement Complex

Figure 5.70: Occupation III floor

Figure 5.68: Occupation I floor

Occupation II is dated to AD 1033 – 1296 (ISGS – 4360, 820+/-70 BP) (Annex B). 5.7.3 Occupation III Occupation III, at 0.50 m, is represented by the same range of remains as in occupation II deposit; a floor extended over the whole probe surface (fig. 5.69). There is however a disturbance in the southeastern corner of the excavated unit where part of the floor was destroyed. Occupation III is dated to AD 1223 – 1441 (ISGS – 4358, 630+/-90 BP). 5.7.4 Occupation IV Finally, occupation IV remains were exposed at 0.25 m at the bottom of the top deposit. They consist of few large sherds, one relatively large laterite block, found on a circular house floor extended over more than half of the probe surface (fig. 5.70). The exposed house floor measures 3.00 m in length west-east, and 2.00 m in width north-south. Unfortunately, no adequate dating material was found in occupation IV deposit. The mound seems to have been settled for two and half to three centuries, from ca AD 9001200 to ca AD 1200-1450. This length of settlement use has resulted in a 1.60 m thick cultural sequence subdivided into four occupation episodes.

Figure 5.69: Occupation II floor

m (fig. 5.67). The recorded cultural remains are constituted exclusively of pottery material distributed into five more or less large sherds clusters. Two of the clusters are found along the southern side of the probe, and three are located along the eastern wall. The area seems to have been used for refuse dumping during occupation I which is dated to AD 904 – 1222 (ISGS – 4362, 960+/-70 BP) (Annex B).

5.7.5 Material Culture The total amount of cultural remains collected from ToraSira-Tomo-6 probe is rather small, with 514 potsherds weighting some 20.55 kilograms (tab. 5.11) and a handful of animal bones. The largest quantity of sherds was found at the bottom of the archaeological sequence in what seems to have been used as a refuse dumping area. The sherds frequency varies from 97 to 9 in the other levels of the

5.7.2 Occupation II Evidence for occupation II was exposed at 1.10 m. The uncovered floor is extended all over the probe surface, with an elongated ash dump in the south (fig. 5.68). The surface was probably part of the courtyard of a habitation unit.

71

Archaeology of West African Mound-Clusters

archaeological deposit. 484 out of the 514 collected sherds are decorated with twisted roulette impression, followed by combined twisted roulette impression/grooved line. Modal sherds’ thickness varies from one occupation to the next and fluctuates for most cases between 9.5-11 and >14.5 mm classes (tab. 5.11). Three vessels shapes were reconstructed from occupation I sherds’ sample (fig. 5.72). This includes a relatively large slightly elongated and globular pot with constricted neck and slightly everted rim, decorated with twisted roulette impression and oblique grooved lines; an elongated pot with constricted neck and everted rim decorated with twisted roulette impression and red slip on rim; and finally, a hemispheric pot, probably a large jug which measures more than 50 cm in maximum diameter with an elaborate decoration. This decoration consists of vertical bands of comb stamped impression delineated by parallel grooved lines outlined by parallel red slipped bands. These vertical slipped bands are underlined by a horizontal band of short grooved lines framed with parallel lines of twisted roulette impression and flanked with parallel red slipped bands.

Figure 5.71: Occupation IV floor

All the reconstructed vessels appear to have been relatively large in size, their mouth diameter ranging from 18 to 30 cm and their maximum diameter from 30 to >50 cm. A small

Figure 5.72: Material culture from TST 6

72

Tora-Sira-Tomo Settlement Complex

Table 5.11: General distribution of potsherds from TST-6 Occupation

Depth (m)

Rim

Body

Base

Level

n

W (kg)

D

ND

U

D

ND

U

-1 2 2 2 2 2 3 3

12 41 97 9 32 16 51 29 227

0.4 1.0 3.2 0.45 1.0 0.5 1.9 1.8 10.3

1 5 4 1 3 2 4 6 34

4 1

1 2 4 1 1 1

6 23 38 6 23 14 41 19 164

2 3 7

4 8 45 2 4 2 4

1 3 2 1 1 5

514

20.55

60

5

10

334

12

69

13

D

ND

TR

GL

TR/GL

U

12 38 94 9 32 16 47 29 207

3 4 20

5 26 35 2 23 12 33 20 152

2 2 2 5

2 2 8 3 5 4 8 5 32

5 10 49 4 5 2 -

484

27

308

11

80

75

5-7

7.5-9

9.5-11

11.5-14

>14.5

5 3 1 1 2 1 6

3 12 22 1 3 3 10 7 53

3 6 15 2 9 5 20 9 64

5 9 19 3 11 6 13 6 65

1 9 38 2 8 2 6 6 29

19

114

133

137

101

Sample Composition VI surface

III II

I Total

0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40 1.40-1.60 1.60

Decoration techniques VI

III II

I Total

Surface 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40 1.40-1.60

-1 2 2 2 2 2 3 3

1.60

Thickness classes (mm) VI

III II

I Total

Surface 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 1.20-1.40 1.40-1.60 1.60

1 2 2 2 2 2 3 3

Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line; U = Unknown.

Summary

clay figurine, 3.5 cm in height and 3.6 cm in width, which seems to represent a “cat-like’’ animal (fig. 5.71-4) was also found in occupation I deposit. A fragment of a portable. sandstone grindstone was found on occupation IV living surface (fig. 5.71-5).

Tora-Sira-Tomo 6 occupation may have started as a refuse dumping area which was later built on. The excavated probe exposed small portions of what were very likely larger household units, comprised of a number of round houses and courtyard. The mound occupation lasted from AD 900-1220 to 1220-1440, or more precisely AD 1500 if one takes into account the undated occupation IV at the top of the mound stratigraphic sequence.

The sample of faunal remains amounts to 30 pieces. Sheep/ goat bones are found in occupation II and IV and cattle bones in occupation I and III, with frequencies varying from 5 to 1. Shells of land-snails and freshwater oysters as well as birds –very likely guinea fowl – are confined to occupation I deposit. It is very probably an artifact of the excavation sampling but it is nonetheless interesting to note that fish bone are totally absent from the TST-6 faunal remains.

5.8: Tora-Sira-Tomo – 7 Tora-Sira-Tomo-7 is located at one hundred meters south of Tora-Sira-Tomo 5. It is an elongated oblong-shaped mound

73

Archaeology of West African Mound-Clusters

Figure 5.73: Stratigraphie of TST 7

Figure 5.75: View of the vessels arrangement

Figure5.74: Occupation I floor

74

Tora-Sira-Tomo Settlement Complex

Figure 5.76: Occupation I burials

from 1.00 to 1.55 m. The recorded cultural remains consist of a floor with an elaborate sherds pavement found in the northwestern part of the excavated probe (fig. 5.73). The floor portion measures 1.75 m in length west-east, and 1.25 m in width north-south. The sherds pavement is smaller in extent. It measures 1 .00 m long and 0.60 m wide, found along the probe western section. It is associated to a very poorly preserved burial including a few bones and three superimposed clay vessels, two bowls and one pot (fig. 5.74 and 5.75). Another pot was found in the eastern half of the tested unit. What was very likely part of the burial facilities was recorded slightly above the level with the sherds pavement at 0.90 to 1.00 m. The uncovered remains are divided into several large laterite blocks, two pieces of stone artifacts, and four clay vessels, organized into nested semi-circular patterns measuring 2.00 m in maximum radius (fig. 5.76). It can be suggested that the described burial was a tumulus-like structure, with its perimeter outlined by laterite blocks. The deceased is a 15-17 years old sub-adult of unknown sex, with calculus flecks and carious lesion on RM1.

Figure 5.77: View of the burials

oriented north-south, measuring 50 m in length and 25 m in maximum width. The tested unit measures 12 m2, and the exposed stratigraphic sequence is 1.55 m thick, with four major sedimentary units (fig. 5.72): 4 - 0.20 – 0.00 m: Dark gray soft sandy silt topsoil. 3 - 0.75 – 0.20 m: Light gray silty clay with a moderate amount of gravel. 2 - 1.25 – 0.75 m: Yellow clay with less gravel. 1 - 1.55 – 1.25 m: Bottom yellow silty clay with gravel.

Two primary burials were uncovered at 1.55 m below occupation I floor (fig. 5.74). One of the burials was partially excavated. The lower half of the skeleton was still in the stratigraphic section after two successive extensions of the test unit. The deceased, a 35 – 50 years old middle adult female, is oriented north-south, laid on the right side, facing west, with the hands before the face. She had heavy calculus as well as moderate to heavy teeth attrition. She presented four well-healed depressed cranial fractures on the frontal, left parietal, and occipital. Interestingly enough,

5.8.1 Occupation I The cultural sequence is comprised of three occupation episodes. The earliest one that of occupation I, was exposed

75

Archaeology of West African Mound-Clusters

Table 5.12: General distribution of potsherds from TST - 7 Rim

Body

level (cm)

n

W (kg)

D

ND

U

D

ND

U

1 1 2 2 2 3 3

19 27 62 48 32 330 64

0.8 1.5 2.2 1.7 2.0 7.2 1.3

1 4 6 5 5 3 -

2 1 1

1 1 3 1 5 -

14 15 37 31 25 49 32

2 1 -

3 5 13 6 221 30

1 2 3 1 1 1

582

16.7

24

4

11

203

3

278

9

D

ND

TR

GL

TR/ GL

U

18 27 60 46 31 328 62

1 2 2 1 2 2

13 16 39 31 21 42 11

2 5

2 3 7 6 9 9 17

3 6 14 9 1 277 29

Total

572

10

161

7

53

339

Thicknessclasses(mm)

5-7

7.5-9

9.5-11

11.5-14

>14.5

4 18 1 24 24

2 6 23 8 10 19 19

8 4 12 9 15 80 12

5 8 16 10 3 168 8

4 9 7 3 3 38 1

71

87

140

218

65

Occupation

Depth

Base

SampleComposition III II

Total

0-20 20-40 40-60 60-80 80-100 100-120 120-150 155

PatternsofDecoration III II

I

III II

I Total

0-20 20-40 40-60 60-80 80-100 100-120 120-150

0-20 20-40 40-60 60-80 80-100 100-120 120-150 155

1 1 2 2 2 3 3

1 1 2 2 2 3 3

Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line; U = Unknown.

she also presented evidence for possible healed porotic hyperostosis (MAES & WALKER, Annex 1).

consists of a middle adult male (45+ years), a middle adult female (35-50 years), and a 15-17 years old sub-adult point to a “coherent” family unit. Both adults present evidence of head injuries. The unsuccessful attempts at trephination suggest that the adult male died of his second head injury.

The other burial was exposed entirely (fig. 5.74). The deceased is oriented west-east, laid on his back with arms folded on both sides, the face turned south, and the legs tightly flexed on the right with the heels at pelvis level. He is a middle adult male, 45+ years old, with an impressive statute of 198.38 cm, or 78.1 in (6’6’’) and severe health problems. He presented evidence of malocclusion (overbite) causing abnormal wear patterns, heavy calculus, abscess, carious lesion, ante-mortem tooth loss, as well as crowding of the anterior mandibular dentition. He also had evidence for multiple trephinations with no sign of healing on the frontal, left and right parietals as well as a well healed circular defect in the occipital (Maes & Walker, Annex 1). No grave goods were included in the burials.

5.8.2 Occupation II Cultural remains from occupation II were found at 0.80 m. They consist of three more or less connected portions of circular house floors, certainly part of a larger household unit (fig. 5.77). The largest house floor with a series of small laterite blocks, located in the east of the probe, measures 2.80 m in length along the north-south axis and 1.60 m in maximum width west-east. The second in exposed size found in the southwestern part of the probe, measures 2.00 m in length, and 0.90 m in maximum width. Three clay vessels as well as many laterite blocks were found scattered on the house floor. And finally, the smaller floor specimen, measuring 1.00 m in maximum exposed radius, was exposed in the northwestern corner of the probe.

During occupation I, Tora-Sira-Tomo-7, or at least the area of the tested unit, appears to have been used as a burial ground. The demographic profile of the deceased, which

76

Tora-Sira-Tomo Settlement Complex

Figure 5.78: Occupation II Floor

Figure 5.79: Burial in occupation III deposit

5.8.3 Occupation III Occupation III, found at 0.20 m almost at the bottom of the top deposit (fig. 5.78), is also represented by the burial of a 5-7 years old child. The uncovered bones are very poorly preserved. The deceased is laid on the right side, oriented east-west, facing north, with the legs slightly flexed. Fragments of sheep/goat bones were found embedded in the child skeleton but appear to have resulted from unintentional association. The location of the burial in the top 20 cm of the stratigraphic sequence points to its intrusiveness. In other words, people settled elsewhere on the mound or even on another mound may have used this uninhabited spot to bury one of their deceased child. 5.8.4 Material Culture Some 582 sherds weighting 16.7 kilograms were collected from Tora-Sira-Tomo-7 probe (tab. 5.12). Occupation I deposit, at the bottom of the sequence, contained the largest amount of potsherds and complete vessels. Twisted roulette impression is largely predominant among the used decoration techniques, with combined twisted roulette and grooved lines in second position (tab. 5.12). Complete and restorable vessels from occupation I and II display the trends in patterns of decoration suggested by the potsherds analysis which shows the predominant use of twisted roulette impression in combination with grooved lines. All vessels from occupation I were collected from a mortuary context (fig. 5.79). They may have been used as offering containers disposed off at the time of internment. They are distributed into straight rim bowls (fig. 5.79: 1-2); small hemispheric pots with slightly everted rim (fig. 5.79: 3-4); flat base globular pot with straight rim (fig. 5.79: 5); slightly elongated globular vessel with everted rim (fig.

Figure 5.80: Pottery from occupation I

5.79: 6-7); elongated pot with inward oriented rim; and finally, globular to elongated pot with constricted neck and everted rim. Most of the recorded vessels measure 30 to 10 cm in mouth diameter (Tab.5.13) with a quite simple decoration syntax: The series of grooved lines when present are always located on the shoulder/neck/rim portion of the clay container (fig. 5.80). The sample from occupation II

77

Archaeology of West African Mound-Clusters

Table 5.13: Characteristics of TST-7 vessels Vessel No

MD (cm)

MAD (cm)

H (cm)

HMD (cm)

Th (cm)

16.5 10 10 12 12 8.5 14 20 11 14

20 10 10 16 14 8.5 16 20 11 18?

16 6.0 7.0 10 10 5.8 10 16 10 10?

10 6 10 -

0.8 0.9 0.7 0.8 0.8 0.8 0.9 0.9 0.9 0.9

Twisted roulette impression and grooved lines Twisted roulette impression Twisted roulette impression Twisted roulette impression and grooved lines Twisted roulette impression Twisted roulette impression and grooved lines Twisted roulette impression and grooved lines Twisted roulette impression Twisted roulette impression Twisted roulette impression and grooved lines

22.5 12.5 12

30 12.5 18

25 6.5 12

13 6.0

1.3 1.0 0.9

Twisted roulette impression and grooved lines Twisted roulette impression Twisted roulette impression and grooved lines

Decoration

Occupation I V-6 V-7 V-8A V-8B V-9 V-10A V-10B V-13 V-12 V-11 Occupation II V-2 V-3 V-4

was collected from a house floor and includes a relatively large storage globular pot with everted rim (fig. 5.80: 1, Tab.5.13); a spherical possibly cooking pot with everted rim decorated with triangular bands of twisted roulette impression (fig. 5.80: 2); and finally, a small serving bowl with straight rim decorated with twisted roulette impression (fig. 5.80: 3). Additional stone and iron artifacts were found throughout the deposit. All the recorded stone pieces, in granite and sandstone, were used for grinding purposes (fig. 5.81). All but one of the pieces (fig. 5.81: 7) were collected from occupation II habitation unit. They are divided into one granite grinder (fig. 5.81: 1) and five fragments of portable. sandstone grindstones. The fragment from occupation III (fig. 5.82: 7) is from a granite grindstone. All the iron artifacts come from the mortuary complex but without any direct association to a specific buried individual. They were dropped in the sediments that may have been part of the mound set above the sherds pavement and the “altar” of offering vessels. They are arranged into a ring, an elongated unidentified iron piece, a coiled flattened ring, the proximal end of an axe, and an undetermined “trident” (fig. 5.83). Tora-Sira-Tomo 7 probe was not particularly rich in faunal remains. All the collected material comes from occupation II deposit and includes one proximal end of cattle (Bos taurus), one molar of sheep/goat, 1 (fish) vertebra of Tilapia sp., three complete valves of fresh water oyster, 12 undetermined long bones fragments, and four undetermined splinters. Despite its rather small size, the recorded faunal sample does point to a reliance on livestock and aquatic resources from the nearby Mouhoun River.

Figure 5.81: Pottery from occupation II

78

Tora-Sira-Tomo Settlement Complex

Figure 5.82: Coarse lithic artifacts

Summary No dating material was collected from Tora-Sira-Tomo-7 probe. Its pottery, basically decorated with twisted roulette impression combined with grooved lines, does not yet allow for a more precise chronology to be worked out. The area seems however to have started and ended as a burial locale probably at the end of the first millennium AD. The mound has witnessed one important habitation phase during occupation II.

Figure 5.83: Iron artifacts

5.9.1 Occupation I The earliest, occupation I, documented at 0.60-0.80 m, is represented by two burials associated to a series of large laterite blocks (fig. 5.85). One of the burials consists of an isolated adult skull found in the northeast of the tested probe. A series of eight large laterite blocks were found next to the cranium, 0.20 to 1.00 m west. The second burial is located at the opposite northwest corner. It was not excavated entirely. The skull and the upper body part were stuck in the probe’s section (fig. 5.86). The state of preservation of the uncovered skeletal remains was poor in general. The deceased is a 18+ years old young adult individual laid on the back, legs flexed on the left, with both arms extended along the body. A group of four large laterite blocks found on the right side of the deceased seem to be associated with the burial. Later construction activities seem to have disturbed both graves’ superstructures.

5.9: Tora-Sira-Tomo – 8 Tora-Sira-Tomo-8 is found at slightly less than one hundred meters southwest of Tora-Sira-Tomo-5. It is a small circular low mound measuring 30 m in maximum diameter. An excavation unit of 9 m2 was sunk in the central part and probed down to 1.25 m below the surface. Its exposed stratigraphic section is comprised of three layers as follows (fig. 5.84): 3 - 0.25 – 0.00 m: Brown-gray sandy silt in partially reworked topsoil. 2 - 0.80 – 0.25 m: Softer yellow brown gray silty clay with a moderate amount of gravel. 1 - 1.25 – 0.80 m: Natural bottom yellow brown silty clay with gravel.

5.9.2 Occupation II Occupation II deposit was exposed at 0.20 to 0.60 m below the surface. It includes three superimposed laterite floors found at 0.55, 0.47, and 0.36 m. Floor IIa at 0.55 m is extended almost over all the 9 m2 of the probe (fig. 5. 87). It seems to have belonged to a relatively large round house which may have measured some 6 m in diameter. Floor IIb at 0.47 m is an irregularly shaped laterite surface. It is a portion of a partly destroyed circular house floor, exposed

Considering the shallowness of the uncovered cultural deposit, the excavation of the probe was carried down to 1.25 m below the surface to eliminate the possibility of an overlooked deeper occupation. The archaeological deposit strictly speaking is 0.80 m thick, comprised of a cultural deposite made of three occupations.

79

Archaeology of West African Mound-Clusters

Figure 5.84: Stratigraphy of TST 8

Figure 5.85: Occupation I floor

Figure 5.86: Occupation IIa floor

80

Tora-Sira-Tomo Settlement Complex

on 2.60 m in the east, 1.00 m in the west and 3.00 m in the south, with a laterite block found lying on the house floor. (fig. 5.88). And finally, floor IIc at 0.36 m below the surface presents the northern half of a round house (fig. 5.89) exposed on three meters in the south, with a maximum radius of two meters. Two small symmetric installations including varying number of laterite blocks and large sherds arranged around shallow holes in the ground appear to be door-sockets. The house door was thus open to the north. 5.9.3 Occupation III Occupation III remains were uncovered at 0.10 m, and consist of a poorly preserved floor stretched along the diagonal axis of the probe, and measuring two meters in maximum width (fig. 5.90). Unfortunately, no dating material was collected from TST – 8 probe. In general the quantity of archaeological data is significantly low compared to the other probes. But house construction, or at least house floors maintenance, appears to have been particularly intensive at the tested spot, with virtually a house floor every ten centimeters of deposit.

Figure 5.88: Occupation IIc floor

5.9.4 Material Culture The sample of collected potsherds amount to 210 pieces, weighting 6.27 kilograms. 2/3 of this material comes from occupation II deposit (tab. 5.14). The handful of sherds from occupation I deposit are very likely intrusive. Twisted roulette impression is largely predominant as decoration technique, followed by far by combined twisted roulette impression and grooved lines. The paucity of archaeological remains is easily understandable in the case of Tora-Sira-Tomo-8 probe. The rooms’ floors were certainly maintained and cleaned

Figure 5.89: Occupation III floor

constantly, making the accumulation of large quantity of worn-out pieces of material culture unlikely. A few significant objects were nonetheless recorded throughout the deposit. An iron arm ring was found in occupation I extended burial with in addition two broken fragments of iron objects (fig. 5.91 a, b, and c). Three pieces of stones, two in sandstone and one in dolerite, previously used as grindstones (fig. 5.91 d, e, and f) were reused in house floor IIc as door-socket. And finally, two significant vessels were found in occupation III deposit. One is a spherical pot with

Figure 5.87: Occupation IIb floor

81

Archaeology of West African Mound-Clusters

Table 5.14: General distribution of potsherds from TST-8 Occupation

Rim

Body

Level

n

weight (kg)

D

N

U

D

N

U

0–20 20–40 40–60 60–80

1 2 2 3

48 63 95 4

1.3 1.55 3.22 0.2

2 2 5 -

1 3 -

4 2 4 -

29 29 55 4

2 1 -

10 29 25 -

1 1 -

80

-

210

6.27

9

4

10

117

3

64

2

D

ND

TR

GL

TR/GL

Depth (cm)

Base

Sample composition III II I Total

Decoration techniques III II I Total

0–20 20–40 40–60 60–80

1 2 2 3

45 62 90 4

3 1 5 -

27 28 51 4

2 -

4 9 -

80

-

201

9

110

2

13

5-7

7.5-9

9.5-11

11.5-14

>14.5

1 6 4 -

8 24 11 -

24 13 19 1

12 15 18 -

3 5 3 3

11

43

57

45

14

Thickness classes (mm) III II I Total

0–20 20–40 40–60 60–80 80

1 2 2 3

Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line

Figure 5.90: Material culture from TST 8

Figure 5.91: Stratigraphy of TST 9

82

Tora-Sira-Tomo Settlement Complex

constricted neck measuring 26 cm in maximum diameter, 26 cm in height, and 10 cm in mouth diameter (fig. 5.91 g). It is decorated on shoulder by occuli-like grooved lines and twisted roulette impression. The other is the base of a threelegged bowl (fig. 5.91 h). Such bowls have been found in an area stretched on 1000 kilometers along the Niger River from Niani the capital city of the medieval Mali Empire in the south to Gourao and the Bandiagara cliff in the north (BEDAUX 1980, FILIPOWIAK 1979). They range in date from the 6th to the 17th century and appear to have been used in both domestic and ritual contexts. Closer to the study area, “a fragment of a bowl on three feet was found at Tou, Upper Volta [Burkina Faso] .. in a hillock mainly built of stone artifacts. .. this hillock may resemble the ritual places used for rain-making by the Dogon” (BEDAUX 1980: 253). This new find and others to be described later increase the geographic range of tri-legged bowls which appear to have reached the Mouhoun Bend in Burkina Faso. The recorded specimen may have been obtained through exchange or copied by local potters.

south to 1.20 m in the north (fig. 5.93). The bottom level is in fact the bed-rock, here the laterite crust, reached at 0.50 to 0.80 m below the surface in the western half of the excavated unit. The deposit is thicker in the north end of the probe where the excavation was halted at 1.20 m. The site’s stratigraphy is constituted of three layers: 3 - 0.05 – 0.00 m: Thin reddish layer of crushed laterite capping the ring. 2 - 0.75/0.80 – 0.05 m: A brown-gray sandy clay with gravel. 1 - 1.20 – 0.75/0.80 m: A yellow brown-gray silty clay with a large amount of gravel, from the weathered laterite crust. Tora-Sira-Tomo 9 is a single occupation site which appears to have been in used at the very end of the settlement history of the Tora Sira Tomo complex around AD 1470 – 1650 (ISGS-4593). It was a special cemetery restricted to a certain category of individuals (fig. 5.92). The recorded cultural remains which includes several burials, numerous clay containers, and many laterite blocks can be organized into four more or less distinct clusters.

A handful of faunal remains, all belonging to sheep-goat, were collected from occupation II and III deposits; one molar and two long bones fragments in the former and four fragmented molars and one undetermined splinter in the latter.

5.10.1 The northern cluster The northern cluster includes two burials, a group of four clay vessels, and two more or less parallel lines of laterite blocks (fig. 5.93), stretched on 4.5 m NW-SE and 3.5 m NE-SW. The skeletal remains are very poorly preserved and crumbled into powder. Burial 21 at the NW end of the cluster is oriented SW-NE. The deceased buried on the right side, facing SE, with arms before the face, and slightly flexed legs, was more than 18 years old in age. Burial 22, 1 m east, shares the same pattern of disposal in terms of orientation as well as body positioning. He was an adult, probably male, 25-50 years old. The strong similarity in patterns of disposal of the deceased recorded in this cluster points to a strong connection between these two adults. The four clay vessels, feature 23, 24, 25, and 26, found at the eastern end of the cluster were probably part of “offerings” installations.

Summary The tested spot of Tora-Sira-Tomo-8 was first used as a burial area. House construction came later during occupation II when three successive building phases took place. The shape of occupation III floored surface is unusual. The discovery of a large but crushed pot and the base of a three legged footed bowl may point to a special use of that area during the later part of Tora-Sira-Tomo-8 occupation. Some ritual activities involving the use of legged bowls may have been performed there. 5.10: Tora-Sira-Tomo – 9 Tora-Sira-Tomo 9 is located at the center of a shallow depression, 400 m east of the iron-smelting site, and surrounded on all sides by shallow mounds. It is not strictly speaking a mound but an elliptically shaped ring oriented northwest-southeast, measuring 30 m in length in the long axis, and 15 m in width in the short one, with a sunken central area. Its perimeter is delineated by a ring of accumulated sediment, 3.00 m in width, rising at 0.50 to 0.75 m above the surrounding ground level. This ring is capped by a thin layer of crushed laterite which may have been geared toward the stabilization of the site’s ring.

5.10.2 The central cluster The central cluster measures 4 m in length and 2 m in width and includes five features associated with six laterite blocks arranged into two loosely parallel west/east lines (fig. 5.92). Four of the exposed features are burials with the remaining one, a relatively large sized clay vessel. Feature 15 at the center of the cluster is a secondary burial made of a series of selected long bones (fig. 5.94). Burial 16, 1 m east, contains the remains of an adult of unknown age and sex. The deceased was buried on the left side, oriented SW-NE, facing NW, with legs and arms flexed (fig. 5.92). Feature 18 is a secondary burial with the skull of a 9-11 years old person. Finally, feature 20 on the western flank of the cluster contains the skeletal remains of a 12-14 years old person of unknown sex. The deceased, oriented SW-

Slightly less than the west half of the site was excavated. The tested probe measures 11 m N-S in the west, 7 m N-S in the east, 10 m W-E in the north, with in addition a 3 by 1 m extension in the north (fig. 5.92). The total excavated surface amounts to 113 m2. The exposed stratigraphic sequence varies in thickness from 0.30 m in the east and

83

Archaeology of West African Mound-Clusters

Figure 5.92: Excavation of TST 9

Figure 5.93: Stratigraphy of TST 9

84

Tora-Sira-Tomo Settlement Complex

Figure 5.95: View of the southern cluster with burial 10, 11, and 12 Figure 5.94: View of the secondary burial from the central cluster

years old person of unknown sex was buried facing east, oriented south-north, lying on the right side. 5.10.4 The eastern cluster

NE and facing the southeast, was laid on the back with stretched legs and arms flexed on the left side. In general, the central cluster appears to include the burials of younger individuals, all less than 18 years old.

The eastern cluster consists of a 3 by 2.5 m reddish laterite floor with a series of clay vessels that may have been part of an “offering” installation. There is in addition one secondary burial containing a poorly preserved adult skull (fig. 5.93). The secondary burial, feature 14, is located in the south. Feature 1, 2, 3, 4, 5, and 13, set on the laterite floor are more or less evenly spaced, with nearest neighbor distances ranging from 0.5 to 1 m. The eastern cluster may have been a sheltered shrine dedicated to the performance of rituals and ceremonies.

5.10.3 The southern cluster The southern cluster also measures 4 m in length N-S and 2.5 m in width W-E. It is comprised of series of superimposed features that seem to have been related to different activities. Feature 7 is a 2 by 2 m laterite floor partly desturbed by later burials. It is probably associated with the large storage vessels 8 and 9 found on its south flank (fig. 5.92 - 94). It is however not clear how these features are related to the burial site. They may have been part of a prior habitation unit incorporated later in an expanding cemetery. The remaining five features are part of the mortuary complex and include two clay vessels, feature 6 and 17, as well as three primary burials in features 10, 11, and 12..

5.10.5 Profile of the deceased population The sample of mortuary remains from Tora-Sira-Tomo-9 provides an interesting entry into the lives of the individuals buried in the cemetery as well as their age and sex patterning. In terms of dental pathology, the degree of dental attrition was generally slight to moderate. Individual 22 for example had calculus flecks, a moderate dental attrition, and lost his lower left incisive. Individual 18, a 9-11 years old sub-adult, presents evidence of trauma with multiple peri-mortem depressed cranial fractures (sharp object) on the right parietal. While Individual 20, the other sub-adult 12-14 years of age, had “possible neoplastic condition of non-specific infection – endocranial defects on the occipital associated with an irregular-shaped free floating ossified mass” (Maes & Walker, Annex 1). Finally, individual 11, 33-38 years old adult female had her thoracic vertebras T5 to T9 completely fused, a condition that may have impaired the mobility of the upper torso, and caused serious and chronic back pain.

Burial 10 and 11 are intimately combined but are not simultaneous inhumations (fig. 5.95). Individual 10, a 35-40 years old adult female, was buried first, laid on her back, oriented northeast-southwest, with stretched legs and arms “raised” on both sides on the head (fig. 5.95). At individual 11’s death, the grave was re-opened and individual 10 femurs were removed to make room for the newly deceased. The latter, a 33 -38 years old adult female, was buried with flexed legs, oriented southeast-northwest, lying on her right side, face down, with arms collapsed on the chest. Individual 10 femurs were piled along the back of individual 11. Finally, burial 12 is oriented north-south. It was clearly a primary burial with a pit dug through the hard reddish laterite floor. The skeleton is so poorly preserved that most of the vertebras, ribs, pelvis, lower arms, and lower legs were missing. The deceased, a more than 18

The age and sex patterning is particularly interesting. Adult males (Individual 21 and 22) are found in the northern cluster. Sub-adults (Individual 18 and 20) are buried in the central cluster. And finally, adult females (Individual 10 and

85

Archaeology of West African Mound-Clusters

11) are found in the southern cluster. The manipulation of the bones and the re-arrangement of the skeletons suggest that these two females, of almost the same age, were directly related one way or the other. They may have been sisters, twins, or co-wives, re-united in the grave. In addition, secondary burials (feature 14, 15, and 18) are found in the central and eastern clusters. One can but speculate on the meaning and rationale behind the patterning disclosed from Tora-Sira-Tomo-9 cemetery. The relatively small number of burials points to the use of this cemetery by a very restricted and select social group.

extensive twisted roulette impression; 2) globular jar with constricted neck (fig. 5.97), plain or decorated with twisted roulette impression in combination with grooved lines; 3) globular pot with slightly constricted neck and everted rim (fig. 5.96b and 5.97c) decorated with twisted roulette impression, combined with grooved lines in one case; and finally, 4) a straight side pot with flat base (fig. 5.98a) decorated with two horizontal bands of twisted roulette impression; 5) a base of a three legs footed bowl belonging to the Upper Niger tradition documented by Bedaux (1980), and finally, 6) a relatively large hole-mouth basin decorated with twisted roulette impression measuring 47 cm in mouth diameter and 20 cm in height. A thick and heavy handle of a storage jar decorated with twisted roulette impression was also part of the collected cultural remains (fig. 5.98d).

5.10.6 Material culture The recorded cultural remains are almost exclusively made of clay vessels and potsherds. Potsherds are heavily concentrated in the top 0-20 cm part of the deposit. Roulette impression was by far the predominant decoration technique and sherds tend to be thicker than the average because of the predominance of large clay containers in the site assemblage (tab. 5.15). All facts supported by the analysis of complete and restorable vessels. The recorded vessels are generally large in size and distributed into six major categories. 1) Large elongated jar with slightly constricted neck and everted rim (fig. 5.96), plain or decorated with

Four stone artifacts were collected: one quartzite flake which may have been used as fire-starter, a sandstone grindstone fragment, a dolerite adze and a sandstone grinder fragment. In addition, one iron slag was found in the excavated deposit. Summary The special nature of Tora-Sira-Tomo-9 as a “special

Figure 5.96: Pottery from TST 9

86

Tora-Sira-Tomo Settlement Complex

Figure 5.97: Pottery from TST 9

purpose site” and in this case, a cemetery with restricted access for a narrow segment of the broader society, is also enhanced by its central location, surrounded by settlement mounds. Approximately half of the site was excavated. There is a preferential clustering of the deceased along age and sex lines. The elaborate re-arrangement of burial 10 and 11 points to the fact that some individuals may have had such peculiar relationships that required the rest of the community to put them to rest in the same burial. None of the individual was buried with grave-goods but the number and location of large clay containers points to the frequent practice of ritual “offerings”.

The excavated material consists essentially of large clay vessels crushed in-situ and associated with relatively large laterite blocks. The sherds clusters are arranged into 12 features concentrated in the southern half of the excavated probe (fig. 5.99). The arrangement of laterite blocks seen in the north consists of two more or less parallel lines oriented SW-NE. The southern line delineates a rectilinear shape, possible remains of a disturbed feature (fig. 5.100). Such stone installations are generally built to support storage platform to preclude termite infestations. The excavated probe thus appears to have been part of a domestic area devoted to the storage of food and other supplies. Each of the features, from 1 to 12, represents a large clay storage pot or jar, most of them decorated with twisted roulette impression (tab5.16). The set of storage vessels was simply left on the site when people decided to move. All other finds revolve around food preparation activities with a sandstone grinder, a sandstone grindstone, and another sandstone grindstone fragment.

5.11: Tora-Sira-Tomo -10 Tora-Sira-Tomo-10 is a small sub-circular mound, measuring 20 m N-S and 18 m E-W, located at some 50 m south of the special cemetery of Tora-Sira-Tomo-9. It is a shallow single occupation site which may have been inhabited by one family group. The excavation probe set at the centre of the mound measures 15 m2, 5 m N-S and 3 m E-W, with a 0. 40 m thick cultural deposit. The sediment protecting the cultural remains resulted from the collapse and leveling of house construction material which is made of yellow brown gray sandy silt.

Summary TST-10 was inhabited around AD 1270 – 1385 (Annex B). The exposed storage space was very likely part of a

87

Archaeology of West African Mound-Clusters

Figure 5.98: Pottery from TST 9

Table 5.15: General distribution of potsherds from TST-9 Occupation

Depth (cm)

Level

n

weight (kg)

Rim

Body

Base

0–20 20–40 40–60

1 2 2

352 78 25

27.09 4.3 1.1

11 7 3

277 70 22

2 -

455

32.49

21

369

2

SampleComposition I

Total

60

Decorationtechniques I

Total

0–20 20-40 40–60

1 2 2

60

Thicknessclasses(mm) I

Total

0–20 20–40 40–60 60

1 2 2

D

ND

TR

GL

TR/GL

346 72 23

6 6 2

280 52 19

-

41 8 2

441

14

351

-

51

5-7

7.5-9

9.5-11

11.5-14

>14.5

5 3 -

10 8 8

100 9 8

203 46 9

32 11 -

8

26

117

258

43

Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line

5.12: Tora-Sira-Tomo - 11 and 12

broader domestic unit that may have included a number of houses not reached by the excavation probe. The excavated cultural remains, as suggested by their crushed and spread out nature, seem to have been left exposed for a certain amount of time before being buried by the leveling effects of erosion agencies.

Tora-Sira-Tomo-11 and Tora-Sira-Tomo-12 are twin mounds. The former, subcircular in shape and measuring 20 m N-S and 19 m W-E, has a few potsherds scatters on the surface but the excavation of a 3 by 3 m probe down to 0.40 m below the surface failed to reveal any archaeological

88

Tora-Sira-Tomo Settlement Complex

Figure 5.99: The occupation surface of TST 10 Table 5.16: General distribution of potsherds from TST-10 Depth (cm)

n

Rim

Body

D

ND

TR

GL

Slip

Other

Composite

0-20 20-40

109 150

8 29

100 121

92 127

17 23

82 116

6 4

6 10

-

4 7

Total

259

37

221

219

40

198

10

16

-

11

1 53 9 44 50 3 49 2 25 5 20 20 5 16 3 206 6 200 201 5 200 4 47 3 44 41 6 28 5 9 3 6 9 9 6 7 2 5 7 7 7 186 17 169 111 71 7 9 8 132 17 115 82 50 44 10 9 25 4 21 25 23 10 12 3 9 12 8 11 10 2 8 8 2 8 12 69 69 69 64 Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line

3 1 5 2 2 -

6 11 -

2 4 1 2 26 2 4 5

Stratigraphy

Feature

deposit. The excavation was thus halted and the probe backfilled.

revealing a stratigraphic sequence with a single layer of yellow brown silty sand with two occupations.

The excavation was more successful on Tora-Sira-Tomo-12 north of Tora-Sira-Tomo-11. It is a low mound measuring 27 m in length E-W and 18 m N-S with a relatively shallow 0.70 m archaeological deposit (fig. 5.100). The tested probe, 4 m long and 3 m wide, was set at the center of the mound. The laterite crust was reached at 0.70 m below the surface

5.12.1 Occupation I Occupation I floor was found at 0.60 m below the surface. The uncovered remains consist of a small portion of a hearth in the center south of the probe, a series of scattered laterite blocks, and a small complete pot in the southwest

89

Archaeology of West African Mound-Clusters

Figure 5.100: View of TST 10 occupation floor I

(fig. 5.101). The exposed area appears to have been part of the courtyard of a domestic unit. 5.12.2 Occupation II Occupation II surface was exposed at 0.40 m below the surface (fig. 5.102). It includes a small portion of a round hut floor in the north (feature 1), the exposed part measuring 2.3 m in length and 0.80 m in width. Two poorly preserved large storage jars set in upside down position (feature 2 and 3), as well as a number of laterite blocks were found scattered all over the exposed courtyard. Both storage jars, of a very common shape and design, are globular with slightly constricted neck and everted rim, decorated with twisted roulette impression.

Figure 5.101: Occupation I at TST 12

5.12.3 A Blacksmith shop An exceptional collection of unused iron artifacts was found associated with occupation II deposit (fig. 5.103). It includes a complete elaborate kneeled arm-ring 6.8 cm in maximum diameter, a fragment of kneeled arm-ring, a small cutlass blade 13.5 cm long and 2.5 cm wide, a knife blade 11.1 cm long and 1.7 cm wide, and finally, a lozenge-shaped spearhead 22 cm long and 1.1 to 2 cm wide. All three pieces have a bent stem indicating that they were not yet put into use. Local workmen explained that the bent stem is a device used by blacksmiths to hang their craft products on a rope for sale in their workshops. The small cutlass was found in feature 3, and all the others all around between the house floor and the large clay vessels, suggesting that the latter may have been used for the storage of freshly made iron artifacts. Accordingly, the exposed portion of occupation II may have been part of a blacksmith habitation unit and/or workshop. A sample from the iron spearhead run directly at the Lawrence Livermore National Laboratory (Berkeley, California) by A. C. Cook et al (2003) dates the occupation to AD 1280-1400. This reading is supported by another radiocarbon date from a standard charcoal sample which indicates AD 1260 – 1380 (ISGS-4596) (Annex B).

Figure 5.102: Occupation II at TST 12

90

Tora-Sira-Tomo Settlement Complex

5.12.4 The sherds population

standard domestic use, with the possibility for a blacksmith habitation unit during the second occupation.

The archaeological deposit was not particularly rich in cultural remains. 146 sherds were collected along with two stone artifacts fragments and a 100 grams iron slag. One of the stone artifacts is a sandstone grinder weighing 0.5 kg, and the other, a relatively small quartzite fragment of unknown use. The largest amount of sherds, 101 out of a total of 146, was collected from occupation I deposit. Virtually all the sherds but 8 are decorated, largely with twisted roulette impression technique (tab. 5.17). The distribution of sherds thickness classes points to the predominant use of large thick vessels, emphasizing the use of the area for storage.

5.13 Tora-Sira-Tomo -13 Tora-Sira-Tomo-13, located at some 100 m east of the twin Tora-Sira-Tomo-11-12, is the easternmost mound of the whole complex. It is a sub-circular site measuring 40 m N-S and 50 m W-E with a 0.70 m thick cultural deposit. The exposed stratigraphy includes five layers as follows (fig. 5.104): 5 - 0.0 – 0.25 m: loose dark brown fill of sandy silt with gravels. 4 - 0.25 – 0.60 m: harder light brown grey silty clay with gravels. 3 - 0.50 – 0.55 m: lenticular white ash fill. 2 - 0.60 – 0.70 m: yellow brown bottom silty fill with gravels. 1 - >0.70 m: laterite bedrock.

Summary Tora-Sira-Tomo-12 was very likely settled during the first half of the second millennium AD, probably between AD 1000 and 1400. The recorded cultural remains point to a

Three successive occupations have been recorded with the latest one heavily disturbed. 5.13.1 Occupation I Occupation I floor was exposed at 0.60 below the surface. It is 7 m long and 3 to 5 m wide, with the whole probe measuring 33 m2. The uncovered cultural remains consist of a relatively large oblong-shape house floor oriented NW-SE, measuring 5 by 4.5 m (fig. 5.105), with a dense cluster of crushed storage vessels on its south flank. There are series of laterite blocks scattered all over the house floor but arranged into two loose groups. One, in the northwest, includes four relatively large blocks and the other in the south has seven smaller-sized blocks. Laterite blocks arranged in two distinct lines are part of the storage area. They may have been used to support and stabilize the storage vessels. A post-hole measuring some 30 cm in diameter was found in the east of the floor, and a fire pit exposed in the south. The latter has a central installation made of large potsherds that may have been used to support cooking pots. A charcoal sample collected from the fire-pit dates occupation I to AD 1160-1280 (ISGS-4597) (Annex B). 5.13.2 Occupation II Occupation II deposit was found at 0.20 – 0.40 m below the surface. The probe measures 4 by 4 m and the floor was exposed at 0.40 m. A large portion of a circular hut floor measuring 4 by 3.5 m with a series of laterite blocks was uncovered (fig. 5.106). The laterite blocks in an L-shape arrangement appear to have been part of a large clay vessel stand. Occupation II is dated to AD 1160-1280 (ISGS-4594) (Annex B).

Figure 5.103: Material culture of TST 12

91

Archaeology of West African Mound-Clusters

Table 5.17: General distribution of potsherds from TST-10 Sample Depth (cm)

Decoration

Thickness class (mm)

n

Rim

Body

D

ND

TR

GL

TR/GL

8.5-10

10.5-13

>13

13 32

5

13 27

9 30

4 2

9 22

4

4

8

2

13 22

40-60

101

-

101

99

2

98

-

1

-

6

95

Total

146

5

141

138

8

129

4

5

8

8

130

OccupationII 0-20 20-40 OccupationI

Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line

Figure 5.104: Stratigraphy of TST 13

5.13.3 Occupation III

A handful of faunal remains was collected: a right tibia from a rooster (Gallus gallus) and one cowry shell (Cyprea monetaria) in occupation I; a shell of freshwater oyster (Etheria elliptica) and a sheep/goat (Ovis/Capra sp. ) phalanx in occupation II; and finally, 4 undetermined long bones splinters and a sheep/goat molar in occupation III.

Occupation III deposit was unfortunately extensively disturbed; no feature of interest has been recorded in the topmost 25 cm of the cultural deposit. However, a charcoal sample collected at 0.20 m below the surface dates the occupation to AD 1300 – 1425 (Annex B).

Summary

5.13.4 Material culture

Tora-Sira-Tomo-13 was settled during the first half of the second millennium AD. The excavated material points to a standard domestic context, with house floors and part of the courtyard. The presence of a cowry shell (Cyprea monetaria) is worth mentioning, indicating that this “currency” percolated down to the Mouhoum Bend at the very beginning of the second millennium AD.

The sherds sample collected from Tora-Sira-Tomo-13 cultural deposit is comprised of 288 pieces more or less evenly distributed among the occupations (tab. 5.18) with frequency ranging from 89 to 104. Most of these sherds are decorated, predominantly with twisted roulette impression, with the number of red-slipped sherds decreasing from occupation I to III. Data on sherds thickness classes, with an overall predominance of the >13 mm class, pointing to the frequent use of thick walled vessels, support the general inference derived from the analysis of dwelling features. Two of the storage vessels from occupation I were reconstructed (fig. 5.107a). One is a straight-side globular and elongated jar and the other a slightly constricted neck and straight rim one, both decorated with twisted roulette impression. A series of iron artifacts was also recorded in occupation II and III; a bent stem and a hook from occupation II, an unknown square section artifact from occupation III (fig. 5.107b), and finally, a sandstone grinder.

5.14: Tora-Sira-Tomo-14 Tora-Sira-Tomo-14 is an elongated mound located at approximately 100 m northwest of Tora-Sira-Tomo-13. It is oriented W-E and measures 77 m in length and 20 m in width. A 3 by 4 m excavation unit was set in the central part on the site and revealed a 0.60 m cultural deposit with two layers (fig. 5.108): 3 - 0.0 – 0.20 m: dark brown-grey topsoil; loose silt with gravel

92

Tora-Sira-Tomo Settlement Complex

Figure 5.105: Occupation I floor

Figure 5.106: Occupation II floor

93

Archaeology of West African Mound-Clusters

Figure 5.107: Material culture of TST 13

Table 5.18: General distribution of potsherds from TST-13 Sample Decoration Depth n (cm) Rim Body D ND TR Slip TR/GL Occupation III 0-20 104 8 96 82 22 64 1 17 Occupation II 20-40 95 17 78 73 22 41 7 24 Occupation I 40-60 89 8 81 81 8 60 18 18 Total 288 33 255 236 52 45 26 59 Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line

2 - 0.20 – 0.60 m: yellow brown-grey silty clay with small amount of gravel 1 - >0.60 m: laterite crust.

Thickness class (mm) 6.5-8 8.5-10 10.5-13

>13

4

12

14

74

1

5

11

77

2 7

18 35

29 54

40 191

at 0.60 m dates the mound occupation to AD 1280-1390 (ISGS-4599). And the other, collected higher in the deposit at 0.30 m, points to AD 1040-1250 (ISGS-4598). The latter date comes from a less secured context. The charcoal tested may have been removed from its original context and redeposited as part of the used building material for example.

Tora-Sira-Tomo-14 is a single-occupation mound with dwelling features recorded at 0.60 m below the surface. The uncovered remains consist of a round hut floor measuring 3.5 m in maximum exposed radius with a number of scattered laterite blocks (fig. 5.109), and a small portion of the courtyard in the southeast. A deeper probe set at the NW corner of the excavated unit was tested down to 1.00 m below the surface to eliminate the possibility of an overlooked occupation level. Two charcoal samples were collected throughout the deposit. One from the house floor

5.14.1 Material culture The density of cultural remains was particularly low in the tested unit. The total sample of potsherds amounted to 77 pieces distributed in decreasing frequency from the house floor to the topsoil (tab. 5.19). Most of the recorded sherds are decorated with twisted roulette impression. Sherds’

94

Tora-Sira-Tomo Settlement Complex

Summary

thickness ranges from 8.5-10 mm to >13 mm, peaking in the latter class. Coarse stone tools, all made of sandstone and used for heavy duty hammering and grinding, are well represented (fig. 5.110 and 5.111). They are distributed into two specimens of complete or broken hammer-stone (fig. 1.110a and b), one elongated grinder (fig. 5.110c), and three grindstones, one complete (fig. 5.110d) and two broken (fig. 5.111a and b). Finally, an iron hook, 8 cm in length, 2.8 cm in maximum width, and 0.4 cm in thickness (fig. 5.111c) completes the repertoire of cultural remains collected from Tora-Sira-Tomo-14 excavation. The recorded faunal remains consist of sheep/goat mandible fragment and molar.

Tora-Sira-Tomo-14 was a residential mound which may have been inhabited by one or two extended families in the 13-14th century AD. The exposed house floor contained very few cultural remains. It may have been used as a sleeping space carefully maintained and frequently cleaned. 5.15: Tora-Sira-Tomo -15 Tora-Sira-Tomo-15 is a small circular mound located at approximately 50 m north of Tora-Sira-Tomo-14. It measures 20 m in diameter with a 0.40 m thick cultural deposit. The excavation probe set at the center of the mound measures 42 m2, 6 m east-west and 7 m north-south. The recorded cultural deposit is made of two layers (fig. 5.112): 3 - 0 – 0.20 m: dark brown grey loose silty sand with gravels 2 - 0.20 – 0.40 m: yellow brown silty clay with small amount of gravel 1 - >0.40 m: laterite crust.

Figure 5.108: Stratigraphy of TST 14

Figure 5.109: TST 14 occupation floor

95

Archaeology of West African Mound-Clusters

Figure 5.110: Coarse stone artifacts

Table 5.19: General distribution of sherds from TST-14 Depth (cm)

Sample n

Rim

Body

Decoration D

ND

TR

Thicknessclass(mm) GL

0-20 20 1 19 19 1 19 1 20-40 23 3 20 16 8 12 1 40-60 34 4 30 25 9 18 1 Total 77 8 69 60 18 49 3 Key: D = Decorated; ND = Non-decorated; TR = Twisted Roulette; GL = Grooved line

TR/GL

8.5-10

10.5-13

>13

3 6 9

3 2 4 9

7 7 5 19

10 14 25 49

measuring more than 6 m in diameter (fig. 5.113), with a number of vessels crushed in-situ and scattered laterite blocks. Two post-holes measuring 0.30 m in diameter were found in the northwestern portion of the floor. Large storage jars, feature 4, 5, 6, and 9 are crushed in the northeast quadrant of the house. Two smaller pots (feature 8 and 15) were recorded in the west. Feature 4 contains two vessels, 9, and 15, all decorated with twisted roulette impression and red slip.

Two occupations have been recorded at Tora-Sira-Tomo-15. Occupation I is documented at 0.40 m below the surface and occupation II at 0.20 m. 5.15.1 Occupation I Occupation I is made of two distinct and successive construction episodes. The first construction episode is indicated by feature 2, a large laterite round house floor

96

Tora-Sira-Tomo Settlement Complex

clay vessels. Occupation II is dated to AD 1280-1390 (ISGS-4600) (Annex B). 5.15.3 Material culture The population of sherds is important and divided into two groups: the sample collected throughout the deposit and that made of the crushed vessels. Occupation I deposit has a higher amount of sherds, 614 compared to 201 from occupation II (tab. 5.20). Twisted roulette impression is largely predominant as decoration technique but red-slipping is more common in occupation I sample. The vessels recorded in both occupations belong to four morphological categories: 1) large open plain bowl with red slip (fig. 5.115a); 2) hemispheric pot with slightly everted rim decorated with twisted roulette impression (fig. 5.115b); globular pot with constricted neck and everted rim with twisted roulette impression (fig. 5.115c); and finally, 4) large globular and elongated storage jar with or without red slip, decorated with twisted roulette impression, combined with grooved lines in a number of cases (fig. 5.115d). Vessels a to c belong to feature 3 in occupation I and vessel d comes from feature 1 in occupation II. The stone artifacts, all in sandstone, are distributed into grinder (fig. 5.116a), grindstone or grindstone fragments (fig. 5.116b, c, and d, fig. 5.117c and d), hammer-stone (fig. 5.120a), and an undetermined piece (fig. 5.117b). All but one (fig. 1.117d) belong to occupation II. The sample of iron artifacts is made of 9 specimens, four (fig. 5.118a, b, c and d) from occupation II and five (fig. 5.118e, f, g, h and i) from occupation I. All occupation II iron artifacts belong to the category of personal adornment items with a possible belt buckle and three rings. Occupation I specimens on the other hand are all arrow-heads.

Figure 5.111: Coarse stone and iron artifacts

Summary Tora-Sira-Tomo-15 mound site was settled between the 12th and the 14th century AD. The sampled area was probably part of a larger one to two households’ domestic unit. The presence of a hearth in feature 3 in occupation I points to its use as a kitchen hut. Feature 2, the large round house was certainly use for multiple purposes, from storage, to reception, and sleeping.

Figure 5.112: Stratigraphy of TST 15

The second construction episode is documented in the southeast by feature 3 (fig. 5.114). It is also a laterite house floor, oblong in shape, and measuring 4 m E-W and 3.75 m N-S. It is associated with a dump of white ash (feature 7) and five clay vessels in two distinct clusters. Feature 12, the eastern cluster is comprised of three cooking and serving vessels. The western cluster includes feature 13 and 14, both smaller storage jars decorated with twisted roulette impression, with red slip for the latter specimen. Occupation I is dated to AD 1160-1280 (ISGS-4601) (Annex B).

5.16: Tora-Sira-Tomo-16 Tora-Sira-Tomo-16 in the northeast flank of the settlement complex is located at approximately 100 m west of ToraSira-Tomo 15. It is a relatively narrow and elongated mound, oriented NW-SE, measuring 30 m in length and 10 m in maximum width. An archaeological probe measuring 4 by 3 m was sunk at the center of the mound. It revealed a 1 m thick stratigraphic sequence made of three depositional units accumulated on the laterite crust (fig. 5.119), and arranged as follows from bottom to top:

5.15.2 Occupation II Occupation II is represented by feature 1 in the northwest (fig. 5.114). It is a round hut laterite floor measuring approximately 3.5 m in exposed diameter with four crushed storage jars arranged along a SW-NE axis. A well preserved portable sandstone grindstone was found amidst the crushed

3 - 0.0 – 0.20 m: dark brown grey soft silt with gravels

97

Archaeology of West African Mound-Clusters

Figure 5.113: Occupation I floor

width. A clay vessel and two laterite blocks were found on the floor. The clay container (fig. 5.122c) is a small flat base basin measuring 40 cm in mouth diameter and 25 cm in height, decorated with twisted roulette impression. Occupation I is dated to AD 1280-1390 (ISGS-4604) (Annex B).

2 - 0.20 – 0.40/0.80 m: light brown grey silty clay with gravels 1 - 0.80 – 1.00 m: yellow brown clay with gravels. The recorded cultural deposit includes evidence for two occupations, with house floors or habitation surfaces, exposed at 0.30 and 0.60 m below the surface.

5.16.2 Occupation II

5.16.1 Occupation I

Occupation II surface at 0.30 m revealed a series of laterite blocks scattered all over the probe associated with two large sherds clusters (fig. 5.121). The laterite blocks arrangement was disturbed but it was still possible to perceive three parallel NE-SW lines. In the Mouhoun Bend, such laterite blocks arrangements can still be seen

Occupation I remains consists of a small portion of a round house laterite-plastered floor and a portion of courtyard. The floor, confined to the north/northeast of the probe (fig. 5.120), measures 3 m in length and 1.5 m in maximum

98

Tora-Sira-Tomo Settlement Complex

Figure 5.114: Occupation II floor

Table 5.20: Tora-Sira-Tomo mound cluster at phase I (650 BC – AD 800) Mound

Size (m2)

Probe (m2)

Deposit (m)

Number of occupations

Nature

TST-1 TST-2 TST-3

255 1,250 ???

65 2 10.50

0.30-1.50 0.30-2.00 0.45

2 ? 1

Iron-smeltingsite Quarry Forge

99

Archaeology of West African Mound-Clusters

today and are generally built to support granaries. This installation was very likely built for the same purpose. A charcoal sample collected in the loose silty topsoil at 0.20 m below the surface, and slightly above occupation II surface provides a reading of AD 1000 – 1160 (ISGS-4603). The dated sample probably resulted from remobilized sediments dug from older deposits and used for the construction of granaries. It consequently does not provide an accurate date for occupation II deposit. 5.16.3 Material culture Tora-Sira-Tomo 16 deposit was not particularly rich in cultural remains. Occupation I deposit was surprisingly devoid of sherds. All the recorded 202 sherds were collected from occupation II sediment. 155 of the sherds are decorated, predominantly with twisted roulette impression (123 specimens) with frequent slipping (16 cases). Thick sherds, belonging to the 11.5-13 and >13 mm thickness classes, are largely predominant. The faunal material consists of one sheep/goat first phalanx and 7 undetermined long bone splinters. The remaining portion of the material culture repertoire is made of iron and stone artifacts, all collected from Occupation II deposit. The iron artifacts consist of a double-hooked anchor-shaped piece (fig. 5.122a) and a flat narrow metal sheet which may have been a finger-ring (fig. 5.122b). The stone artifacts, both in sandstone, belong to the grinding equipment. One (fig. 5.122d) is a fragment of a broken grindstone, and the other (fig. 5.122e) an elongated grinder. Summary Tora-Sira-Tomo-16 was very likely a single household mound settled in the 13th -14th century AD. All the excavated remains point to simple and standard daily-life facilities, a house floor and a portion of courtyard in occupation I and the supporting rocks of a granary in occupation II. Figure 5.115: Pottery from TST 15

Figure 5.116: Coarse stone artifacts

100

Tora-Sira-Tomo Settlement Complex

5.17: Tora-Sira-Tomo – 17 Tora-Sira-Tomo-17 is located at approximately 50 m east of the main TST-3 mound. It is a slightly elongated mound, oriented SE-NW, measuring 26 m in length and 20 m in width. A 12 m2, 4 by 3 m excavation probe was sunk at the center of the mound. It revealed a 0.60 m thick stratigraphic sequence accumulated on the laterite crust and comprised of three layers (fig. 5.123): 4 - 0.0 – 0.5/0.10 m: Soft dark brown grey silty sand with laterite gravels 3 - 0.5/0.10 – 0.35 m: Brown red grey silty clay with lateruite gravels 2 - 0.35 – 0.60 m: Light yellow grey clay with laterite gravels. 1 - >0.60 m: Laterite crust. The probe was excavated down to the “bedrock” to eliminate the possibility of another occupation underneath. The uncovered cultural remains consists of two superimposed house floors, a portion of courtyard, a number of clustered and scattered laterite blocks (fig. 5.124), and potsherds. The house floors, found in the eastern half of the excavated unit and made of crushed laterite, are both circular in shape. The earlier and larger specimen, exposed at 0.35 m below the surface, measures approximately 3 m in radius. The later and smaller floor at 0.30 m confined at the northeastern

Figure 5.117: Iron artifacts

Figure 5.118: Additional iron artifacts

Figure 5.119: Stratigraphy of TST 16

101

Archaeology of West African Mound-Clusters

Figure 5.120: Occupation I floor

Figure 5.121: Occupation II floor

102

Tora-Sira-Tomo Settlement Complex

further clarification. It is however very likely that these constructions were part of another kind of “special purpose” site, in this case a karité oil production workshop. 5.17.1 Material culture The probe was not particularly rich in archaeological remains but potsherds, few pieces of coarse lithic and some faunal material were collected. Most of the sherds, 104 out of 133, were concentrated in the occupation level at 0.200.40 m below the surface. 107 of these sherds are decorated, predominantly with twisted roulette impression in 83 cases. The sherds tend to belong to thick walled vessels as suggested by the fact that 91 out of 133 are thicker the 13 mm. Thinner service vessels sherds are absent from the sample. The processing and storage of karité oil may have required large thick-walled vessels. The handful of rock artifacts backs the idea suggested above. The 3 recorded coarse lithics are all broken sandstone pieces which may have been used as nutcrackers. They measure 7.6 to 4.4 cm in length and weigh 500 to 200 grams. The faunal sample is made of 11 pieces: 1 horn-core and 1 pelvis fragment of Bos taurus (cattle); 3 ribs and 3 proximal metapodials of Ovis/Capra sp. (sheep/goat); and finally 3 shell fragments of Etheria elliptica (fresh water oyster). Summary Tora-Sira-Tomo-17 founded and settled in AD 1280-1380, was very likely another special-purpose site devoted to the processing of karité nuts for the production of oil. The task may have been carried out by a specific household which used the rest of the mound for their other dwelling facilities.

Figure 5.122: Material culture of TST 16

corner of the probe measures 1.20 -1.75 m in radius. A pile of burnt kernels of karité nuts (Butyrospermum parkii) was found at 0.40 m in the courtyard portion of the excavation. It was however not associated to any installation. Many other smaller kernel pieces were found scattered all over the courtyard of the test unit. The area may have been used for different activities including the roasting of karate nuts prior to its occupation which is dated to 1280 – 1380 AD (ISGS-4605) (Annex B). The dwelling evidence is non-diagnostic. The small size of the probe precludes any

5.18: The dynamics of Tora-Sira-Tomo settlement complex The pattern of growth of the inhabited space within a single mound cannot be assessed with the field methodology adopted in this case, articulated most of the time on one

Figure 5.123: Stratigraphy of TST 17

103

Archaeology of West African Mound-Clusters

Figure 5.124: Occupation I floor

Table 5.21: Tora-Sira-Tomo mound cluster in phase II (AD 800 – 1100) Mound

Size (m2)

Probe (m2)

Deposit (m)

Number of occupations

TST-1 TST-2 TST-3 TST-4 TST-6 TST-12 TST-13 TST-15

255 1,250 37,200 345 1,500 486 1,963 314

65 2 35 72 9 12 33 42

0.30-1.50 0.30-2.00 0.40-0.85 0.80 2.00 0.70 0.70 0.40

2 ? 2 2 2 2 3 2

Nature Iron-smelting site Quarry Residential Weaving/Cloth dyeing workshop Residential Blacksmith workshop Residential Residential

documented in three mounds, at TST-1, TST-2, and TST-3-East (tab. 5.20), started in the middle of the first millennium BC. The earliest traces of occupation are found at TST-1, an iron-smelting site dated to 650-395 BC. The quarry site from TST-2 which provided raw material for house construction and iron production was probably kept in use during the whole span of the settlement complex occupation. And finally, the blacksmith workshop exposed at the bottom of TST-3-East probe may have been in use during the same period. That level is unfortunately not dated directly and the handful of associated potsherds is un-diagnostic.

test unit per mound. A general time line for the formation of the settlement complex can nonetheless be outlined. Tora-Sira-Tomo settlement complex is spread over an area measuring 900 m west-east and 500 m north-south. The settlement as whole thus measures 45 hectares in total extent. It includes standard residential mounds and special purpose sites (tab. 5.20 - 5.23). The development of ToraSira Tomo settlement complex took place in four successive phases, from the middle of the first millennium BC to the middle of the second millennium AD. 5.18.1 Phase I (650 BC – AD 800)

The initial core of the settlement was thus situated along the northwestern flank of the settlement complex arranged

The initial phase of settlement at Tora-Sira-Tomo,

104

Tora-Sira-Tomo Settlement Complex

Table 5.22: Tora-Sira-Tomo mound cluster in phase III (AD 1100 – 1400) Mound

Size (m2)

Probe (m2)

Deposit (m)

Number of occupations

TST-1 TST-2 TST-3 TST-5 TST-4 TST-6 TST-7 TST-8 TST-10 TST-12 TST-13 TST-14 TST-15 TST-16 TST-17

255 1,250 37,200 1,963 345 1,500 625 706 360 486 1,963 1,400 314 300 520

65 2 35 9 72 9 12 9 15 12 33 12 42 12 12

0.30-1.50 0.30-2.00 0.40-0.85 1.25 0.80 2.00 1.55 1.25 0.40 0.70 0.70 0.60 0.40 0.60 0.60

2 ? 2 2 2 2 2 2 1 2 3 1 2 2 1

Nature Ritual site Quarry site Residential site Burial and Residential Cloth dyeing workshop Residential site Burial and Residential Burial and Residential Residential site Blacksmith workshop Residential site Residential site Residential site Residential site Oil production workshop

Table 5.23: Tora-Sira-Tomo mound cluster in phase IV (AD 1400 – 1650) Mound

Size (m2)

TST-1 TST-2 TST-3 TST-4 TST-5 TST-6 TST-7 TST-8 TST-9

255 1,250 37,200 345 1,963 1,500 625 706 225

Probe (m2) 65 2 35 72 9 9 12 9 113

Deposit (m)

Number of occupations

0.30-1.50 0.30-2.00 0.40-0.85 0.80 1.25 2.00 1.55 1.25 0.30–1.20

2 ? 6 2 2 2 2 2 1

Nature Ritual site Quarry site Residential site Cloth dyeing workshop Residential Residential site Burial and Residential Burial and Residential Cemetery

along a SW-NE axis. TST-2 the quarry site was located at the southwestern end, TST-1 the iron-smelting site in the middle, and finally, TST-3-East, the blacksmith’s workshop at some 300 m NE at the eastern end of TST-3 residential mound (fig. 5.125). The extent of the inhabited space around the TST-3-East blacksmith’s workshop is not known. The production of iron artifacts appears clearly as the direct common link between these three Phase I sites. The initial settlement phase may thus have been motivated by the exploitation of the local high grade iron ore.

mounds, at TST-3 and TST-6, and two special purpose sites, at TST-1 (iron-smelting and ritual) and TST-2 (quarry). The eastern sub-cluster forms a triangular arrangement of nearly equidistant mounds with one at the near center. TST-13 and TST-15 are residential sites while TST-4 and TST-12 are special purpose sites. TST-4 was used as weaving and cloth dyeing workshop and TST-12 was possibly a blacksmith workshop. In general, residential mounds tend to be larger, 37,200 to 314 m2 in surface extent, with thicker cultural deposits, 2.25 to 0.40 m.

5.18.2 Phase II (AD 800 – 1100)

5.18.3 Phase III (AD 1100 – 1400)

Phase II of the cycle of settlement growth witnessed the foundation of five new mounds between AD 800 and 1100 (tab. 5.21), reaching a total of 8 mounds. The resulting mounds arrangement was made of two sub-clusters of four sites each: TST-1, TST-2, TST-3, and TST-6 in the west, and TST-4, TST-12, TST-13, and TST-15 in the east (fig. 5.125). It is during this phase that TST-3 became the main and largest residential mound of the settlement complex. The western sub-cluster thus included two residential

The Tora-Sira-Tomo settlement complex reached its maximum extent during phase III, when seven new mounds sites were founded between AD 1100 and 1400. The number of settled mounds amounted to 15 (tab. 5.22, Fig. 5.126). These new units were all located within the framed space delineated during phase I and II. The general impression is that of evenly distributed residential mounds filling the space between the western, and the eastern subclusters. The new additions to the settlement complex are

105

Archaeology of West African Mound-Clusters

Figure 5.125: Tora-Sira-Tomo in phase I and II

set in two kinds of arrangements. TST-5, TST-7, TST-8, and TST-10 present a rectilinear arrangement of almost equidistant mounds set at approximately 100 m from one to the next. They measure 1,963 to 360 m2 in surface extent with 1 to 4 occupation levels in 0.40 to 1.25 m thick cultural deposits (Tab.5.22). TST-5, TST-7, and TST-8 are all residential mounds which developed out of what appears to have been a cemetery in the south flank of TST-3. Each of these mounds sits on burials which were dug deep in the laterite crust. TST-14, TST-16, and TST-17 are set in a linear east-west arrangement along the northeastern flank of the complex, at 100 m from one to the next. Their size ranges from 1400 to 300 m2, with 1 to 2 occupations levels in 0.60 m thick cultural deposit (Tab.5.22). Two of the mounds, TST-14 and TST-16, were residential and the last one, TST-17 was a karité oil production workshop.

size varying from 1,500 to 37,200 m2 (tab. 5.22). In fact, multi-occupation residential mounds, TST-3, TST-5, TST-6, TST-7, and TST-8 were located on the higher portion of the landscape. Small one to two occupation mounds were founded in the east flank of the complex during phase III. They measure 225 to 1,400 m2, with cultural deposit thickness varying from 0.30 to 1.20 m. These small mounds were one to two household sites but some, like the karité oil production mound (TST-17), were special purpose localities. 5.18.4 Phase IV (AD 1400 -1650) Phase IV witnessed important changes in the size and components of the Tora-Sira-Tomo settlement complex. All the one to two occupation mounds located in the eastern part of the complex were abandoned (tab. 5.23, Fig. 5.127 - 128). The settlement complex shrunk and consisted then of 9 mounds with the single case of TST-9 as the only newly founded site. The latter mound was used as a “restricted access” cemetery located in the eastern periphery of the complex. Six of the mounds, including TST-3, TST-4,

In AD 1400, Tora-Sira-Tomo settlement complex was made of four special purpose sites (fig. 5.126 - 127): TST-1 as a possible ritual site, TST-2 as quarry site, TST-12 as a blacksmith workshop, and finally, TST-17 as a karité oil production site. The remaining 11 mounds were all residential. They were larger in general, their

106

Tora-Sira-Tomo Settlement Complex

Figure 5.126: Tora-Sira-Tomo in phase III

Figure 5.127: Tora-Sira-Tomo in phase IV

TST-5, TST-6, TST-7, and TST-8, were residential. TST-1 may have still been used for ritual purpose, and TST-2 for quarrying the laterite. TST-3 was the largest main inhabited mound located in the north of the complex. The remaining components were then concentrated in the south. With the exception of TST-2 the quarry site, all the remaining mounds were set in more or less patterned arrangements; TST-1, TST-6, TST-5, and TST-9 were almost equidistant from one to the next along a west-east axis. TST-5, TST-

9, TST-4, and TST-7 were set in a rectilinear pattern of equidistant mounds. With the notable exception of TST-8, the distance between neighboring mounds oscillated around 100 m. It is this built-in regularity which gives the overall impression of evenly distributed mounds when one looks at the map of Tora-Sira-Tomo settlement complex. In summary and with all the required caution, Tora-SiraTomo can be said to have been settled in the middle of

107

Archaeology of West African Mound-Clusters

Figure 5.128: Tora-Sira-Tomo mounds size in phase IV

the first millennium BC by groups with strong focus on iron working. Their settlement was made of three distinct units, a quarry, a smelting site, and a forge. This small settlement of iron producers which lasted for more than a millennium was probably connected to the other hamlets and villages located in this part of Burkina Faso and along the Mouhoun River valley (Holl and Kote 2000, Kiethega 2006, Coulibaly 2006).

workshop was founded. And new residential mounds were founded in the south flank of TST-3, the largest and main residential mound, in an area formerly used for burial. ToraSira-Tomo settlement complex shrunk considerably in the middle of the second millennium AD (AD 1400-1650), in terms of size, number of components, and diversity (fig. 5.128). All the eastern flank of the complex was abandoned. The number of mounds dropped from 15 to 8. Craft activities were very likely still carried out in the residential mounds themselves. The ritual and quarry sites were still in use and a new “restricted access” cemetery was founded in the eastern periphery of the cluster. Five out of the 8 recorded mounds were residential.

The pace of growth and diversification accelerates relatively at the end of the first millennium AD. In AD 800-1100, the cluster was made of 8 mounds, four residential and four special purpose sites. The quarry and ritual sites were for the whole community use. Specialized craft activities took place in two mounds, one used by cloth dyers/weavers and the other by blacksmiths/iron workers. There are two equally probable options: in one, both mounds were workshops, used exclusively for craft activities carried out by people living in any of the residential mounds. In the other, craft specialists households were constrained to live in their own space, separated from the rest of the community. Each of these options has important social implications but the material at hand does not allow any clear cut choice.

5.19 Gnambakouon Sira Tomo Gnambakouon Sira Tomo (GST) (12o 35‘36“N and 3o 20‘ 48“) at 284 m asl., is a small hamlet located away from the river valley and made of three relatively small mounds. It was probably connected one way or the other to the Tora-Sira-Tomo settlement complex situated at some 600 m west. The largest GST-3 mound was tested. The sample of cultural remains collected is rather small but deserves nonetheless to be presented thoroughly.

Despite evidence for violence, in this case fire, which may have resulted from acts of warfare in TST-3-West Occupation I and II dated to AD 997-1158 (ISGS-4590), the first half of the second millennium AD (AD 11001400) witnessed a significant acceleration in the growth of Tora-Sira-Tomo settlement complex, in fact the peak of its growth. A larger and more populated settlement was probably the most effective deterrent to potential aggressors. The complex was then made of 15 mounds, 12 residential and 3 special purpose sites. The quarry and ritual site remained in use. The weaving/cloth dyeing area was turned into a residential space. The blacksmith workshop continued to be in operation. A new karité oil production

5.19.1 The Site The mounds are set in a linear northeast-southwest arrangement in a relatively low lying zone at approximately 50 m from one to the next (fig. 5.129). GST-1 at the northeastern end of the complex is sub-circular in shape, and measures 34 m N-S, 31 m E-W and approximately 1 m in height. GST-2, a small circular 25 m in diameter mound is found in the middle. And finally, GST-3, the tested and largest mound of the complex is situated at the southwestern end. It is a relatively elongated shallow mound, 55 m long, 30 m wide, and 1 m high.

108

Tora-Sira-Tomo Settlement Complex

5.19.2 Gnambakouon Sira Tomo-3 A small excavation probe measuring 9 m2 was sunk in the southwestern half of Gnambakouon-Sira-Tomo-3. The probe was excavated down to the laterite crust at 1 m below the surface and revealed a single occupation documented at 0.40 m. The exposed stratigraphic sequence is made, from bottom to top, of the following sedimentary units (fig. 5.130): 4 - 0.00-0.10 m: Brown grey loose sand and silt with a few gravels. 3 - 0.10 – 0.50 m: Brown red and compact laterite with gravels 2 - 0.60/0.50 – 0.70 m: Brown grey clayey sand 1 - 0.70 – 1.00 m: Yellow brown grey sandy clay Evidence of occupation was exposed at 0.40 m below the surface in the southern half of the probe (fig. 5.131). It includes a series of relatively large laterite blocks delineating the perimeter of a more or less circular feature. There are, in addition, patches of a strongly burnt light grey surfaces and two relatively large iron slag. All this material points to the presence of an iron-working workshop, very likely a small part of a forge or a smelting site. The crushed laterite accumulated in the 40-50 cm thick layer 3 was probably the by-product of the exploitation of this raw material for the production of iron. 5.19.3 Material culture A sample of 548 sherds was collected from GST-3 probe. Most of the material, 483 sherds, comes from layer 2 and

Figure 5.129: Map of Gnambakouon Sira Tomo

Figure 5.130: Stratigraphy of GST 3

109

Archaeology of West African Mound-Clusters

Figure 5.131: Occupation I floor

Figure 5.132: Pottery from GST 3

110

Tora-Sira-Tomo Settlement Complex

Figure 5.133: GST 3 pottery: rims

Summary

appears to be associated with iron-working activities. The topsoil and the bottom layer 1, with respectively 30 and 35 sherds, have a marginal amount that may have “spilled” from layer 2. The recorded sherds are overwhelmingly decorated with twisted roulette impression. Probably because of the context, there is a large proportion of decorated but worn out sherds which cannot be unambiguously assigned to a specific technique. The sherds’ thickness record indicates the presence of the whole spectrum of vessels categories, including relatively thin-walled service wares, medium thick cooking and service pots, and finally thicker short and long-term large size storage vessels (Fig. 5.132 and 5.133).

Data collected from the tested GST-3 point to a specific activity area devoted in this case to iron-working. The rest of the mound may have been used for daily-life dwelling installations for the group of iron-workers who may have operated the workshop exposed in the tested excavation unit. GST-1 and GST-2 could have been the residence of small family units which may have been linked to the main settlement complex of Tora-Sira-Tomo. Unfortunately, no datable material was found during the excavation of GST-3. The settlement seems to have been founded in the first half of the second millennium AD, at the peak or end of the Mouhoun Bend mound dwelling tradition.

111

Chapter 6 Kerebe-Sira-Tomo Settlement Complex

Introduction

along the south flank of the main habitation cluster. They are all arranged along a roughly ENE-WSW axis (fig. 6.1). The material obtained indicates that the settlement complex was inhabited from the second half of the first millennium BC to the second half of the thirteenth century AD.

Kerebe-Sira-Tomo (KST) settlement complex is located at six kilometers southeast of Diekono, approximately two kilometers from the Mouhoun River. The complex of twelve components, with eleven more or less distinct mounds is stretched on 350 m west-east and 600 m north-south (fig. 6.1). The quarry (KST-2) is located at approximately 800 m in the east. The complex is comprised of three distinct sets of sites: the quarry, iron-working stations, and habitations mounds. All habitations mounds are clustered in the north, in what may have been a relatively large 12 - 13 hectares village site. All iron-working stations including ironsmelting and forge workshops are located at 100 to 200 m

6.1: Kerebe-Sira-Tomo-1A Two probes, KST-1A and KST-1B, set 50 m apart, were sunk in the southwestern part of the habitation mounds cluster (fig. 6.1). KST-1A measures 3 by 3 m set around a circular feature protruding from the ground. It was initially thought to be an iron-smelting furnace base. The tested area presents a shallow 0.20 m thick archaeological deposit. The

Figure 6.1: Map of Kerebe-Sira-Tomo settlement complex

112

Kerebe-Sira-Tomo Settlement Complex

exposed circular feature measures 2 m in diameter (fig. 6.2 and 6.3). Its wall was preserved up to 20 cm above the original floor. It is built with mud bricks with a wall measuring 0.15 m in maximum thickness in the east and southeast side. The west and north side are partly eroded. The feature opening is situated in the southeast leading inside to a well preserved strongly burnt laterite floor. The exposed floor is pitch-black at the center shifting to dark red on the sides, suggesting a regular use as a firing place. An intentional deposit of two large cattle molars was found concealed under a large sherd in the northern portion of the floor. A large charcoal piece also found on the feature’s floor is dated to 2050+/-70 BP (ISGS- 3941), cal 348 BC – 117 AD (Annex B).

one hand, this installation may have been a secure pottery firing area. A closely similar feature was observed at Tora, a Bwa village located at a little less than 10 kilometers west of the excavated site (fig. 6.4). An aluminum sheet is generally used to cover the installation during the pottery firing, creating a confined combustion chamber. Such a confinement creates firing conditions close to that of kilns. On the other hand, the installation may have been part of a karité oil production platform, as seen at Koussiri in the southeast of the study area. The karité nuts grinding platform is build on top of a circular cylindrical shaped construction, 1 to 1.2 m high. A number of grinding stations, each with a grindstone and a grinder, are set all along the perimeter of the platform. Karite oil production of this kind is a collective undertaking. Each grinding station is operated by a woman. The specimen seen at Koussiri has 9 grinding stations (fig. 6.5). A fire is lit inside the installation, under the platform. The ground karité nuts are pushed to the center on the hot platform surface. Oil is liquefied, separated from the rest of the ground nuts, and flows through a drain to a large container. The offerings found and deposited to enhance a smooth firing of pottery or a successful karité oil production, could have been useful in either case, making the issue “un-decidable” in the present circumstances. No other feature or any special concentration of cultural remains was found in the rest of the 9 m2 probe.

This feature is too small to be used for human habitation. It could have been used as a chicken coop but the strong evidence of frequent use of fire does not fit with such an activity. Field observations suggest two possibilities. On the

Figure 6.2: KST 1A: view of the circular installation

Figure 6.4: The pottery firing “kiln” at Tora

Figure 6.3: Plan of the circular installation

Figure 6.5: The karité nuts grinding platform of Koussiri

113

Archaeology of West African Mound-Clusters

6.1.1 Material culture

Summary

Beside the cattle molars referred to above, all the cultural remains collected from this special purpose area consist of potsherds (tab.6.1) arranged in two samples. One of the samples is made of the material collected from the probe 0.20 m thick archaeological deposit. The other includes the sherds from the in-feature fill. Both samples combined are made of 191 sherds weighting 6.3 kilograms. There are in fact very few differences between both samples as can be seen in sample composition, decoration techniques, and sherds’ thickness. Most of the rim and body sherds are decorated (tab. 6.1) and decorated sherds are largely predominant. Twisted roulette impression is the favored decoration technique, sometime combined with grooved lines. Sherds thickness presents an overall uni-modal distribution with a peak in the 11.5-14 mm class in one sample, and 9.5-11 mm in the other. Thicker sherds grouped in the >14.5 mm class and belonging to large storage vessels are well represented.

KST-1A probe provides evidence for the earliest occupation of Kerebe-Sira-Tomo settlement complex. It was a pottery firing area or a karité oil production installation located on the edge of the mound. The feature may have been in use for a long time during the second half of the first millennium BC but it is not known for how long. Standard dwelling facilities contemporary with KST-1A installation have not been found in the probes sunk in the habitation mounds cluster. One has to be careful not to make too much of the evidence of absence particularly when a miniscule proportion of the settlement complex has been tested. This having been said, KST-1A probe points to the special purpose use of this part of the settlement complex during the last centuries of the first millennium BC. 6.2 Kerebe-Sira-Tomo 1B Kerebe-Sira-Tomo 1B is located at approximately 50 m southwest of the previous one. The excavation unit measures 9 m2 with a 1.2 m thick archaeological deposit (fig. 6.8). The exposed stratigraphy is made of the following layers:

No complete or restorable vessel was found in the excavated probe. A number of pieces provided clues for a range of rims, vessels shape, and decoration patterns represented in the assemblage (fig. 6.6 and 6.7). There are globular pots with restricted neck and everted rim (fig. 6.6-1 and 6.7-1), small bowls with straight to slightly restricted side (fig. 6.6-4, and 6.7-2), and finally, a flat base pot specimen (fig. 6.6-8). Some patterns of decoration are more elaborate than others. Twisted roulette impression is frequent and found on different part of the vessels, on rim and body (fig. 6.6 and 6.7). It is sometime combined with grooved lines. The latter are exceptionally used alone but most of the time in association with other motifs like comb impression (fig 6.7-7).

7 - 0.35-0.00 m: Compact brown gray silty sand with gravels. 6 - 0.40-0.35 m: Dark red house floor 5 - 0.50-0.40 m: Yellow brown gray clayey sand 4 - 0.60-0.50 m: Two dark red house floors 3 - 0.90-0.60 m: Yellow brown gray clayey sand 2 - 1.00-0.90 m: Three yellow, red, yellow, gray house floors 1 - 1.20-1.00 m: Brown-yellow sandy clay

Table 6.1: Characteristics of KST-1A sherds’ population n

Weight (kgs)

0.00-0.20 0.20 (feature)

91 100

Total

191

Stratigraphy (m)

Rim

Body

D N

U

D N

3 3.3

5 6

1

3 -

54 45

6.3

11

1

3

99 12

U

Base

Sample composition

Decoration techniques

5 7

24 39 63

2 2 Composite

D

N

TR

GL

I

Other

TR-GL

Other

5 90

51 10

35

1

1

4 1

11

4

Total

176

15

86

1

1

1

15

4

Thickness classes (mm)

5-7

7.5-9

9.5-11

11.5-14

>14.5

0.00-0.20 0.20 (feature)

2 1

20 12

21 34

28 26

20 27

Total

3

32

55

54

47

0.00 – 0.20 0.20 (feature)

86

Key: D = Decorated; N = Non-decorated; U = Unknown; TR = Twisted Roulette impression; GL = Grooved line; I = Incision.

114

Kerebe-Sira-Tomo Settlement Complex

Figure 6.6: Pottery from KST 1A sample 1

Figure 6.7: Pottery from KST 1A sample 2

115

Archaeology of West African Mound-Clusters

6.2.2 Occupation II Occupation II floor was exposed at 0.60 m below the surface. The recorded cultural remains include laterite flooring all over the excavated probe (fig. 6.10). It is comprised of a relatively large 2.5 m radius round hut floor adjoined on the north flank by a small courtyard portion. The remains of a storage bin made of unfired clay measuring 1 m in diameter were found collapsed on its four supporting laterite blocks. A small portable sandstone grindstone was still on the house floor in the east side of the exposed surface. The remaining material consists exclusively of pottery of different size and shape used for storage (feature 8, 10 and 13), cooking (feature 12), and serving (feature 11). The excavated portion of this domestic unit appears to have been devoted to the storage of supplies and food preparation activities. Feature 15, an occupation II burial was found in the western part of the excavated probe in occupation I deposit (fig. 6.9). The deceased was buried in a tightly flexed position, lying on the left side with hands before the face, and oriented southeast-northwest (fig. 6.11). The head was protected by a large bowl used as “helmet”. Two additional bowls were found associated with the burial. Occupation II is dated to AD 1020-1280 (ISGS-3932, 880+/-70 BP) (Annex B).

Figure 6.8: Stratigraphy of KST 1B

6.2.3 Occupation III Occupation III deposit is at the top of the exposed stratigraphic sequence, from the surface to 0.40 m below. Occupation III habitation surface was found at 0.40 m and consists of a large portion of a house floor abutting a courtyard (fig. 6.12). The exposed round house floor, made of crushed laterite “cement”, is extended virtually all over the excavated probe. It measures at least 3 m in radius, suggesting the presence of a large house, part of a larger household complex. With the exception of one

Figure 6.9: Occupation I floor

Three occupation episodes have been documented 6.2.1 Occupation I Occupation I deposit is found at 0.90 below the surface. It includes a series of three superimposed house floors exposed at 1.00 - 0.90 m, spread all over the excavated probe (fig. 6.9). Large potsherds are found along the northsouth axis with the remains of a 1 m long, 0.20 m thick mud-brick wall associated with a door-socket in dolerite. Occupation I is dated to AD 240-550 (ISGS-3936, 1660+/70 BP) (Annex B).

Figure 6.10: Occupation II floor

116

Kerebe-Sira-Tomo Settlement Complex

Figure 6.11: Occupation II burial

laterite block, all the excavated cultural remains consists of abandoned complete vessels, in more or less good state of preservation (fig. 6.13). 9 such containers were exposed. Two rather large specimens in the south and west of the exposed house floor are large storage jars (fig. 6.14). One, feature 3 at the center of the probe, is a sophisticated steamer and the remaining six specimens clustered in the west half of the probe are relatively large cooking and storage vessels. Remarkably, all the recorded pottery is found on the house floor with none recorded on the exposed small portion of the courtyard (fig. 6.13). A charcoal sample collected from the upper part of Occupation III deposit at 0.20 m below the surface dates the episode to AD 980 – 1260 ( ISGS-3937, 930+/-70 BP) (Annex B).

Figure 6.13: Occupation III floor

of courtyard with domestic gear found throughout. The occupation sequences brought to light at KST-1B spans from AD 230-550 to 1020-1280 (Annex II). The radiocarbon chronology points to a half millennium occupation hiatus between Occupation I and II. 6.2.4 Material culture Besides the exposed built installations and dwelling features, pottery and a small amount of heavy duty coarse stone pieces are the essential components of the material culture recorded in KST-1B. The sherds population is made

The domestic character of the units uncovered in KST-1B is clear and straightforward. In all the documented three occupations, a habitation floor is connected to a portion

Figure 6.12: View of occupation III floor

117

Archaeology of West African Mound-Clusters

Table 6.2: Characteristics of KST-1B sherds’ population Stratigraphy (m)

n

Rim

Body

Weight (kgs)

D

N

U

D

N

U

Base

Sample composition Surface 0.00-0.20 0.20- 0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20

111 4 122 29 19 132 23

1.8 1.9 4.07 0.89 1.1 8.35 0.75

4 2 15 2 6 15 1

3 4 1 -

1 4 4 -

59 2 74 18 12 72 14

3 3 3 1

41 18 4 28 7

1 1 2 -

Total

440

18.77

45

8

9

251

10

98

4

D

N

TR

GL

I

Other

TR-GL

Other

Surface 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20

105 4 111 28 18 115 22

6 11 1 1 17 1

45 63 15 10 63 11

4 -

6 1 1

1 -

6 4 14 4 8 24 3

8 4 1 -

Total

403

37

207

4

8

1

63

13

Thickness classes (mm)

5-7

7.5-9

9.5-11

11.5-14

>14.5

Surface 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20

5 11 1 -

26 16 6 3 19 5

42 47 7 14 67 4

27 24 9 1 28 8

11 4 24 7 1 17 6

Total

17

75

181

97

70

Decoration techniques

Composite

Key: D = Decorated; N = Non-decorated; U = Unknown; TR = Twisted Roulette impression; GL = Grooved line; I = Incision.

6.14, Tab.6.4). The recorded domestic set includes three vessels categories: a small bowl (fig. 6.14-2); the base of a large footed plain bowl (fig. 6.14-5); three relatively large pots with restricted neck and everted rim, decorated with twisted roulette impression (fig. 6.14-7 and 8), combined with grooved lines in one case (fig. 6.14-6); and finally, a large elongated storage jar with restricted neck and everted rim decorated with twisted roulette impression and grooved lines (fig. 6.14-9).

of 440 pieces weighting 18.77 kilograms (tab. 6.2) unevenly distributed throughout the deposit. The largest amount per occupation is found in Occupation I and the lowest in Occupation II. Occupation II figures are explainable by the large proportion of complete vessels found in-situ on the house floor. Most of the rim and body sherds are decorated. Twisted roulette impression is the predominant decoration technique followed by far by combined twisted roulette and grooved line. The distribution of sherds throughout thickness classes is characteristically uni-modal with peaks in the 9.5-11 or 11.5-14 mm classes, those of cooking and storage vessels. Thicker sherds of the >14.5 mm class are however well represented confirming the insight gained from the general distribution of complete and restorable vessels throughout the cultural deposit.

Occupation II sample is comparatively large with 11 specimens, including the three vessels from burial 15 (fig. 6.15, Tab.6.4). The latter are divided into an elaborately decorated large bowl (fig. 6.15-1) used as the deceased “helmet” and 2 small footed bowls (fig. 6.15-3 and 4). The domestic set is made of 8 specimens distributed into service, cooking, and storage sub-sets (fig. 6.15). The service subset includes three vessels, a shallow open bowl decorated with vertical bands of twisted roulette impression and grooved lines (fig. 6.15-1), a small carinated pot decorated

Occupation I complete or restorable vessels sample is made of two distinct sets, one comprised of three vessels coming from occupation II burial in feature 15 and the other with 5 specimens found in the rest of the exposed surface (fig.

118

Kerebe-Sira-Tomo Settlement Complex

Figure 6.14: Pottery from Occupation I

Figure 6.15: Pottery from Occupation II

119

Archaeology of West African Mound-Clusters

Table 6.3: Measurements of vessels from KST-1B Vessel No

Feature

MD (cm)

MaD (cm)

DB (cm)

HT (cm)

HMD (cm)

Th (cm)

Decoration

29 22 20 9

? ? 23 11

? ? 6.5

13(?) 11(?) 15(?) 9

? ? ? 4

1.1 0.85 1.1 0.70

TR TR+GL TR+GL TR+I

41 10 19 25 23 11 10.5

70 12.5 23 38 23 11 10.5

5.5 4.5

45 9.5 20 24 9 6.5 6.5

25 3 9 4 9 6.5 6.5

1.8 0.70 1.3 1.3 0.75 0.60 0.85

TR+GL TR+GL TR TR TR+I+GL TR+I TR+I

large pot small pot large pot large pot large bowl footed bowl footed bowl

24 39 22 21 16 13.5 18 20 32 23.5

26 ? 28 25 21.5 13.5 18 27 43 32.5

10 10 6 -

9 25(?) 30 25(?) 18 6 9 24 40 29

7 ? 20 6 10 6 9 8 20 10

1.0 1.7 1.3 1.2 0.80 0.70 0.80 1.2 1.1 1.1

TR+GL TR+GL TR+C+GL TR+C+GL TR TR+GL TR+I TR+GL TR TR+GL

small basin large pot large pot steamer footed pot bowl large bowl large pot large pot large pot

Nature

Occupation I (1.00 m) 1 2 3 4

14 -

large pot large pot large pot bowl

Occupation II (0.60 m) 5 6 7 8 9 10 11

10 11 12 13 15(burial) 15(burial) 15(burial)

Occupation III (0.40 m) 12 13 14 15 16 17 18 19 20 21

1 2 3 4 5 6 7 8 9

Key: MD = Mouth diameter; MaD = Maximum diameter; BD = Base diameter; HT = Height; Th = Thickness; TR = Twisted roulette impression; C = Comb impression; I = Incision; GL = Grooved line.

with combined twisted roulette impression and grooved lines (fig. 6.15-2), and a hemispheric pot decorated with bands of twisted roulette impression (fig. 6.15-3). The cooking sub-set is made of two specimens of globular to slightly elongated pots with restricted neck and everted rim, decorated with twisted roulette impression with and without grooved lines (fig. 6.15-4 and 5). And finally, the storage sub-sets is comprised of three very large globular to elongated jars with restricted neck and everted rim, decorated with twisted roulette impression (fig. 6.15-6 and 7), or combined twisted roulette impression and horizontal grooved lines (fig. 6.15-8). Interestingly, Occupation II pottery presents a new decoration syntax characterized by the use of a narrow horizontal band of twisted roulette impression at the vessel’s neck (fig. 6.15). It emerges at the end of the first /beginning of the second millennium AD. The recorded stone artifacts are all part of grinding equipment in sandstone. They are divided into broken (fig. 6.16-1 and 4) or complete (fig. 6.16-5) portable grindstones, one grinder (fig. 6.16-2), and an undetermined flake (fig. 6.16-3).

Figure 6.16: Coarse stone artifacts from Occupation II

Occupation III has a large and diverse pottery sample and a few specimens of stone artifacts (tab. 6.4, fig. 6.17-19).

120

Kerebe-Sira-Tomo Settlement Complex

Figure 6.17: Pottery from Occupation III

Figure 6.18: Pottery from Occupation III

121

Archaeology of West African Mound-Clusters

Table 6.4: Characteristics of the sherds’ population from KST-3 Rim

Body

n

Weight (kgs)

D

N

U

D

N

U

291 191 129 141 644 778

23.62 26.60 5.50 5.40 36.70 41.02

27 11 11 15 56 65

2 2 5 14

8 2 4 3 24 10

192 154 84 108 437 579

2 1 1 1 11 14

60 10 28 11 105 87

1 13 1 7 9

2,175

134.84

185

23

51

1,554

30

301

31

D

N

TR

GL

I

Composite TR-GL

219 178 95 123 498 644

64 13 33 17 145 134

163 128 77 87 336 477

1 1 7 15 8

4 -

44 49 17 29 98 168

1,757

406

1,268

32

4

405

Thickness classes (mm)

5-7

7.5-9

9.5-11

11.5-14

>14.5

0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20

6 3 3 5 22 25

29 3 17 23 95 83

63 21 27 27 161 225

105 109 52 46 191 264

82 55 309 40 176 189

Total

64

250

524

787

572

Stratigraphy (m)

Base

Sample composition 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 Total Decoration techniques 0.00-0.20 0.20-0.40 0.40-0.60 0.60-0.80 0.80-1.00 1.00-1.20 Total

Key: D = Decorated; N = Non-decorated; U = Unknown; TR = Twisted Roulette impression; GL = Grooved line; I = Incision.

comb impression is new in this occupation which spans the first few centuries of the second millennium AD.

The collected pottery includes 7 service/food consumption vessels: a shallow open bowl (fig. 6.17-1), a large flat base bowl (fig. 6.17-2), a plain footed bowl (fig. 6.18-1), a small plain bowl (fig. 6.18-2), a small globular pot with restricted neck (fig. 6.18-3), a footed globular pot with restricted neck (fig. 6.17-3), and finally, a flat base carinated (?) pot (fig. 6.18-7). Four cooking vessels have also been recorded. One is a bell-shaped steamer (fig. 6.17-7) decorated with twisted roulette impression over-lined on the neck by a horizontal band of comb impression. The remaining specimens (fig. 6.17-4-6) are all globular to elongated with restricted neck and everted rim, each with a distinct pattern of decoration which includes twisted roulette impressions, grooved lines, and comb impressions. The storage category is comprised of three specimens, a large flat lid (fig. 6.18-4), the upper part of a likely globular liquid container with constricted neck and everted rim (fig. 6.18-5) decorated with a horizontal band of twisted roulette impression on the neck, and finally, the upper portion of a large globular and elongated jar with restricted neck and everted rim (fig. 6.18-8).

The coarse stone artifacts are divided into one broken piece of grindstone (fig. 6.19-1), a small portable grindstone (fig. 6.19-3), and a spherical hammer-stone (fig. 6.19-2). The relative high number of sherds (tab. 6.2) and the presence of large rim pieces allowing for the reconstitution of partial vessel’s shapes in the top 0.20 m of the stratigraphic sequence suggest the presence of an eroded occupation IV deposit. The reconstructed shapes includes a small bowl (fig. 6.20-1) decorated with comb impression and grooved lines, an elongated pot with restricted neck and everted rim (fig. 6.20-2), and finally, a large elongated storage jar with everted rim combining twisted roulette impression and grooved lines (fig. 6.20-3). Summary KST-1B presents evidence of an interesting occupation history which started around the middle of the first half of the first millennium, ca 250 AD, and ended in the middle of

The narrow horizontal band of twisted roulette on the neck is found on 7 of Occupation III vessels. The use of bands of

122

Kerebe-Sira-Tomo Settlement Complex

Figure 6.19: Coarse stone artifacts from Occupation III

Figure 6.20: Pottery from the eroded Occupation IV

the first half of the second millennium, ca. 1250 AD. There was however an occupation hiatus of approximately half a millennium between occupation I and II. The possibility of a disturbed and eroded occupation IV deposit is genuine. If taken into account, the occupation history of KST-1B may well have been stretched to the middle of the second millennium AD. All the components of the exposed and collected cultural remains belong to the realm of standard domestic activities, with house floors, portions of courtyard, storage, cooking, and service vessels. The 9 m2 of the probe sampled a tiny proportion of more encompassing domestic units in all three occupations. Faunal remains were not well represented in the tested probe. They amount to a handful of pieces: 1 bovidae sp. molar fragment and two undetermined

birds’ long bones in occupation I, and two fresh water mussels’ shells (Aspartharia sp.) in occupation II deposit. 6.3: Kerebe-Sira-Tomo-2 Kerebe-Sira-Tomo 2 is a relatively large quarry site located at a little less than one kilometer in the East-northeast of the main mound cluster. This site was not strictly speaking excavated but was cleaned of grass and shrubs to be measured and mapped (fig. 6.21). It is a sub-circular excavation measuring approximately 30 m in diameter, 5 m in depth, with a concave bottom. The quarried laterite deposit is made of three thick strata (fig. 6.22). The upper stratum is 2 m thick in the north and 1 m thick in the south. The middle stratum presents the same thickness at both

123

Archaeology of West African Mound-Clusters

Figure 6.21: View of KST 2, the quarry

Figure 6.22: View of the laterite deposits

Figure 6.23: Section of the quarry

124

Kerebe-Sira-Tomo Settlement Complex

Figure 6.24: Stratigraphy of KST 3

Figure 6.25: Profile of Occupation I laterite floors

125

Archaeology of West African Mound-Clusters

1 - 1.60 – 1.30 m: The bottom natural yellow silty clay deposit. Four occupation episodes have been documented all along the KST-3 stratigraphic sequence. 6.4.1 Occupation I Evidence from occupation I was exposed at depth ranging from 1.00 m to 1.20 m. The uncovered remains consist of a thick dark-red layer of superimposed house floors, overlain by a thick ash deposit, charcoal pieces, and a burnt wooden

Figure 6.26: Occupation I floor

ends, ranging from 1 to 1.5 m. The third and lower stratum does not seem to have been exploited. It is exposed at the bottom of the quarry and carries large laterite blocks left from the quarrying of the overlying middle and upper strata (fig. 6.23). It is very likely that all the material used for houses and courtyards flooring as well as iron ore was collected from this quarry. The upper stratum generally exposed to the elements is in most cases extensively oxidized providing poor iron ore but good flooring material. The lower strata are the main target of iron smelters; they tend to be relatively softer with higher iron content on the average (Coulibaly 2006, Kiethega 2006). The raw material extracted from KST-2 quarry amounts approximately to 3,534.3 cubic meters.

Figure 6.27: Occupation II floor

6.4 Kerebe-Sira-Tomo - 3 Kerebe-Sira-Tomo 3 is located on the eastern side of the largest single mound of the complex (fig. 6.1). It is a 3 m by 3 m excavation unit. The exposed stratigraphy measures 1.60 m in thickness comprised of eight layers arranged from bottom to top as follows (fig. 6.24 and 6.25): 8 - 0.10 – 0.00 m: Brown gray loose sandy silt. 7 – 0.50 – 0.10 m: Light brown sandy clay with gravel. 6 - 0.60 – 0.50 m: A series of superimposed red house floors. 5 - 0.75 – 0.60 m: Dark gray sandy clay with charcoal pieces. 4 - 1.00 – 0.75 m: Brown–gray sandy clay. 3 - 1.20 – 1.00 m: Thick deposit of light-gray to dark gray ash with charcoal and a large burnt log. 2 - 1.30 – 1.20 m: A series of superimposed dark-red house floors

Figure 6.28: Occupation III floor

126

Kerebe-Sira-Tomo Settlement Complex

Table 6.5: Faunal remains from KST-3 Taxa

I

II

III

IV

Total

44 -

29 31

30 1

-

103 32

17

76

5 38

-

5 131

-

3 7

-

-

3 7

2

5

2

-

9

1 6 90

45 97

1 50 74

-

2 101 261

-

50

13 8 -

-

13 8 50

160

343

222

-

725

Mammals Domestic animals Bos Taurus Ovis/Capra sp. Wild animals Felix sp. Bovidae sp. Reptiles Crocodylus niloticus Varanus niloticus Shells Etheria elliptica Fish Lates niloticus Gymnarchus niloticus Others Birds Numida meleagris Undetermined Undetermined Total

log (fig. 6.25 and 6.26). It is clear from the material record that the excavated house fell victim of fire. The exposed portion of the circular house floor measures 3 m in the east along the north-south axis, and 2 m in the north along the west-east axis (fig. 6.26). One grindstone and two grinders in sandstone as well as two large laterite blocks were found in the northwestern corner of the excavated unit. Finally, few large sherds were uncovered along the southern section. Occupation I is dated to AD 1042-1379 (ISGS 4319, 790+/70 BP) (Annex B).

for the lines, and three for each of the rows. The distance between neighboring blocks varies from 0.30 to 0.75 m. The recorded installation represents the supporting blocks of a granary built with mud-bricks. It is raised above the ground to avoid humidity and termites that plague the region. The installation was probably built in the courtyard of a domestic unit, and the exposed surface was hard and compact. Occupation III is dated to AD 1070-1385 (ISGS – 4310, 770+/-70 BP) (Annex B). 6.4.4 Occupation IV

6.4.2 Occupation II

Finally, evidence from occupation IV found at 0.20 m at the bottom of the top deposit appears to have been partly disturbed. The uncovered material remains consist of a crushed but restorable pot exposed in the northwestern corner of the probe and a scatter of large sherds found one meter southwest (fig. 6.29). Occupation IV is dated to AD 1039-1376 (ISGS – 4302, 800+/-70 BP) (Annex B).

Occupation II was exposed at 0.75 m and consists of a scatter of potsherds, fragmented stones pieces and animal bones (fig. 6.27). The amount of ash and charcoal found in the deposit do not suggest a residential use. The probe 3 portion of the site appears to have been used as a refuse dumping area during occupation II, which is dated to AD 1039-1376 (ISGS 4311, 800+/-70 BP) (Annex B).

For less than three centuries, from AD 1039/1070 – 1376/1385, the build-up of mound III seems to have occurred at a relatively fast speed. The use of the space in the tested area shifted from habitation in occupation I, to a refuse dumping area in occupation II, a storage area in occupation III, and finally, a portion of courtyard in occupation IV. The fire that seems to have destroyed the house from occupation I may have had different causes: a domestic accident, a purposeful destruction, or a violent act of warfare. This issue will be considered later.

6.4.3 Occupation III Occupation III was uncovered at 0.40 to 0.60 m. The exposed cultural remains consist of a series of large laterite blocks, large potsherds probably from jars, and one crushed large storage jar found along the northern side of the probe. The uncovered large laterite blocks are arranged into two parallel SW-NE lines and three NW-SE rows (fig. 6.28). They are comprised of five and three blocks

127

Archaeology of West African Mound-Clusters

Table 6.6: Kerebe-Sira-Tomo 4: Patterns of distribution of vessels from Occupation II and III Vessel No

MD (cm)

MaD (cm)

DB (cm)

HT (cm)

HMD (cm)

Th (cm)

Decoration

20 12.5 25 25 28 30 11

22 13 34 35 40 35 24

5 -

7 8.5 32 34 33 32 ?18

6 5 18 17 14 17 ?

0.7 0.7 1.3 1.3 1.4 1.5 1.0

TR+Redslip TR+GL TR+GL TR+GL TR TR+GL Awldots

16 16

22 16

5

14.5 5

6.5 ?

0.8 0.7

TR+GL TR

Hemispheric Pot Flat base basin

30 37 28 31 35 15 20 15 50 14

50 50 40 31 35 24 32 20 70 30

8 9 10

?29 ?43 ?22 12 17 20 13 14 20 30

? 20 ? 12 17 ? ? ? ? 12

1.3 1.5 1.4 0.8 1.0 1.1 1.0 0.7 1.5 1.0

TR+GL TR+GL TR+GL TR TR+GL TR+GL TR+GL TR+GL TR TR+GL+

Elongated jar Elongated jar Elongated jar Flat base basin Flat base basin Globular pot Globular pot Globular/elongated pot Large spherical jar Globular pot

10 16 20 12

20 28 32 12

-

10 ?13 36 4

7 ? 14 ?

0.6 0.8 1.5 2.0

TR TR TR Plain

Globular pot Globular/Elongated jar Large globular jar Footed bowl

20 20

20 32

10 -

4 18

?

1.2 1.0

Plain TR+GL

flat base basin Large globular jar

30 8.5 16

41 10.5 16

5

40 3 7

20 7.5 16

1.0 0.8 0.8

TR+GL TR+GL TR

Elongated jar Hemispheric pot Flat base large bowl

8

27

-

25

13

0.8

TR+GL

Large jug

22 30

30 35

-

35 35

15 15

1.1 1.3

TR+GL TR+GL

Large globular jar Large globular jar

18 16 23

20 28 31

-

10 24 28

8 9 14

1.0 0.9 0.8

Plain TR+GL TR+GL

Small hemispheric pot Globular/Elongated pot Globular/elongated pot

10.5 14 18 11

13 14 18 24

6 6 -

9 10 10 18(?)

-

0.6 0.9 0.7 1.0

TR TR+GL CS TR

Small globular pot Flat base bowl large flat base bowl large globular pot

Nature

Occupation II House 1 (storage) 1 2 3 (jar 6) 4 (jar 8) 5 (jar 10) 6 (jar 7) 7 House E3 1 2 House E5 1 2 3 4 5 6 7 8 9 10 House E7 1 2 3 4 House E9 1 2 House E10 1 2 3 House E11 1 House W1 1 2 Others 1 2 3 Occupation III 1 2 3 4

128

Hemispheric steamer Flat base pot Globular jar Globular jar Globular/elongated jar Globular/elongated jar Large spherical jug

Kerebe-Sira-Tomo Settlement Complex

6.4.5 Material culture A sample of 2,175 potsherds weighing 13.5 kilograms was collected from KST-3. The general distribution points to a decreasing frequency of sherds from occupation I to IV (tab. 6.6). The earlier occupation I and II deposits have 778 and 785 potsherds respectively. The later occupation III and IV have 320 and 291. More than ¾ of the collected sherds are decorated, predominantly with twisted roulette impression. Twisted roulette impression combined with grooved lines is also well represented in all the occupations (tab. 6.6). All the thickness classes are found in all the samples with however varying frequencies. Medium to thick walled vessels from the 9.5-11 to >14 mm thickness classes are largely predominant. Thin walled vessels are relatively rare in occupation III and IV deposits. Pottery fragmentation was particularly high in KST-3 probe. Very few complete or restorable vessels were recorded (fig. 6.30). Two specimens were found in occupation IV. One is a fragment of a relatively large storage jar with restricted neck and everted rim, measuring 22 cm in height, 25 cm in mouth diameter, 45 cm in maximum diameter, and 1.3 cm in wall thickness (fig. 6.30a). It is entirely decorated with oblique twisted roulette impressions. The other is a small globular plain bowl, 12 cm in mouth diameter, 8 cm in height, and 0.7 cm in wall thickness (fig. 6.30b). One small flat base pot with straight side was found in occupation III. It measures 15 cm in mouth diameter, 9 cm in height, 7 cm in base diameter, and 0.6 cm in wall thickness, with its

Figure 6.29: Occupation IV

Figure 6.30: Pottery from KST 3

129

Archaeology of West African Mound-Clusters

Figure 6.31: Base of a tri-legged bowl

Figure 6.33: Coarse stone artifacts from Occupation I

6.5 cm long and 0.6 to 0.5 cm in diameter. The remaining three pieces include a small axe (fig. 6.32b), a hoe-blade socket (fig. 6.30e), and an undetermined iron sheet fragment (fig. 6.32f). The miniature axe measures 7.8 cm long, 4 cm in maximum width, and 0.6-0.4 cm in thickness. It has a triangular blade, a round distal end, and a 2-3 cm stem. The preserved portion of the hoe-blade socket measure 5 cm in length, 3 cm in width, and 0.3 cm in thickness.

Figure 6.32: Iron artifacts

Coarse stone artifacts in sandstone, all used for grinding purposes, were found in three of the four occupations. The sample from occupation I found on the house floor includes two complete specimens and one fragmented grindstone (fig. 6.33a). Both complete pieces are grinders. One is crescent-shape, 37 cm long, 12 cm wide, 6 cm thick, weighing 3,5 kilograms (fig. 6.33c). The other is smaller, 11.8 cm long, 5.1 cm wide, 4.7 cm thick, weighing 0.3 kgs (fig. 6.33b). The samples from occupation II and III (fig. 6.34) are all relatively large grindstone fragments. They measure 9 to 23 cm in length, 9 to 16 cm in width, 4.8 to 3.2 cm in thickness, and weigh 0.6 to 1.7 kgs.

lower part decorated with twisted roulette impression (fig. 6.30c). The round base of a footed bowl is the last item found in occupation II deposit (fig. 6.31). It is similar to the wares common along the lower Niger and the Inland Niger delta from the 6th to the 17th century, and may have been obtained through the inter-regional long-distance exchange network. It could also have been a local copy. All the iron artifacts were collected from the trash dumping area of occupation III. Three of the pieces (fig. 6.32a, c, and d) are probably fragments of spear or arrow tangs, 5 to

130

Kerebe-Sira-Tomo Settlement Complex

The faunal material from occupation II, the largest sample of the tested probe, was collected from a refuse dumping area (tab. 6.5). A broader range of species is represented, including cattle and sheep/goat for domestic animals, catfish and others, fresh water oysters, as well as two reptiles, the monitor lizard (Varanus niloticus) and the crocodile (Crocodylus niloticus). The 23 sheep/goat remains, distributed into 3 teeth, 2 mandibles, 3 vertebras, 2 distal metapodials, 5 long bones, and 6 phalanges, tend to indicate that a single mature animal was butchered, consumed, and the bones thrown in the refuse heap. Nine taxa are represented in the faunal sample from occupation III which amounts to 222 pieces. The mammalian component consists of cattle, sheep/goat, Bovidae sp., and wild cat (Felix sp.) bones, with frequency varying from 1 to 29 (tab. 6.5). The wild cat remains include 1 right humerus, 1 atlas, 1 thoracic vertebra, and 2 mandibles. Birds’ remains are divided into 8 undetermined fragments and 13 guinea fowl (Numida meleagris) bones. The latter are thinner then chicken (Gallus gallus) bones with however a certain overlap. Aquatic resources are partitioned into shells and fish. The former consists of 2 shells of fresh water oyster (Etheria elliptica) and 125 fish bones, from nile perch (Lates niloticus), different kinds of catfish (Gymnarchus niloticus, Synodontis sp., Heterobranchus sp.), as well as undetermined fish species.

Figure 6.34: Coarse stone artifacts from Occupation II and III

Keeping in mind the small size of the tested probe, the absence of wild mammals from the recorded samples is remarkable. Animal proteins and fat seem to have been obtained from two main spheres of activities, livestock husbandry of cattle and sheep/goat on the one hand and the exploitation of aquatic resources on the other.

Despite the small sample size, the repertoire of the material culture collected in KST-3 is representative enough of the range of activities performed in domestic units. 6.4.6 Faunal Remains

Summary

The sample of faunal remains collected from KST-3 amounts to 723 pieces recorded from occupations I, II and III (tab. 6.5). Occupation IV deposit at the top of the stratigraphic sequence was devoid of animal bones.

The part of the mound complex in which KST-3 probe was sunk was settled from the beginning to approximately the middle of the second millennium AD (Annex B). The excavated area was used for different activities all

The faunal material from occupation I is partitioned into two main components, domestic animals and aquatic resources. The latter includes one nile perch (Lates niloticus) vertebra, six fragments of catfish bones (Gymnarchus niloticus), a few shell fragments of fresh water oyster, and 90 undetermined fish bones. The domestic animals part is made exclusively of cattle bones. The recorded bones point to a specific portion of the animal which may have been roasted. The sample is made of 5 fragments of both proximal and distal ends of the left femur, 3 fragments of the left femur head, 4 large fragments of the mesial part of the femur, 6 pelvis fragments, as well as 26 smaller fragments of femur and coccyx bones. The distal and proximal ends of the left femur are burnt; a possible signature of meat roasting. However, considering that occupation I has witnessed a fire which very likely burnt most of the dwelling installations, the tossed away cattle bones could have been burnt unintentionally.

Figure 6.35: Stratigraphy of KST 4

131

Archaeology of West African Mound-Clusters

Figure 6.36: Plan of Occupation II features

connected directly to the multiple aspects of domestic life. The dwelling facilities from occupation I were burnt by a fire. It could have been a simple accident of ordinary life or the result of a hostile expedition. An important amount of burnt karité kernels (Butyrospermum parkii) was found among the burnt debris of occupation I house floor. The presence of a large crescent-shape grinder used in the karité oil production process, precisely for the grinding of boiled nuts, adds to the evidence and suggests the production of karité oil at the beginning of the second millennium AD, around 1000 AD. The area was used for refuse disposal during occupation II, for a grain storage installation during occupation III, and as part of an open courtyard during occupation IV

Figure 6.37: View from the north of the exposed occupation II features

6.5: Kerebe-Sira-Tomo - 4 Kerebe-Sira-Tomo-4 probe is located at the center a small distinct mound situated at 40 m southeast of the main mound cluster. It measures 50 m in length north-south and 30 m in width west-east. The excavation started with a 9 m2 probe, in a 3 x 3 m grid. It was extended sequentially to reach 120 m2 in order to expose the maximum of well preserved features from occupation II. Occupation II

deposit was found at 0.70 - 0.80 m below the surface. The probe was again narrowed back to a 3 x 1 m trench in the south side of the excavated unit to complete the exposure of KST-4 archaeological sequence (fig. 6.35). The recorded stratigraphic sequence measures 1.50 m in thickness with four layers and four occupation levels:

132

Kerebe-Sira-Tomo Settlement Complex

4 – 0.20 – 0.00 m: Brown gray partially reworked sandy silt. 3 – 1.00 – 0.20 m: Dark gray sandy clay with large amount of charcoal and ash. 2 – 1.20 – 1.00 m: Brown sandy clay with gravel sandwiched between two red floors. 1 – 1.20 – 1.60 m: The bottom natural yellow silty clay.

north-south (fig. 6.37). The quality and the high resolution of the recorded data allow for an interesting analysis of the structure of households in terms of features arrangements and distribution of domestic gear. The exposed houses are arranged into three distinct clusters, the central, northern, and southern clusters. 6.5.2.1 The Central cluster

6.5.1 Occupation I

The central cluster includes five round huts, House (H)-1, H-2, H-E1, H-E3, and H-W1 (fig. 6.36 and 6.37) delineated by a pise wall. The houses are distributed into two subsets in the west and east of an “all-paved” courtyard. The western sub-set is comprised of three houses, H-1, H-2, and H-W1. These round huts, with their burnt walls still preserved to a height of 0.50 m, were found buried under a thick deposit of ash and charcoal (fig. 6.38).

In order to preserve occupation II exposure, occupation I deposit was reached through a small 3 x 1 m trench located at the south end of the excavation unit. The probed occupation level was rather poor in archaeological remains. The latter consist of a few potsherds and a dark-red laterite floor uncovered at 1.20 m and extended all over the 3 m2 trial trench. A charcoal sample dates occupation I to AD 1191-1399 (ISGS – 4345, 720+/-70 BP) (Annex B).

H-1, in the south, measures 2.90 m in diameter. It includes an earth platform built in the west half, supported by a series of large laterite blocks, and filled with hundreds of large sherds set in a clay matrix. It was probably the bed, and in this case, H1 was a sleeping house.

6.5.2 Occupation II Occupation II, exposed at 0.70 to 0.80 m, consists of relatively well preserved remains of a series of round huts clusters. The excavated unit is L-shaped. It measures 16 m west-east and 12 m north-south (fig. 6.36). The southern part of the probe, oriented west-east, is irregular in shape, and measures 16 m in length and 3 to 7 m in width. The eastern portion, 5 m long and 4 m wide, is rectangular and oriented

The smaller northern H-2 measures 2.50 m in diameter. The house floor was carefully maintained as shown by the number of superimposed layers. An adult was laid to rest on the floor without grave pit in the southern half of the hut (fig. 6.38). He/she was oriented west-east, laid on the right

Figure 6.38: View of house H-1 and H-2 with a burial

133

Archaeology of West African Mound-Clusters

Figure 6.39: View of the eastern hut cluster

A mud-brick wall connecting H-1, H-E3, and H-E1 closes the sub-sets in the south. The unit was probably accessed from the north, and from there one would have proceeded south through a passage between H-2 and H-E1. The presence of additional huts in the northern side of the exposed cluster is highly probable. One large pot and a jar were found next to H-2 on the southeastern side (fig.

side, facing south, the right arm extended and the left one folded, with the legs slightly flexed. A large pot was laid on top of the grave above the feet. H-2 was accordingly a sleeping hut which was later turned into a grave. H-W1 in the northwest measures 2.70 m in diameter. A large portion of a pisé wall was preserved in the south. Large pieces of storage jars and four large laterite blocks set at right angle were found in the south and east of the feature. The latter were probably part of a storage installation, suggesting that this hut may have been used for the storage of the domestic unit supplies and cooking purposes. The eastern sub-set includes H-E1 and H-E3, both connected by a mud brick wall (fig. 6.39). H-E1 measures 1.90 m in diameter with its doorway opening to the south. A large number of large sherds associated with large laterite blocks were found scattered on the eastern half of the hut, suggesting the presence of a bed similar to the specimen from H-1. H-E1 thus appears to have been another sleeping room on the opposite side of the courtyard.

Figure 6.40: View of the chicken egg-nest

H-E3 on the southeast flank of the previous one is the largest hut of the cluster. It measures 3.20 m in diameter and contained an important amount of artifacts including several clay vessels, 1 large grindstone and grinder both in sandstone, a large fragment of a portable hearth, as well as a large concentration of sherds and laterite blocks in the centre and north of the house. The recorded pottery belongs to the cooking and service categories. One is a hemispheric pot with elaborate decoration made with twisted roulette impression combined with grooved lines and the remaining two, flat base basins. An arm-ring and bell both in iron were found among the scattered material culture remains. H-E3 was very likely the main “living room” of the unit where meals may have been served.

Figure 6.41: View of the food storage vessels

134

Kerebe-Sira-Tomo Settlement Complex

6.39). All the courtyard has a smooth carefully crafted floor extended all over the excavation unit. A scatter of large sherds and one complete bowl were exposed in the courtyard. An egg-nest containing the remains of five chicken eggs (fig. 6.40) was found at the foot of the wall in the corner between H-E1 and H-E3. The kitchen and storage area are located between both houses extending in the west and northwest. Nine large pots and jars were found with their burnt content (fig. 6.41). Three of the pots contained two different species of beans, cowpeas (Vigna unguiculata) and Bambara groundnuts (Voandzeia subterranea) (fig 6.42); and five of the jars were full of fonio (Digitaria iburua), an African cereal (fig. 6.43). It is the largest sample of botanical macro-remains ever found in West Africa, along with the discovery of fonio in precise and well dated archaeological context. The same thing holds for the chicken eggs. The fire-place is located next to the storage vessels. The three large laterite blocks found amidst burnt clay lumps were probably used as seats.

Figure 6.42: View of the burnt cowpeas

The vessels used for food storage vary in size and decoration, but share the same general shape with minor variations, that of large globular pot with restricted neck and more or less everted rim. They are all decorated with twisted roulette impression combined, in three cases, with grooved lines (fig. 6.44 and 6.45). Their mouth diameter varies from 25 to 30 cm, their maximum diameter from 34 to 40 cm, their height from 32 to 34 cm, and their wall thickness from 13 to 15 mm (tab. 6.6). 6.5.2.2 The northern cluster The northern cluster, located in the northeast of the excavated unit, was not reached by the fire which burnt down the central cluster. It is comprised of three sunken and partially superimposed house floors arranged along the north-south axis (fig. 6.46). H-E9 at the southern end was the first to be built. It measures 2.5 m in diameter and had an important concentration of large laterite blocks in its southwestern half, with scattered large sherds. The laterite blocks were very likely part of the bed platform installation.

Figure 6.43: View of the burnt fonio

Figure 6.44: Storage vessels

135

Archaeology of West African Mound-Clusters

Figure 6.45: Storage vessels

Two clay vessels were found on the hut floor (tab. 6.6 and fig. 6.43). One is a flat base basin decorated with twisted roulette impression. The other is the upper half of a large globular and elongated storage jar. H-E9 was probably used as a sleeping and occasional storage room. H-E5 at the center of the cluster was built on top of the northern edge of H-E9. It is clearly the largest hut of the whole excavated unit, measuring approximately 5.5 m in diameter. The western half of the feature was exposed and includes a large number of clay vessels piled up against the wall (fig. 6.47). This part of the large hut was clearly devoted to the storage of supplies. The recorded vessels are divided in two main categories: more or less globular and elongated large storage pots on the one hand and flat base large basins on the other hand.

Figure 6.46: View from the north of the northeastern hut cluster

The large globular and elongated storage vessels vary in size and shape (tab. 6.6). In general however, they tend to have restricted neck, and everted rim, and are all decorated with twisted roulette impression combined in almost all the cases with grooved lines (fig. 6.48). The flat base basins are divided into two sub-categories. Two of the specimens have rounded sides (fig. 6.49a and b) and the remaining two have straight oblique sides (fig. 6.49d and e). They are all decorated with twisted roulette impression combined in one case with grooved lines. A carefully crafted iron arm-ring was also found among the debris of H-E5 floor

Finally, a minute portion of H-E10 floor was exposed at the northern end of the cluster. It was built 20 - 25 cm above H-E5 floor and included a number of scattered sherds and large laterite blocks. Three complete to almost complete vessels were collected from H-E10 floor (tab. 6.6). One is a relatively large storage jar with restricted neck and everted rim, decorated with twisted roulette impression and grooved lines (fig. 6.50f). The remaining two are service vessels: a small spherical pot decorated with twisted roulette impression on body, zig-zag grooved lines on shoulder,

136

Kerebe-Sira-Tomo Settlement Complex

and an horizontal line of awl stabbing at the rim base (fig. 6.50a), and a large flat base bowl decorated with oblique bands of twisted roulette impression (fig. 6.50b). The small size of the exposed H-E10 floor precludes any precise assessment of its use. However, the sample of vessels points to its use for storage and food consumption which may have been combined with sleeping. 6.5.2.3 The southern cluster The southern cluster along the southern side of the excavated unit includes two contiguous house floors. H-E7 measures approximately 3 m in diameter. It has an important concentration of features, vessels, and laterite blocks in its northern half (fig. 6.36). Large pieces of 7 pots, one very large storage jar, as well as an unfired clay storage bin were recorded. The latter measures 0.80 m, with the 15 cm thick bin’s wall still preserved to a height of 30 cm. The sample of vessels ranges across a broad-use spectrum. There are two relatively large storage jars, both globular and elongated with restricted neck and everted rim, decorated with extensive twisted roulette impression (fig. 5.51a and c); two specimens of plain footed bowls and one small spherical pot decorated with twisted roulette impression (tab. 6.6 and fig. 5.51b, d, and e). H-E7 appears to have been used for the storage of food supply and domestic gear.

Figure 6.47: Concentration of storage vessels in house H-E9

H-E11 on the east flank of H-E7 was barely exposed with an arch measuring 2 m in length and 0.50 m in height (fig. 6.36). Projections from the arch indicates that the hut may have measured 3 m in diameter. A large but crushed hemispheric jug was found on the floor. It measures 8 cm in (?) mouth diameter, 27 cm in maximum diameter, 25 cm in height, and 0.8 cm thick wall. It is elaborately decorated with twisted roulette impression. The design is organized into three parts. The lower third of the vessel is decorated with oblique twisted roulette impression over-lined by a narrow horizontal band of twisted roulette impression (fig. 6.52e). The upper part of the body is decorated with four equidistant pairs of vertical/oblique bands of twisted roulette impression. And the neck base has a horizontal band of twisted roulette impression. The precise use of H-E11 cannot be assessed reasonably. 6.5.2.4 The structure of the habitation units The exact size, shape, and structure of the habitation complex exposed in occupation II are unknown. All the courtyards are carefully paved with dark-red laterite floor. The partition of the dwelling features into central, northern and southern clusters is convenient but does not necessarily reflect what happened in the past. In fact, the northern cluster is a succession of three huts and not a series of three huts. It appears to have been an additional wing to the huts from the central cluster. In that perspective, the central and northern cluster belonged to the same domestic unit. The southern cluster appears to have been built after the destruction by fire of the central cluster. A number of vessels were collected from the courtyard (fig. 6.53c, d,

Figure 6.48: Vessels from H-E5

137

Archaeology of West African Mound-Clusters

Figure 6.49: Vessels from H-E3

Figure 6.50: Vessels from H-E10, H-E11, and the courtyard

138

Kerebe-Sira-Tomo Settlement Complex

Figure 6.51: Vessels from H-E7

g, and h). They include a hemispheric steamer decorated with a horizontal band of twisted roulette impression of shoulder (fig. 6.53c), a plain small hemispheric pot (fig. 6.53d), and large globular and elongated jars with more or less restricted neck and everted rim, decorated with twisted roulette impression and grooved lines (fig. 6.53g and h). Four burials, all intrusive to occupation II deposit were also excavated. They were probably derived from occupation IV.

measuring 95 m2. It is also L- shaped (fig. 6.54) made with two partially overlapping perpendicular blocks. The westeast block measures 10.5 by 6.5 m, and the north-south one 8 by 4 m. Part of a badly damaged burial was uncovered along the western side of the excavated probe. Few long bones crumbling into powder were patiently exposed but could not be saved. Judging from the position of the arms bones, the deceased may have been buried facing northeast, laid on the left side, and probably oriented northwest-southeast.

Occupation II is dated to AD 1042-1379 (ISGS – 4344, 790+/-70 BP). Part of the habitation complex is also burnt as is the case for KST-3 occupation I, and dates exactly to the same time sequence (keeping in mind the uncertainty range of radiocarbon dates). The same range of possibilities is equally valid in this case.

The rest of the cultural remains consists almost exclusively of pottery. A number of clay vessels was found scattered over the exposed surface. Three specimens of complete bowls, set at 3.00 m from one to the next along the northsouth axis, have been recorded in the eastern half of the excavation (fig. 6.54). They measure 14 to 20 cm in mouth diameter, two with flat base and one rounded, and 10 cm high. Two are decorated with twisted roulette impression

6.5.3 Occupation III Occupation III deposit was found at 0.40 to 0.50 m. The exposure is slightly smaller than that of occupation II,

139

Archaeology of West African Mound-Clusters

Figure 6.52: Coarse stone tools from occupation II

Figure 6.53: Occupation III surface

(fig. 6.55b and e) and one with vertical bands of combrocker stamping impression (fig. 6.55c). Two globular pots were also recorded. One is decorated with extensive twisted roulette impression (fig. 6.55d) and the other, a small steamer, is also decorated with twisted roulette impression complemented by two horizontal bands of red slip (fig. 6.55h). Large storage vessels are represented by two truncated specimens. One is a globular pot with constricted neck decorated with triangular motifs and narrow bands of twisted roulette impression (fig. 6.55f). The other is the top of a globular jar with restricted neck and everted rim decorated with extensive twisted roulette impression (fig. 6.55). Finally, there are three scatters of large sherds in the north and eastern side of the probe from which a well preserved plain convexo-concave lid with a handle (fig. 6.55a) was found. Occupation III is dated AD 1244-1418 (ISGS – 4320, 660+/-70 BP) (Annex B).

its open side oriented northeast. It measures 0.40 m in length and 0.40 in width, with a 0.15 m high wall. Four burials exposed in occupation II deposit are assignable to occupation IV. Burial E2 cuts through the southern wall of the central huts cluster (fig. 6.56). The grave pit is oblong in shape, oriented west-northwest / East-southeast, and measures 1.5 m in length and 0.5 m in width. The deceased, an adult of unknown sex, was laid on the left side, head in the north, facing east, with legs and arms flexed. The femurs and tibia/ fibula measure 48 and 45 cm in length. This individual may have been 1.65 - 1.75 m tall. Burial E4, on the south flank of H-E3 (fig. 6.56) is a secondary burial containing two poorly preserved long bones, one miniature iron axe and one iron ring (fig. 6.56a and e). They were laid in a rectangular pit, 0.75 m long and 0.50 m wide oriented west-northwest / east-southeast.

6.5.4 Occupation IV Occupation IV at the top of the stratigraphic sequence is particularly loose. The exposed surface is extended over 25 meters square in a 5 by 5 m grid. Even the amount of sherds in this presumably reworked top deposit is low. The recorded cultural remains are reduced to a surprisingly well preserved horse-shoe-shaped hearth built in clay with

A minute portion of burial E6 was exposed in the southeast corner of the tested probe, overlain by a concentration of large potsherds (fig. 6.56). Its orientation is similar to that of the previous graves and the uncovered remains include poorly preserved tibia/fibula and feet bones.

140

Kerebe-Sira-Tomo Settlement Complex

Finally, burial E8, in the east of the excavation unit cut through H-E9 floor (fig. 6.56). The grave pit is also oriented west-northwest/east-southeast and measures 1.50 m in length and 0.45 m in width. The deceased, an adult individual 1.65 to 1.78 m tall, was laid on the right side, the head in the southeast, facing northeast, legs slightly flexed, and hands before the face. The base of a three-legged bowl was buried with the deceased (fig 6.57d). This kind of vessels is assumed to have been used in different rituals all

along the Inland delta and the upper Niger lands (filipowiak 1979, R. mcintosh 2002). All the recorded graves share the same general orientation. There are three primary and one secondary burials. It is very likely that KST-4 mound was used intermittently after the destruction of occupation II features. It may have been used as a convenient burial place during occupation IV. A charcoal sample collected in the deposit without any precise

Figure 6.54: Pottery from Occupation III

141

Archaeology of West African Mound-Clusters

Figure 6.55: Additional occupation III pottery

association with any specific feature dates occupation IV to AD 1195 – 1401 (ISGS – 4340, 710+/-70 BP). Even if this reading is within the acceptable uncertainty margins of radiocarbon dating techniques it does not fit in the stratigraphic coherence of the dated material obtained from habitation features.

recorded. The collected potsherds population is divided into two sub-sets. One includes all the sherds collected from the fill accumulated on occupation floors and the other is made of the sherds collected from individual features. The sample of sherds collected from the site stratigraphy amounts to 2,730 pieces. Occupation I deposit was probed with a narrow 3 by 1 m trench; the amount of collected potsherds is thus understandably small (tab. 6.7). In general, sherds frequency per occupation varies from 534 to 831. Most of the collected sherds are decorated. Twisted roulette impression is by far the predominant decoration

6.5.5 Material culture Pottery, as complete vessels and sherds, is the largest component of the material culture collected from KST-4 excavation. Iron and coarse stone artifacts were also

142

Kerebe-Sira-Tomo Settlement Complex

Figure 6.56: Occupation III iron artifacts

impression is completely absent from this large sample. The distribution of sherds in thickness classes is completely skewed towards thicker walled vessels of the 9.5-11 mm class and above belonging to the cooking and storage categories. It is only in H-E3 that thinner walled service vessels are represented.

technique used throughout the mound occupation history. It is followed by composite decoration combining twisted roulette impression with grooved lines. Comb impressions, a handful of incisions, and red ochre slip are also present. The distribution of sherds into thickness classes is bimodal in four samples out of five. Cooking and storage vessels belonging to the 9.5-11 to >14.5 mm thickness classes are largely predominant. Service wares which tend to be thinner and belong to the 5-7 and 7.5-9 mm thickness classes are consistently represented in all the samples with a particularly significant proportion in occupation III deposit at 0.40-0.60 m.

Different kinds of pot-lids were recorded in occupation II essentially. Most are broken and range from plain to extensively decorated with twisted roulette impression (fig. 6.57). Coarse stone artifacts were recorded in all occupation levels. A fragment of sandstone grindstone was retrieved from occupation I floor. It measures 22 cm in length, 20 cm in width, 4.1 cm in thickness, and 2.65 kgs in weight. In occupation II, H-E3 had an elaborate set of grinding tools, all in sandstone, including two small and one large grinder and one large but broken grindstone (fig. 6.58). H-E5 had two large fragments of sandstone grindstones (fig. 6.58). Occupation III and IV had fewer pieces; a thick grindstone fragment and a stone polisher in occupation III (fig. 6.59b and c), and a sandstone grinder in occupation IV (fig. 6.59a). A number of iron artifacts collected from burials and houses have already been presented. There is in addition one specimen of iron rod which may have been

Occupation II is a special case. Sherds belonging to each of the exposed huts were recorded separately. On the average these sherds tend to be larger in size and amount to 1,609 pieces. Their distribution is uneven, with frequency ranging from 8 to 698 (tab. 6.8). The patterns outlined for the stratigraphic sample are replicated in this case too. Most of the sherds are decorated with twisted roulette impression as the predominant decoration technique. It is followed by composite decoration which combines twisted roulette impression and grooved lines. Red-ochre slipping comes third followed by incisions. Remarkably enough, comb

143

Archaeology of West African Mound-Clusters

Table 6.7: General characteristics of the sherds population from Kerebe-Sira-Tomo 4 Stratigraphy

n

Rim

Body

Base

821 534 541 818 16

105 46 60 133 4

712 477 457 694 12

6 4 12 16 -

2,730

348

2.352

38

Decoration techniques

ND

D

TR

GL

In

Slip

Composite

0.00-0.20m 0.20-0.40m 0.40-0.60m 0.60-0.80m 0.80-1.00m 1.00-1.20m 1.20-1.40m

76 83 141 28 -

745 451 470 786 16

376 365 292 352 6

8 10 6 9 1

2 3 17 4 -

23 3 19 -

150 81 127 153 -

Total

328

2,468

1,391

34

26

45

511

Thickness classes (mm)

5-7

7.5-9

9.5-11

11.5-14

>14

0.00-0.20m 0.20-0.40m 0.40-0.60m 0.60-0.80m 0.80-1.00m 1.00-1.20m 1.20-1.40m

6 5 23 7 1

59 100 191 70 2

302 79 99 174 8

176 109 106 362 5

278 223 151 178 6

Total

42

422

662

758

836

Sample composition 0.00-0.20m 0.20-0.40m 0.40-0.60m 0.60-0.80m 0.80-1.00m 1.00-1.20m 1.20-1.40m Total

Key: D = Decorated; N = Non-decorated; U = Unknown; TR = Twisted Roulette impression; GL = Grooved line; In = Incision.

of four distinct habitation episodes, found in a 1.20 m thick cultural deposit. Roughly, an average of 50 years per occupation in this case which seems to have been more resilient than that from KST-3. The extent and intensity of occupation I is not known. Occupation II features seem to have been extended all over KST-4 mound. Huts used for sleeping, cooking, storage, and food consumption are arranged into coherent habitation units with paved courtyards. Part of occupation II houses and food supplies were destroyed by fire. The probability of a domestic accident is high in these thatched roofs round huts but the eventuality of a hostile raid is one reasonable option.

part of a spear (fig. 6.60). It was collected from H-2 in occupation II and measures 10 cm in length and 0.4 cm in section diameter. 6.5.6 Faunal remains Besides the chicken egg-nest from occupation II very few animal remains were collected from KST-4 excavation. One shell of fresh water mussel (Aspartharia sp.) was collected from the floor of H-E7 in occupation II and 1 bird phalanx and 15 splinters of undetermined bird bones from H-1. The rest of the faunal remains comes from occupation IV and consists of 3 shells of fresh water oysters (Etheria elliptica), one land-snail shell (Limicolaria sp.) from a specimen which may have died its natural death, and a series of goat bones. The latter include one molar, three radio/ulna and three tibia fragments

6.6: Kerebe-Sira-Tomo - 5 Kerebe-Sira-Tomo-5 is situated in the northernmost eminence of the mounds cluster. It measures 50 m in length northwest-southeast and 30 m in width west-east. The probe is of the standard 3 by 3 m size. The exposed stratigraphic section measures 1.40 m in thickness and consists of three major sedimentary deposits (fig. 6.61):

Summary Kerebe-Sira-Tomo-4 archaeological sequence spans from ca. AD 1000/1244 to ca. AD 1379/1453. It is comprised

144

Kerebe-Sira-Tomo Settlement Complex

Table 6.8: Kerebe-Sira-Tomo 4: Characteristics of the sherds population from Occupation II FeatureNo

n

Rim

Body

Base

Unknown

H-1 H-E3 H-E5 H-E9 H-E10 H-E11 H-E12

162 689 257 20 271 193 8

14 112 42 1 35 18 2

147 553 214 19 249 162 6

1 3 1 2 -

1 3 -

Total

1,609

224

1,350

7

4

D

ND

TR

GL

In

Slip

Composite

H-1 H-E3 H-E5 H-E9 H-E10 H-E11 H-E12

155 620 220 19 259 192 5

7 78 37 1 31 1 3

66 395 112 12 191 162 2

3 -

2 24 2 1 3

7 44 11 -

79 104 97 2 65 31 -

Total

1,470

158

940

3

32

62

378

5-7

7.5-9

9.5-11

11.5-14

>14

5 5

65 4 69

56 316 91 6 82 6 557

34 161 29 5 56 1 1 287

84 243 137 7 171 188 1 631

Sample composition

Decoration techniques

Thickness classes (mm) H-1 H-E3 H-E5 H-E9 H-E10 H-E11 H-E12 Total

Key: D = Decorated; N = Non-decorated; U = Unknown; TR = Twisted Roulette impression; GL = Grooved line; In = Incision.

6.6.2 Occupation II

3 - 0.20 – 0.00 m: A reworked brown sandy silt topsoil. 2 - 1.20 – 0.20 m: A thick light brown gray silty clay with thin horizontal red bands of house floors. 1 - 1.40 – 1.20 m: The bottom yellow clay with weathering laterite blocks. Five occupation episodes are documented in KST-5 archaeological sequence.

Occupation II consists of few large sherds, a large laterite block, and a small complete bowl found at 0.90 m on a moderately hard surface (fig. 6.63). One flexed leg, presumably part of a burial was exposed perpendicular to the probe southern limit. The other leg may have weathered away.

6.6.1 Occupation I

6.6.3 Occupation III

The earliest, occupation I was exposed at 1.00 to 1.20 m (fig. 6.62). The cultural remains consist of a red-pink burnt surface, partly destroyed and of irregular shape, found along the eastern side of the probe. It looks like the material signature of bonfires used to fire pottery. A burial was found in the opposite west side of the excavated unit. The exposed part consists of the flexed legs of an adult, oriented westeast and laid on the right side probably facing south. The burial clearly belongs to a later occupation, more likely occupation III.

Evidence for occupation III has been uncovered at 0.50 to 0.60 m (fig. 6.64). The remains of a round hut measuring 2.80 m in maximum diameter as well as a large oblongshape jar in upside-down position were exposed in the probe and extension. The hut’s floor is well made and maintained. It is surrounded by the remains of a yellowish mud-brick wall measuring 0.20 in thickness. Occupation III is dated to AD 1024 – 1291 (ISGS – 4348, 840+/-70 BP) (Annex B).

145

Archaeology of West African Mound-Clusters

Figure 6.57: Pot lids, basins and bowls bases

6.6.4 Occupation IV

extension, on the east flank of the main probe. Occupation IV is dated to AD 1022 – 1288 (ISGS – 4347, 850+/-70 BP) (Annex B).

Cultural remains from occupation IV were found on a hard and compact surface at 0.40 m below the surface (fig. 6.65). The uncovered archaeological record is located at the center of the excavated probe and consists of a scatter of large to very large sherds probably part of a broken large jar. A large storage jar in its original position was found in the

6.6.5 Occupation V Finally, the remains from occupation V were recorded at the base of the reworked topsoil, at 0.20 m (fig. 6.66). They

146

Kerebe-Sira-Tomo Settlement Complex

Figure 6.58: Grinders and grindstones from H-E3

Figure 6.59: Grindstones from H-E5

are comprised of a cluster of large sherds found along the southern side of the excavated probe in the central part. There are in addition three laterite blocks, two large ones located at two meters from each other along the northwestsoutheast axis, and a smaller one. Occupation V is dated to AD 1037 – 1298 (ISGS – 4346, 810+/-70 BP) (Annex B). Unfortunately, KST-5 occupation I and II at the bottom of the stratigraphic sequence are not dated directly. The upper part of the sequence, from occupation III upwards appears to have accumulated at a relatively fast pace. 6.6.6 Material culture The total amount of archaeological remains collected from KST-5 is relatively small. The sherds sample is made of 742 pieces unevenly distributed among the five occupation deposits. The frequency varies from 15 sherds in occupation I to 217 in occupation V (tab. 6.9). As is the case in the other probes discussed so far, decorated sherds are largely predominant in all the excavated occupations. In all the cases, twisted roulette impression is by far the dominant decoration techniques used all along the mound occupation history. It is followed by the combined twisted roulette and grooved lines impression. Thin walled vessels are well represented at the end of the probe’s occupation. Their

Figure 6.60: Iron artifacts from KST 4 and KST 5

147

Archaeology of West African Mound-Clusters

Figure 6.61: Stratigraphy of KST 5

Figure 6.62: Occupation I floor

Figure 6.63: Occupation II floor

frequency is negligible in occupation I to III. Service and cooking vessels in the 7.5-9 to 11.5-14 mm thickness class range are largely predominant. Large storage vessels of the >14.5 mm thickness are found in all the occupation deposits with however increasing frequencies from occupation I to V.

diameter, and 21 mm in wall thickness (fig. 6.67b). It is decorated with twisted roulette impression combined with series of pairs of vertical grooved lines. The other is more globular with everted rim, measuring 35 cm in mouth diameter, 46 cm in maximum diameter, 45 cm in height, and 30 cm in height of maximum diameter, with 13 mm thick wall. It is decorated with extensive oblique twisted roulette impression combined with “sail-boat” motif made of grooved lines (fig. 6.67c).

Two complete clay vessels were recorded from occupation IV and a large grindstone fragment from occupation II. The sandstone grindstone fragment is 11 cm long, 5.7 cm wide, 9 mm thick, and weighs 0.200 kgs (fig 6.67a), Both vessels are large storage jars. One is elongated and globular with everted rim, measures 40 cm in mouth diameter, 47 cm in maximum diameter, 28 cm in height of maximum

The sample of collected faunal remains consists of two fresh water oyster (Aspartharia sp.) shells. One collected from occupation I deposit and the other from occupation III.

148

Kerebe-Sira-Tomo Settlement Complex

Summary KST-5 was very likely settled during the second half of the first millennium AD up to the end of the 13th century AD. The probe did not reveal any particularly intensive occupation at any stage during the occupation history of this specific area of the Kerebe-Sira -Tomo settlement complex. One hut in one case, and portions of courtyard in all the others were exposed in the excavated probe. In all the cases however, KST-5 probe exposed different facets of domestic units shifting use of space. 6.7 Kerebe-Sira-Tomo-6 Kerebe-Sira-Tomo 6 probe is located in the northern branch of the cluster at 75 m southeast of KST-5 (fig. 6.1). The excavation revealed a 1 m thick cultural deposit accumulated on the laterite crust. The stratigraphic sequence is made of five layers as follows: 5 - 0.40 - 0.00 m: reworked brown-gray sandy silt with laterite gravels.

Figure 6.64: Occupation III floor

Figure 6.65: Occupation IV floor

149

Archaeology of West African Mound-Clusters

4 - 0.60 - 0.40 m: dark-red floors topped with light gray ash deposit. 3 - 0.90 - 0.60 m: dark-red layer of laterite floors. 2 - 1.00 - 0.90 m: brown yellow clay with laterite gravels. 1 - >1.00 m: the bottom laterite crust. Two and possibly three occupations have been documented at KST-6. 6.7.1 Occupation I Occupation I deposit was exposed at 0.60 to 0.90 m below the surface. It includes a series of connected house floors sealed by sediment from collapsed installations (fig. 6.68). The exposed features are quite obviously part a larger domestic unit. All three houses shared light brown yellow clay walls in a bee-hive pattern. The largest house floor is found in the north of the probe. It measures 3 m west-east and 2.75 m north-south and by projection may have been a 5 m in diameter round house. A small clay vessel laying on the floor was found in the east and a circular hearth, 0.75 m in diameter and filled with white ash, exposed in the west side of the floor (fig. 6.68). Two large laterite blocks were scattered in the central part of the floor. The

Figure 6.66: Occupation V floor

Table 6.9: Kerebe-Sira-Tomo 5: Characteristics of the sherds population n

W (kg)

Rim

Body

Base

0-20 20-40 40-60 60-80 80-100 100-120

217 163 142 96 105 15

10 58 4.2 3.1 3.5 1.2

28 20 24 7 20 4

178 139 122 86 84 11

10 2 1 1 1 -

Total

742

27.8

103

619

15

D

ND

TR

GL

TR+GL

0-20 20-40 40-60 60-80 80-100 100-120

208 158 140 85 97 15

9 3 6 9 7 -

106 94 89 54 62 10

8 4 1 1 2 -

46 35 24 27 23 5

Total

703

34

415

16

160

Thickness Classes (mm)

5-7

7.5-9

9.5-11

11.5-14

>14.5

0-20 20-40 40-60 60-80 80-100 100-120

13 18 4 3 4 1

43 35 31 24 29 2

81 49 48 26 26 5

55 44 47 27 35 3

24 27 16 16 11 4

Total

43

164

235

211

98

Stratigraphy (cm) Samplecomposition

Decoration techniques

Key: D = Decorated; N = Non-decorated; U = Unknown; TR = Twisted Roulette impression; GL = Grooved line; In = Incision.

150

Kerebe-Sira-Tomo Settlement Complex

Figure 6.67: Material culture from KST 5

southeastern floor included a number of large sherds with the exposed portion measuring 2 by 1.3 m. The third floor, in the southwest was barely exposed with a 1.75 m long and 0.80 m wide segment.

Three large laterite blocks and a small portable sandstone grindstone were also recorded. This room may have been used for storage and cooking. 6.7.3 Occupation III(?)

6.7.2 Occupation II

A pile of large sherds which may have been part of occupation III material culture was exposed at 0.40 m below the surface without any associated feature. These sherds may have been dropped there and buried or the accompanying installations were eroded.

Occupation II at 0.40 to 0.60 m is made of a succession of house floors overlain by ash (fig. 6.69). The first floor to be built was exposed at 0.45. Its preserved portion measuring 1.0 to 1.5 m in width and 3 m in length is found along the south side of the probe. It was truncated by the second floor which was built at 0.60 m below the surface. This later floor is extended over more than half of the excavated probe and measures 3 m west-east and 2.60 m north-south. A series of large sherds belonging to a crushed storage jar was found dumped with ash in the south of the probe. A round hearth and the remains of an unfired clay storage bin were found in the west part of the floor. A number of clay vessels of different size and shape were scattered all over the floor.

6.7.4 Material culture Pottery is largely dominant in the material culture collected from KST-6. The sample of sherds is relatively small with 338 pieces. The distribution between occupations is quite uneven (tab. 6.10). The largest amount comes from occupation II. The recorded sherds are predominantly decorated with twisted roulette impression as the most

151

Archaeology of West African Mound-Clusters

frequent technique. Sherds with composite decoration obtained through a combination of twisted roulette impression and grooved lines are largely confined to occupation II sample. Storage vessels from the >13 mm thickness class are dominant in all the occupations. Cooking vessels in the 8.5-11 and 11.5-13 mm thickness classes are well represented in the sample from occupation II. And finally, thin-walled service vessels of the 4.5-6 mm thickness class are found exclusively in occupation I in association with thick storage vessels sherds (tab. 6.10). An interesting sample of complete or nearly complete vessels was collected in occupation I and II. Occupation I sample is made of three specimens (tab. 6.11; fig 6.70): one, a large three legged bowl decorated with twisted roulette impression and red slip at the rim base (fig. 6.70a) is a service ware. The other is a large globular and elongated storage jar with everted rim, decorated with twisted roulette impression and grooved lines (fig. 6.70d). And finally, the last specimen, a cooking pot, is a globular vessel with restricted neck and everted rim decorated with twisted roulette impression and grooved line “sail-boat” motif (fig. 6.70e). As suggested by its context, occupation II sample is essentially made of cooking and storage vessels (tab. 6.11; fig. 6.70 and 6.71). They are all decorated with twisted roulette impression, combined with grooved lines with in addition awl-stabbing and red slip bands on one specimen each. The set includes: a flat base globular cooking pot with restricted neck and everted rim (fig. 6.70b); a flat base basin (fig. 6.70c), a large globular and elongated storage jar with constricted neck (fig. 6.71a); a spherical jug-like pot decorated with triangular register of twisted roulette impression framed on the sides by narrow bands of red slip and at bottom and top by horizontal and parallel lines of comb impression (fig. 6.71b). The three remaining specimens (fig. 6.71c, d, e) are all incomplete globular and elongated storage jars.

Figure 6.68: Occupation I floor

A well preserved circular convexo-concave pot-lid (fig. 6.72a) decorated with intersecting bands of comb impression, one grinder and portable grindstone both in sandstone (fig. 6.72b and c) were was also found on occupation II floor. Summary The archaeological record exposed in occupation I and II in KST-6 probe consists essentially of house floors. These dwelling features were cleaned and maintained constantly. It is therefore not surprising that there are no faunal remains in the data collected from the probe. Unfortunately, none of the exposed occupations is dated directly. The pottery is very similar to that of KST-4 occupation II suggesting that the area was settled during the first half of the second millennium AD, very likely from ca. 1000 to 1400 AD. The floor from occupation I is approximately 20 cm thick and was redone several times. The same is true for occupation

Figure 6.69: Occupation II floor

152

Kerebe-Sira-Tomo Settlement Complex

Table 6.10: KST-6: Characteristics of the sherds population Stratigraphy (cm)

n

Rim

Body

Base

Unknown

0-20 20-40 40-60 60-80 80-100

16 165 119 20 18

12 20 5 -

16 154 96 13 -

1 3 2 -

18

Total

338

37

279

6

18

D

ND

TR

IN

Slip

Composite

0-20 20-40 40-60 60-80 80-100

14 153 104 18 18

2 12 14 2 -

7 93 68 12 18

1 -

9 2 -

3 58 21 6 -

Total

307

30

198

1

11

88

Thicknessclasses(mm)

Samplecomposition

Decorationtechniques

4.5-6

6.5-8

8.5-11

11.5-13

>13

0-20 20-40 40-60 60-80 80-100

13

2 -

1 51 25 4 -

2 23 26 5 -

13 94 66 11 5

Total

13

2

81

56

189

Key: D = Decorated; N = Non-decorated; U = Unknown; TR = Twisted Roulette impression; GL = Grooved line; In = Incision.

II. The upper half of the stratigraphic sequence was built through the accumulation of sediment from the surrounding mounds by erosion agencies.

burnt clay furnace wall fragments were found scattered in the feature which was dug in the natural clay deposit (fig. 6.75a).

6.8: Kerebe-Sira-Tomo-7

Feature 2 in the south of the tested probe is circular in shape and measures approximately 4 m in diameter (fig. 6.74). It includes a work space with two attached forge furnaces. It was more eroded than feature 1 but presents essentially the same characteristics. Both forge furnaces A and B, measure 1 to 1.5 m in diameter, built along the north and east sides of the circular shelter. Number of fired furnace wall fragments and broken tuyeres were found scattered all over the exposed feature (fig. 6.74 and 6.75b).

Kerebe-Sira-Tomo-7 is a blacksmith workshop located at 150 m southeast of the main habitation mounds cluster. It measures 11 m north-south and 8 m west-east, approximately 90 m2 in surface extent (fig. 6.73) of a low mound of blacksmithing debris. A series of two small blacksmith workshops were recorded in the southsoutheastern half of the site (fig. 6.74). They are located on a strongly burnt surface and were considerably disturbed by post-depositional agencies, in this case trampling by cattle and sheep/goat herds. The shape and size of the recorded features is nonetheless accessible and presented below as accurately as possible.

6.8.2 The blacksmithing debris Samples of the scattered iron-working debris were collected for further analysis. They include iron ore, slag, tuyeres fragments, pieces of furnace wall, and coarse rock artifacts. Feature 1 sample is comprised of 12 debris distributed into 4 slag, 6 tuyere fragments, 1 block of iron ore, and 1 “brick” made from termite mound material (tab. 6.12). The recorded slag are generally small in size, measuring 6.7 to 11 cm in length, 5 to 5.8 cm in width, 3.5 to 5.8 cm in thickness, and weighing 0.200 to 1.35 kgs. The sample from feature 2 is smaller, made of 10 debris, 4 slag, 3 tuyeres fragments, and 3 coarse stone fragments. The coarse stones shattered

6.8.1 The blacksmith workshop features Feature 1 is elongated and elliptical in shape and oriented north-south. It measures 4 m in length and 2.5 m in width and includes a forge furnace at the northern end of what appears to have been a shelter. The forge furnace measures 1.5 m in diameter and had one tuyere fragment still in its original position in the south flank (fig. 6.74). Numerous

153

Archaeology of West African Mound-Clusters

Figure 6.70: Pottery from Occupation I

to smaller fragments were probably used for hammering purposes. The collected slag are 10-14.6 cm long, 3.8 to 7.5 cm wide, 0.6 to 5.2 cm thick, weighing 0.100 to 1.2 kgs.

1.2 kgs in weight. The third and last one from the general sample weighs 1.3 kgs, and measures 10.6 cm in length, and 5.4 cm in width and thickness. These pieces were collected from iron smelting workshops and used in the forge for hammering and/or sharpening the manufactured iron items.

The third sample collected from the rest of the excavation is made of 12 debris, 6 slag, 4 tuyere fragments, 1 piece of burnt termite mound, and 1 piece of iron ore. The slag pieces are within the same size and weight bracket: 6.5 to 10.6 cm in length, 5.1 to 8.1 cm in width, 2.5 to 5.4 cm in thickness, and 0.125 to 1.3 kgs in weight (tab. 6.12).

Summary The excavated blacksmith’s workshop may have been operated by two crews if both features were in use at the same time. The iron blooms obtained from iron-smelting sites - to be discussed later - were transformed into iron objects in this workshop. The material at hand suggests an interesting pattern of division of labor. Surprisingly, no organic material allowing for the direct dating of the blacksmith workshop was found. Even ordinary potsherds were not available.

There are however, three pieces of cylindrical shaped pieces resulting from molten slag which cooled down and clogged the tuyeres. There is one such piece in each of the sample. The specimen from feature 1 measures 11 cm in length, 5.8 cm in width and thickness, and weighs 1.35 kgs. That from feature 2 is 14.6 cm long, 5.6 cm wide and thick, and 154

Kerebe-Sira-Tomo Settlement Complex

Figure 6.71: Pottery from Occupation II

6.9: Kerebe-Sira-Tomo-8

2 - 0.65 - 0.20 m: slag, furnace walls and tuyeres fragments in dark brown-gray silty clay 1 - > 0.65 m: bottom sterile brown-yellow clay with gravels

Kerebe-Sira-Tomo-8 is located at approximately 250 m west of the blacksmith workshop. It is an oblong-shaped iron-smelting site measuring 10 m in length NW-SE and 5 in width SW-NE, with a single furnace. The site stratigraphic section is 0.75 m thick, with iron smelting debris constituting the bulk of the deposit (fig. 6.76).

A 45 m2 excavation probe, irregular in shape and measuring 10 m north-south and 6 to 8 m west-east, was set on the low mound with scattered by-products of iron-smelting (fig. 6.77). The furnace was not visible from the site surface. It was found at the center-east of the excavated probe associated with one pot and two laterite blocks. The furnace is circular in shape, measures 1.55 m in diameter, and was

3 - 0.20 - 0.00 m: topsoil made of termite mound and dark brown silty sand.

155

Archaeology of West African Mound-Clusters

Figure 6.73: Contour map of KST 7

6.9.1 The smelting by-products Samples of smelting debris were collected from the furnace and its vicinity. They are distributed into slag pieces, tuyeres fragments, furnace wall pieces, and coarse rock (tab. 6.13). The slag pieces are divided into two categories with 4 specimens each. There are four pieces of molten slag which clogged the tuyeres. They are cylindrical in shape, 6.5 to 9.1 cm long, 5 to 5.7 cm wide, weighing 0.400 to 0.750 kgs. These finds tend to indicate that the excavated furnace was operated with a set of four tuyeres. The remaining four slag pieces vary in size and shape. The heavier specimens, 10.5 to 12 cm long, 6.7 to 8.5 cm wide, 3.8 to 5.4 cm thick, weighing 0.600 to 1.300 kgs, may have been removed from the iron bloom. The smaller pieces are light and porous.

Figure 6.72: Pot-lid and coarse stone artifacts

filled with broken tuyeres, wall‘s fragments and slag. The preserved portion is 0.85 m deep. The furnace wall was built with stretched clay lumps. Its mouth, 0.20 m wide, was located in its southeast flank (fig. 6.78 and 6.79).

The collected tuyeres fragments vary in length and width as a result of breakage hazard. The wall thickness is more or less standardized, ranging from 1.9 to 3 cm. The two samples of furnace wall fragments are 2.8 to 4.7 cm thick “bricks” and the 5 pieces of coarse stone were used as hammer-stones.

Interestingly enough, as is the case for TST-1, the furnace from KST-8 was also a dual-purpose installation, used for iron smelting and tuyeres firing. The upper part with the location of the active tuyeres was not preserved. 14 un-fired tuyeres made with light yellow clay were found in up-right position at the bottom in the lower firing chamber (fig. 6.78 and 6.79). They were positioned on top of the bottom slag. An unknown event precluded the firing of the blow-pipes.

Two broken pots and a handful of sherds, all decorated with twisted roulette impression, were found among the smelting debris (fig. 6.80). Both pots are globular and elongated with restricted neck and everted rim. Finally, five land-snail shells (Limicolaria sp.) were found in top layer of the site’s

156

Kerebe-Sira-Tomo Settlement Complex

Table 6.11: KST-6: Vessels mensurations Occupation I

II

1 2 3 1 2 3 4 5 6

MD (cm)

MaD (cm)

H (cm)

Th (mm)

18 25 40 15 24 25 25 30 7

18 37 55 22 24 35 35 40 28

9 20 65 18 11 ?15 ?26 ?32 25

7 10 20 8 8 10 10 15 9

Decoration

Nature

TR + Slip TR + GL TR + GL TR + GL TR + GL TR + GL + AS TR + GL TR + GL TR + CI + Slip

Large three legged bowl Globular pot Large elongated and globular pot Globular flat base pot Flat base basin Storage jar Storage jar Storage jar Spherical storage pot

Key: MD = Mouth diameter; MaD = Maximum diameter; H = Height; Th = Thickness; TR = Twisted roulette impression; GL = Grooved line; AS = Awl stabbing

Figure 6.74: Plan of KST 7 blacksmith workshops

157

Archaeology of West African Mound-Clusters

Figure 6.75: The furnaces sections

sequence. They could have died of their natural death on the site but their use as “snacks” cannot be totally ruled out.

measures 2.20 m in diameter at top. The preserved furnace portion shaped like an upside down bell is 1.80 m deep with a 0.80 m diameter base (fig. 6.85). The wall is particularly thick in this case, measuring 0.20 m made of stretched and smoothed clay “bricks”.

Summary Kerebe-Sira-Tomo-8 was exclusively used for iron-smelting purposes. It was operated as a workshop by people from the main mound complex. The amount of debris accumulated in a 0.45 m thick layer on the site suggests that the furnace was in use for a certain time. This could have been on a seasonal basis, for a few weeks each year. The workshop is unfortunately undated and the recorded pottery is too common to point to a specific time segment.

A single slag piece weighing 0.800 kgs was found in the furnace. The large number of large tuyeres fragments is intriguing. The sampled tuyeres fragments measure 5.7 to 22 cm in length, 5 to 10.2 cm in width, 2.2 to 3.3 cm in thickness, and 0.100 to 1.3 kgs in weight (tab. 6.14) This installation appears to have been devoted to the production of blow-pipes. The firing chamber was filled with un-used and broken tuyeres. Iron smelting may have taken place infrequently as suggested by the small amount of iron slag and the total absence of a bottom slag.

6.10: Kerebe-Sira-Tomo-9 Kerebe-Sira-Tomo-9 is located at approximately 150 m southwest of the KST-7 blacksmith workshop (fig. 6.1). It is a shallow scatter of smelting debris measuring 6 to 6.5 m in diameter, with the remains of a furnace in the center-east of the sampled area (fig. 6.81 and 6.82).

Summary Kerebe-Sira-Tomo-9 is an additional interesting case of the patterns of social division of labor implemented by the inhabitants of Kerebe-Sira-Tomo settlement complex. The workshop seems to have been devoted to the production of tuyeres to be used in other places for iron-smelting and blacksmithing purposes.

The top of the preserved portion of the furnace was visible from the surface, marked by a dense scatter of fired walls fragments. The exposed feature was filled with numerous large tuyeres and wall fragments (fig. 6.82 - 6.84) and

158

Kerebe-Sira-Tomo Settlement Complex

Table 6.12: KST-7: Samples of iron production by-products Nature

Length (cm)

Width (cm)

Thickness (cm)

Weight (kgs)

7.2 11 7.8 6.7

5 5.8 5.8 5.2

3.8 5.8 3.8 3.5

0.250 1.350 0.400 0.200

11 10 11 14.6

6.0 3.8 7.5 5.6

0.6 2.5 5.2 3.3

0,150 0.100 1.000 1.200

7.5 7.9 6.5 10 10.8

5.2 6.9 5.9 3.7 8.1

3.5 4.3 3.6 2.5 4.,5

0.225 0.400 0.250 0.125 1.000

9.8 11.3 11.6 12.1 9.6 9.7

7.4 9.7 9.2 9.3 7.8 7.3

2.4 2.3 2.1 2.5 2.1 2.1

0.300 1.800 1.800 1.800 1.800 1.800

9.2 9.0 6.8

7.0 4.9 5.3

2.1 2.7 2.7

0.600 0.600 0.600

7.0 9.3 6.6 9.2 10.6

4.0 6.9 5.3 4.2 5.4

2.5 2.2 2.1 2.4 5.4

0.100 0.300 0.150 0.150 1.300

5.6 9.4 6.6

6.1 7.3 5.2

4.0 2.7 1.4

0.375 0.800 0.100

10 17 10.5 16

7.3 11.5 7.0 10.3

6.0 7.2 6.2 6.0

0.800 0.850 0.800 0.800

Slag Feature 1 1 2 3 4 Feature 2 5 6 7 8 Others 9 10 11 12 13 Tuyere Feature 1 14 15 16 17 18 19 Feature 2 20 21 22 Others 23 24 25 26 27 Coarse rock fragment 28 29 30 Others 31 - Iron ore 32 - brick 33 - Iron ore 34 - Termite brick

159

Archaeology of West African Mound-Clusters

Figure 6.76: Stratigraphy of KST 8

Figure 6.77: Plan of the excavation

160

Kerebe-Sira-Tomo Settlement Complex

6.11: Kerebe-Sira-Tomo-10 Kerebe-Sira-Tomo 10 is located at approximately 100 m southwest of KST-9. It is a small sub-circular site measuring 4-5 m in diameter. An excavation probe measuring 3 by 2.5 m was sunk around the remains of the furnace, visible from the surface (fig. 6.86). A portable sandstone grindstone was found at less than 1 m NW of the furnace. The latter was filled with broken wall fragments, tuyeres, and some slag (fig. 6.87). It is a U-shaped feature, measuring 1.20 m in diameter at top, 0.70 m at bottom, and 0.90 m in depth. The furnace was made of 0.10 m thick wall and none of the above ground part of the installation was preserved. 38 un-fired blowpipes were found at the bottom of the furnace, set for a firing event which never took place (fig. 6.88). Depending on how one reads the evidence, they are in upright position, tightly packed and arranged in rows (fig. 6.89). 6.11.1 The smelting debris The sampled smelting debris, all collected inside the furnace, are comprised of a handful of slag, tuyeres, and wall fragments (tab. 6.15). The three slag pieces are 7.5 to 9.2 cm long, 5 to 6.7 cm wide, and 4.8 to 6.4 cm thick, and weigh 0.300 to 0.600 kgs. They are all light and porous and do not seem to have resulted from an in-situ iron-smelting episode. The recorded wall fragment is a 11.4 by 9.5 cm brick, 3 cm thick. On the average, three of these bricks were “layered” to built the furnace wall. Finally, 6 fired tuyeres fragments were collected. Their size and shape resulted from the hazard of preservation. The thickness is nonetheless significant. In the case at hand, it ranges from 1.6 to 3.7 cm and points to a blow-pipes manufacturing standard which allow to withstand thermal shock.

Figure 6.78: Plan and section of the furnace

Summary Kerebe-Sira-Tomo-10 was also a tuyere firing workshop. Iron-smelting does not seem to have ever been practiced in this site. The dozens of un-fired tuyeres left at the furnace bottom are intriguing. They were left more or less intact but the rest of the furnace was filled with broken fired tuyeres, wall fragments, and intrusive slag. As is the case at KST-8 and KST-9, it appears that some groups of the KerebeSira-Tomo settlement complex may have specialized in the production of tuyeres and supplied the iron-smelters and the blacksmiths. 6.12: Kerebe-Sira-Tomo-11 Kerebe-Sira-Tomo 11 is located in the southwest of the settlement complex at less than 100 m from the previous KST-10 workshop (fig. 6.1). It is a 8.5 m in diameter scatter of smelting debris distributed around a collapsed central furnace. It does not differ significantly from the other ironworking workshops. A 8 by 7 m excavation unit was opened to be narrowed later to a 3 by 2.5 m probe in the furnace area after documenting the absence of buried debris. The

Figure 6.79: View of KST 8 furnace

161

Archaeology of West African Mound-Clusters

Figure 6.80: Pottery from KST 8 Table 6.13: KST-8: Samples of iron production by-products Nature

Length (cm)

Width (cm)

Thickness (cm)

Weight (kgs)

8.8 8.7 6.5 9.1 10.5 8.0 12 8.5

5.4 5.7 5.0 5.5 8.5 4.3 6.7 7.5

5.4 5.7 5.0 5.5 5.4 3.3 3.8 2.5

0.750 0.400 0.400 0.625 0.600 0.300 1,300 0.350

Slag 1 (tuyere) 2 (tuyere) 3 (tuyere) 4 (tuyere) 5 6 7 8

162

Kerebe-Sira-Tomo Settlement Complex

Nature

Length (cm)

Width (cm)

Thickness (cm)

Weight (kgs)

9.6 6.2 9.5 10.2 7.5 8.7 9.8 11.2 8.5 9.4

6.5 4.8 7.6 7.0 5.1 7.0 5.8 6.5 7.5 6.0

2.1 2.1 3.0 2.3 2.0 1.9 2.2 2.9 2.5 2.4

0.350 0.100 0.500 0.500 0.125 0.200 0.800 1,100 0.350 0.450

7.4 6.1 8.0 7.5 10.6 8.9 11.8

4.0 4.0 7.0 6.0 5.7 6.7 5.7

4.7 2.8 3.2 5.0 3.0 4.2 2.3

0.200 0.100 0.325 0.500 0.400 0.500 0.500

Tuyere 9 10 11 12 13 14 15 16 17 18 Others 19 F. wall 20 F. wall 21 Rock 22 Rock 23 Rock 24 Rock 25 Rock

Figure 6.81: Plan of KST 9

163

Archaeology of West African Mound-Clusters

Figure 6.82: View of the furnace fill top

Figure 6.85: Map of KST 10 furnace

Figure 6.83: View of the accumulation of tuyeres

Figure 6.86: View of the top of the furnace

Figure 6.84: View of the furnace bottom

164

Kerebe-Sira-Tomo Settlement Complex

Table 6.14: KST-9: Samples of iron production by-products Nature

Length (cm)

Width (cm)

Thickness (cm)

Weight (cm)

7.7

6.7

5.5

0.800

12.5 7.1 5.7 6.1 6.5 9.2 22

7.0 6.8 5.4 6.1 5.0 6.0 10.2

2.5 2.7 3.3 2.6 2.6 2.2 2.6

0.325 0.300 0.225 0.200 0.100 0.300 1.300

Slag 1 Tuyere 2 3 4 5 6 7 8

furnace, visible from the surface, was filled with smelting debris, essentially large pieces of blow-pipes. It measures 1.2 m in diameter and was 0.70 m deep for the preserved underground part. 6.12.1 The smelting debris All the sampled smelting debris were collected from the furnace. The material scattered on the site surface was excessively fragmented by centuries of trampling by livestock herds. Some 50 tuyeres specimens were found in the furnace fill. Three well preserved specimens collected at the bottom measure 15.5 to 29 cm in length, 10 to 11.5 cm in diameter, 2.6 to 3.4 cm in wall thickness, weighing 0.570 to 5.5 kgs (tab. 6.16). A dense and heavy 8.8 kgs bottom slag, circular in shape, and measuring 0.80 m in diameter and 5 cm in thickness sealed the bottom of the furnace. The sample of furnace’s wall was rectangular in shape, 16 by 15 cm, 6 cm thick, weighing 1.7 kgs. The recorded debris point to the use of this installation for iron-smelting purpose (fig. 6.90). The shallow scatter of fragmented slag, tuyeres, and other furnace walls fragments suggests an intermittent short-term use.

Figure 6.87: View of the furnace bottom

A few sherds and the upper half of an elongated and globular pot are the only material culture element found on the site. They are all decorated with twisted roulette impression (fig. 6.90). Two of the sherds are thin and may have belonged to service vessels of the bowls category. The third sherd, from a large storage vessel, is more than 2 cm thick. And finally, the partly preserved pot may have been a beverage and/or food container. Summary Kerebe-Sira-Tomo 11 was clearly an iron-smelting workshop. It is an additional indication of the patterns of social division of labor developed by the inhabitants of the Kerebe-Sira-Tomo settlement complex. It is unfortunately not possible to trace the development in time of this pattern.

Figure 6.88: Plan of the bottom of furnace 10

165

Archaeology of West African Mound-Clusters

Figure 6.89: Pottery from KST-11

Table 6.15: KST-10: Samples of iron production by-products Nature

Length (cm)

Width (cm)

Thickness (cm)

Weight (cm)

7.5 8.3 9.2

5.0 5.5 6.7

5.0 4.8 6.4

0.300 0.400 0.600

9.2 6.5 10 4.7 9.7

7.4 5.5 8.0 3.2 8.3

3.2 3.7 2.5 1.6 2.6

0.500 0.200 0.350 0.100 0.350

11.4

9.5

3.0

0.600

Slag 1 2 3 Tuyere 4 5 6 7 8 Others 9 Furnace wall

6.13: The dynamics of Kerebe-Sira-Tomo Settlement complex

6.13.1 Phase I (ca. 350 BC - 150 AD) The earliest evidence for settlement at the Kerebe-SiraTomo settlement complex is dated to 350 BC - 150 AD documented in probe KST-1A in the west-central part of the complex and very likely in KST-2, the quarry site. The domestic units associated with the pottery firing or karité oil production installation of KST-1A has still to be found.

The probes excavated all over the Kerebe-Sira-Tomo settlement complex provide an insight into the evolution of the whole complex during its 2000 years of existence. It goes without saying that the total size of the settled area at any stage in the history of the settlement cannot be known as precisely as one may have wished. The extent of the settled space is therefore inferential. The settlement complex seems to have developed through four main phases.

6.13.2 Phase II (ca. 250 - 550 AD) The settlement grew in two directions during phase II, with domestic units built in KST-1B and KST-3 areas. The

166

Kerebe-Sira-Tomo Settlement Complex

Figure 6.90: Kerebe-Sira-Tomo in phase I and II

Table 6.16: KST-11: Samples of iron production by-products Nature

Length (cm)

Width (cm)

Thickness (cm)

Weight (kgs)

80

80

5.0

8.800

29 20 15.5

11.5 10.5 10

3.4 9.6 2.8

5.500 1.500 0.570

16

15

6.0

1.700

Slag 1 Tuyere 2 3 4 Others 5 Furnace wall

mound was of an irregular potato-shape oriented SW-NE (fig. 6.91). It is assumed that KST-2, the quarry site, was relied upon for the supply of iron ore and construction material during the whole existence of the settlement complex. The 500 years settlement hiatus recorded in KST1B appears to apply to the whole site. Settlement resumed at the beginning of the second millennium AD.

occupation I and KST-4 occupation II. Both habitation units were located along the southeastern flank of the complex and may have been the target of an act of warfare. It is very likely during this phase of relative fast growth that special purpose workshops devoted to iron-smelting, blow-pipes making, and blacksmithing were founded a few hundred meters south of the main habitation mounds cluster.

6.13.3 Phase III (ca. 1000 - 1250 AD)

6.13.4 Phase IV (ca. 1250 - 1450 AD)

Phase III witnessed the accelerated growth of KerebeSira-Tomo settlement complex. Habitation units were then spread all over the site, in all the probes, KST-IA and 1B, KST-2 (quarry site), KST-3, KST-4 (occupation I and II), KST-5, and KST-6 (fig. 6.92). It is during this phase that fires destroyed the domestic features from KST-3

The settled area shrunk considerably during phase IV. An especially loose occupation took place in KST-4 occupation III and IV. No dwelling features have been recorded suggesting that the complex may have been abandoned with however the small KST-4 mound still visited intermittently. That small mound was used for burial during occupation IV.

167

Archaeology of West African Mound-Clusters

Figure 6.92: Kerebe-Sira-Tomo in phase IV

Summary Kerebe-Sira-Tomo settlement complex presents another facet of the dynamics of mounds cluster. The habitation mounds formed a coherent 15 hectares site while craft activities, articulated around iron production, took place at some distance in the south. It is not known if all the installations related to iron-working were operated in parallel by different task-groups or sequentially during the entire occupation history of the settlement complex.

Figure 6.91: Kerebe-Sira-Tomo in phase III

168

Chapter 7 Human Skeletal Remains: Taphonomy, Demography, and Palaeopathology

Introduction

completeness. The number of individuals present was determined and each skeleton (or portion of a complete skeleton) was assigned a number in sequence (from 1 to 20). Age and sex determinations were made when possible. All possible cranial and post-cranial metric observations were made. Dental metric data were recorded from observable dentitions, to be used in geographic and diachronic comparative studies of dental morphology. Dental morphological data were also recorded following the Arizona State University Dental Anthropology System protocol (TURNER et al. 1991). Dental morphological variation within the Mouhoun Bend sample was statistically compared to data from fifteen other sub-Saharan African population dental samples. Results were presented at the 73rd Annual Meetings of the American Association of Physical Anthropologists (MAES et al. 2004). All elements were examined for evidence of skeletal pathology, and descriptive observations were recorded. Finally, two measures of preservation were recorded for each individual, including descriptions of apparent post-depositional changes (taphonomic processes) affecting the bones (see Skeleton Summaries, Annex A).

Skeletal materials from the Mouhoun Bend Archaeological Project 1999 and 2000 field seasons were analyzed by Maes at Dr. Philip Walker’s laboratory in the Department of Anthropology at the University of California, Santa Barbara. Osteological data were used to reconstruct the health status and behaviors of people buried in various settlements of the area, taking into account the larger project goals which aim to document changing subsistence patterns (from hunting and gathering to food production), the onset of iron production, and long-term cultural change including craft specialization (Holl 2009, Holl and Lassina, 2003). Most of the human remains were excavated from Tora Sira Tomo (TST) settlement, with in addition a smaller sample from Kerebe-Sira-Tomo (KST), and a single individual from Diekono. In general, the remains from TST and KST are poorly preserved, while the single burial from Diekono is relatively well preserved. A total of twenty individuals are represented in the studied human skeletal sample. Unfortunately and owing to poor preservation, only 20% of the individuals are represented by a significant portion of the complete skeleton. This fact severely hinders analysis, especially with regard to the observation and description of skeletal pathological conditions. Nevertheless, an exciting observation was made concerning possible evidence of attempted trephination from TST (Skeleton 3), which would constitute the first documented evidence of surgery from an archaeological context in sub-Saharan West Africa. This finding was presented at the 14th Biennial European Members Meeting of the Paleopathology Association in 2002 (GJERDRUM et al. 2002).

7.1 Materials and Methods A total of twenty individuals were analyzed: eleven individuals from TST, eight from KST, and one from Diekono. The individuals from TST and KST were located at different levels of occupation within several excavation probes from both settlement-complexes (holl 2009; holl & kote 2000). The available time sequence represented by these occupations ranges from 7-500 BC to AD 1600. Skeletal remains were initially transferred from their original packaging and placed in sturdy sealable bags with clear labeling. Many bones were still stuck in sediment blocks and required separation and cleaning. To prepare remains for analysis, the bones were cleaned with water, soft brushes, and small bamboo picks. Using water was necessary for the removal of adhering sediments, which were often very dry and cement-like. When bones appeared too fragile to withstand washing, observations were made without complete washing.

Poor preservation of the skeletal material is largely attributed to the destructive effects of termites. This specific taphonomic process is not well documented in the field of osteo-archaeology. Therefore, the Mouhoun Bend skeletons present useful comparative material for future descriptions of tropical taphonomic processes. This preliminary analysis of the human remains from the Mouhoun Bend Archaeological Project followed protocols outlined in Buikstra and Ubelaker (1994). An extensive inventory was made of all skeletal and dental remains present for each individual. Inventory data included the identification of bone or bone type, side, segment, and

Age was determined using the multiple criteria of dental development and eruption, epiphyseal union, pubic symphysis morphology, auricular surface morphology,

169

Archaeology of West African Mound-Clusters

et al. 1991). Following recordation, trait frequencies were determined for the study sample and compared with the 15 samples of known cultural affiliation, employing C.A.B. Smith’s Mean Measure of Divergence (MMD) statistic, using the Freeman and Tukey angular transformation. This multivariate technique provides a quantitative estimate of biological divergence between samples, based on the degree of phenetic similarity of dental traits (irish 1997). Results of this study were presented at the 73rd Annual Meetings of the American Association of Physical Anthropologists in 2004 (maes et al. 2004).

cranial suture closure, and dental attrition (as a secondary criterion used only for relative age determinations). For sub-adults, age ranges were determined within two year increments, while adults were separated into the following six categories: Young Adult (18-34 years); Young-Middle Adult (18-49 years); Middle Adult (35-49 years); MiddleOld Adult (35-50+ years); Old Adult (50+ years); and Adult-Other (18+ years). Sex determination was also based on multiple criteria as outlined in Buikstra and Ubelaker (1994). The most heavily weighted characteristics were those comprising the Phenice method: ventral arc (of the pubis), subpubic concavity, and ischiopubic ramus ridge (phenice 1969). Unfortunately, observation of these characteristics was severely limited by poor preservation.

Cranial and post-cranial metric data were compiled using digital sliding calipers, spreading calipers, and an osteometric board. In one case this data was used in the assessment of stature (Skeleton 3), following formulae outlined in Krogman and Isçan (1986). These data will also be used in future comparative studies and studies of interesting skeletal morphological variation exhibited among the Mouhoun Bend skeletons.

Other sexually dimorphic characteristics include greater sciatic notch, preauricular sulcus, nuchal crest, mastoid process, supra-orbital ridge (glabella), supra-orbital margin, mental eminence, and overall body size and robusticity. These characteristics aided in confirmation of sex determinations.

Data collection for dental pathology was based on the identification of patterns of several pathological conditions: carious lesions, abscesses, antemortem tooth loss, pulp exposure, calculus deposition, alveolar resorption, artificial abrasion, crowding, enamel hypoplasias, trauma, and congenital agenesis and anomalies. In this study, carious lesions were recorded by location and size of each lesion, and hypoplastic lesions were recorded by tooth affected and distance of the lesion from the cemento-enamel junction. Other pathological conditions were documented by presence and degree of expression. Dental attrition was recorded following the Scott (1979) method, which assigns a stage of wear (0-10) for each tooth. Molars are divided into quadrants and each quadrant is assigned a stage of wear. This serves to identify a plane of wear, which was especially relevant in this case because the current assemblage exhibits some abnormal patterns of attrition as well as evidence of occupational use of the teeth.

Individuals younger than 16 years of age were not assessed for sex, as sexually dimorphic characteristics are not manifested in the skeletal remains of pre-pubescent individuals. Also, the remains of some adults did not include any elements exhibiting clear characteristics for sex determination. Thus, several individuals in this skeletal sample of twenty were recorded as “indeterminate” sex. All other adults were designated as male, female, probable male, or probable female. Dental morphometric data were compiled following Buikstra and Ubelaker (1994). Buccolingual diameter, maximum mesiodistal diameter, and crown height dimensions were observed from all teeth using digital sliding calipers (0.01 mm). Diameter dimensions will be used in future comparative studies, while crown height dimensions give an idea of the remaining amount of enamel on each tooth after a lifetime of attrition.

The identification of skeletal pathological conditions was based on the location and degree of bone loss and bone deposition and the pattern of elements affected. Pathological conditions were grouped by etiology, including Joint Disease, Non-specific Stress Indicators, Non-specific Infections, Trauma, Congenital Conditions, Circulatory Disorders, Metabolic Disorders, Neoplastic and Other Conditions (Roberts & Manchester 1995). Cultural modification of the skeleton was also observed in one case of trephination. Observation of pathological conditions, including one unusual condition, was hindered by poor preservation and the “pseudo-pathological” effects of termite activity, which was generally ubiquitous among the current assemblage.

Twelve individuals from the Mouhoun Bend excavations were available for the study of dental morphology. These twelve include two juveniles with mixed deciduous and permanent dentitions, two adolescents with permanent dentitions, and nine adults with permanent dentitions. Only permanent teeth were observed. Unfortunately, substantial dental attrition and poor preservation of osseous material prevented the observation of much potential data. Thus, the sub-samples of most individual traits in the study sample are less than twelve. The 15 comparative samples (796 total dentitions), used in previous studies of African population affinity studies (irish 1993, 1997), are a representative cross section of sub-Saharan African cultural groups. The occurrences of 36 dental morphological traits were recorded for each available individual in the Mouhoun Bend sample, following the ASU Dental Anthropology System (turner

Two measures of preservation, completeness and condition, were recorded for each burial. Completeness (tab. 7.1) is an index quantified by the number of long bones present for each skeleton divided by fourteen (representing both

170

Human Skeletal Remains

Table 7.1: Taphonomic condictions of the Mouhoun Bend skeletons sample Variable

n

Poor

Fair

Good

Excellent

Completness State of preservation

20 20

16 13

2 4

2 3

0 0

Table 7.2: Demography of the Mouhoun Bend skeletons sample 1 - Sex

2 - Age

3 - Individuals with observable dentition

n

Male

Probable male

Female

Probable female

Indeterminate

20

2

4

3

1

10

20

5 – 11 y

12 – 17 y

18 – 34 y

18 – 50 y

35 – 49 y

35 – 50 y

50 y+

18y+ (adult others)

2

2

1

2

2

2

1

8

n 13

Adolescents Adults with Juveniles with permanent with mixed permanent dentitions dentitions dentitions 2

2

9

antimeres of the seven long bones: clavicle, humerus, ulna, radius, femur, tibia, and fibula; walker et al. 1988). These quantitative indices are converted into qualitative categories (Poor, Fair, Good, Excellent), and provide a general idea of how much of the individual is preserved.

probable female, and ten individuals were of indeterminate sex (tab. 7.2). The large number of individuals with either probable or indeterminate sex identifications is attributed to poor preservation and, specifically, lack of sexually dimorphic skeletal elements for the majority of individuals.

The assessment of condition, which also ranges from poor to excellent, is a more subjective measure. It takes into account the amount of cortical bone preserved, evidence of weathering and diagenesis, the adherence of cementlike sediment, and other taphonomic processes that obscure bone surfaces. Skeletons listed as “poor” condition were friable, severely fragmented, and heavily obscured by adhering sediment and/or termite destruction. In contrast, skeletons listed as “good” condition were sturdy, less likely to exfoliate, and possessed a substantial portion of clearlyobserved bone surfaces. No skeletons were preserved in “excellent” condition (tab. 7.1).

In terms of age distribution, this assemblage of twenty individuals contained sixteen adults and four sub-adults. Within these four sub-adults, two were aged 5-11 years and two were aged 12-17 years. The breakdown of ages within the adult group was one Young Adult, two Young-Middle Adults, two Middle Adults, two Middle-Old Adults, one Old Adult, and eight Adult-Other individuals (aged greater than 18 years) (tab. 7.2). Demographic statistics cannot be assessed for this assemblage, as such analyses require large populations, and the Mouhoun Bend skeletons come from separate settlements spanning several centuries.

Poor preservation is directly related to lack of skeletal data, particularly in the realm of pathology. For instance, because vertebral and diarthrodial articular surfaces were so severely under-represented due to poor preservation, almost nothing can be said about frequencies and descriptions of degenerative joint disease. Degenerative joint disease in a skeletal assemblage is particularly useful in the reconstruction of chronic occupational stress (i.e., activity patterns); thus, any description of the life-ways of the people occupying the site under study will lack this aspect.

7.3 Dental Pathology Dental pathology can elucidate patterns of diet, mechanical demands placed upon the teeth, and many processing activities associated with cultural behavior. Dental pathological conditions noted among the Mouhoun Bend skeletons include carious lesions, calculus deposition, abscesses, ante-mortem tooth loss, alveolar resorption, dental crowding, and wear due to processing activities. Attrition, the normal wearing away of tooth crowns due to chronic mastication of food, was observed in both permanent and deciduous dentitions. Abnormal wear patterns (e.g., those that show bilateral asymmetry, difference between maxillary and mandibular teeth, and

7.2 Demography For the complete skeletal sample, two individuals were male, three were female, four were probable male, one was

171

Archaeology of West African Mound-Clusters

Table 7.3: Pathologies in the Mouhoun Bend skeletons sample Dental Pathologies

n

Percentage

Calculus deposition Caries Antemortem tooth loss Abscesses Alveolar resorption Moderate/Severe attrition Artificial wear

7 4 3 2 3 9 5

78% 44% 33% 22% 33% 100% 56%

Skeletal Pathologies

n

Percentage

Spinal joint disease Degenerative joint disease Trauma Non-specific infection Neoplastic condition Non specific stress indicator Other, Trepanation

2 3 3 1 1 1 1

11% 16% 16% 5% 5% 5% 5%

uneven molar wear) were also observed. Such patterns may be attributed to problematic conditions such as malocclusion (e.g., overbite, as observed for Skeleton 3) and dental crowding (as observed for Skeleton 11), as well as to the use of the teeth for occupational processing (as observed for Skeleton 15).

10 years). Skeleton 11 possesses weaker defects on the upper second molars (not present on lower second molars), suggesting that a less severe insult occurred around age 4. It is impossible to say what specific insults occurred, but most likely they involved malnutrition and/or infectious etiologies. Interestingly, it is rare to observe hypoplastic lesions on second and third molars, as most occur on anterior teeth such as upper central incisors and lower canines (Aufderheide and Rodríguez-Martin 1998). These teeth develop at an earlier age (during the first year of life) when an individual is most susceptible to environmental stressors, especially poor nutrition. Thus, it may be that these two individuals suffered acute infectious insults rather than nutritional deficiencies.

Thirteen individuals were available for the study of dental pathology, while seven skeletons lacked preserved teeth. These thirteen individuals include two juveniles (Skeletons 4 and 18) with mixed deciduous and permanent dentitions, two adolescents (1 and 19) with permanent dentitions, and nine adults (2, 3, 5, 7, 11, 12, 14, 15, and 20) with permanent dentitions (Tab.7.2). Of all the individuals with observable dentitions, five are male/probable male, three are female/probable female, and five are of indeterminate sex.

Dental crowding was observed among five individuals (2, 3, 5, 11, and 15). In each case, crowding was evident only in the mandibular dentition, with the incisors most commonly affected, though canines and molars sometimes exhibited rotation. Dental crowding probably occurs most often as a result of a decrease in the size of the bony alveolus which houses the teeth. This condition has been observed among many Holocene skeletal populations that transitioned from hunting and gathering patterns of food procurement to food production (larsen 1997).

The two juveniles (4 and 18) did not exhibit any observable dental pathological conditions beyond slight to moderate attrition of some deciduous teeth. As most pathological conditions are degenerative, sub-adults in skeletal samples often did not live long enough to develop such conditions. Among the adolescents and adults (n = 9), the most common dental pathologies were calculus deposition (seven individuals affected), carious lesions (four individuals affected), and ante-mortem tooth loss (three individuals affected). In addition, abscesses were observed at low frequencies (two individuals affected). The degenerative condition of alveolar resorption was identified by bone loss in the alveolus of the mandible in conjunction with the three observed cases of ante-mortem tooth loss (tab. 7.3).

Five individuals (2, 3, 5, 7, and 15) show possible patterns of artificial dental wear, which may be attributed to the use of the teeth for occupational processing. Skeleton 2 exhibits bilateral asymmetry of wear between lower first molars, with the right antimere having much greater wear—26 versus 34 Scott (1979) wear scores. Skeleton 3’s overbite explains much of his observed abnormal wear patterns; however, a few conditions probably have other etiologies. RI1 has a concave groove on the distal corner of the crown, while LI1 has the same grooves on both mesial and distal corners. Abnormal grooves are present on LC1, LP1, and RM2. Degrees of molar wear have no clear gradient from 1st to 2nd to 3rd molars. Finally, bilateral asymmetry of wear is observed between all mandibular antimeres from canines through third molars, with heavier wear invariably present on the right teeth. Skeleton 5 exhibits bilateral asymmetry of wear between the lower canines, with notable labial wear facets on both. Skeletons 7 and 15 show patterns that may easily be attributed to occupational processing. Concave grooves are present on the mesial occlusal corners of RI1, LI1, and RI1 (LI1 is unobservable) for Skeleton 7. The

In addition to these commonly occurring pathologies, a few other conditions were also noted. Two individuals (7 and 11) possess linear hypoplastic lesions. Dental hypoplasias are defects in the tooth enamel caused by metabolic, systemic insults during the development of the tooth crown (Aufderheide and Rodríguez-Martin 1998). Since dental development progresses in a well-known sequence, we can estimate when an insult occurred by observing the locations and patterns of hypoplastic defects. Skeleton 7 possesses strong defects on all third molars, revealing that a significant insult occurred around age 10 (caution must be taken when estimating the age because the development of third molars can vary significantly among individuals, sometimes beginning when the individual is well older than

172

Human Skeletal Remains

same grooves are present on the mesial occlusal corners of RI1 and LI1 (RI1 and LI1 are unobservable) for Skeleton 15. Thus, one can visualize a roughly circular area of abnormal wear when the anterior teeth of these individuals were centrically occlused. In addition, Skeleton 15 exhibits uneven wear patterns (by quadrant) on upper and lower right first and second molars. Both lower molars have heavier wear on the disto-buccal quadrant, while both upper molars have lighter wear on the disto-buccal quadrant.

9-11 year-old exhibiting what appear to be two perimortem depressed cranial fractures on the right parietal. The pattern suggests a sharp object struck the cranium from behind. Unfortunately, it is not clear if the marks were made or not by excavators’ tools, a “taphonomic process” that often affects archaeological skeletal remains. Further analysis and/or field notes from the excavation may elucidate the true etiology of the marks. Skeleton 2 exhibits multiple (four) well-healed depressed cranial fractures on the frontal, left parietal, and occipital bones. Diameters of the fractures range from 10.15mm to 21.34mm. Complete healing, like that observed in this case, is indicated by bone deposition that has been fully remodeled. At this stage of healing, the injured area exhibits an altered morphology, yet the bone surface is smooth. It is likely that the four fractures of Skeleton 2 were all sustained within a short time span, if not as part of a single episode. Depressed cranial fractures like those observed for Skeleton 2 are often associated with interpersonal violence. That the majority of the observed fractures occur on the frontal and left parietal bones suggests a face-to-face conflict with a righthanded assailant. Conversely, the fractures observed on the cranium of Skeleton 18 are not consistent with face-to-face conflict. Since the pattern of fractures is not consistent with an accidental fall, this may suggest their possible postmortem etiology.

Finally, two cases of dental anomalies with genetic influences were observed. Skeleton 12 possessed an isolate peg-shaped lateral upper incisor (side unknown due to destruction of the alveolar bone). Skeleton 15 possessed a supernumerary tooth between RP3 and RP4. The tooth is shaped like an upper premolar, thought its mesiodistal diameter is significantly reduced. The tooth also was affected with a large carious lesion. Both of these dental anomalies are examples of non-metric traits as defined by the Arizona State University Dental Anthropology System (TURNER et al 1991). The latter case is an example of polydontia, which has somewhat regular frequencies among sub-Saharan African populations (IRISH 1998). 7.4 Skeletal Pathology Nineteen individuals were available for the study of skeletal pathology (Skeleton 20 is represented by isolated teeth only). Several types of pathological conditions were documented. Identifications of pathologies were generally limited to adult individuals, though two sub-adult individuals (18 and 19) exhibited possible pathological conditions (see below). In all, six adult skeletons (2, 3, 7, 8, 11, and 14) possessed observable conditions of six different types: Spinal joint disease, Degenerative joint disease, Trauma, Non-specific infection, Non-specific stress indicators, and “Other” (represented by the case of attempted trephination). As table 7.7 shows, degenerative joint disease (DJD) was the most commonly observed condition, with 50% (3/6) of the adult sample affected. Spinal joint disease (SJD) affected 33% (2/6) of the sample. DJD was diagnosed using the indicators of osteophyte growth, microporosity, enthesophytes, and eburnation (though none of the observed conditions were serious enough to involve eburnation). SJD was similarly indicated by the morphological changes of microporosity and osteophyte growth on vertebral bodies and neural arches. Vertebrae and the articular surfaces of diarthrodial joints were among the most poorly-preserved and under-represented skeletal elements in the Mouhoun Bend assemblage. Thus, it is impossible to estimate the true frequencies and extent to which the Mouhoun Bend skeletons were affected by DJD and SJD.

One case of non-specific infection (periostitis on the left fibula of Skeleton 14) and one possible case of a non-specific stress indicator (healed porotic hyperostosis on Skeleton 2) were also observed. Periostitis is the inflammation of the outermost layer of the bone (periosteum), which results in the deposition of discrete patches of woven (reactive) bone. Porotic hyperostosis is grouped in the category of “Non-specific stress indicators,” along with such conditions as cribra orbitalia, Harris lines, enamel hypoplasias, and delayed long bone growth. The possible causes of these conditions are numerous and often difficult to ascertain, though they usually result from some combination of malnutrition and disease. Skeleton 19, a sub-adult aged 12-14 years, exhibits an interesting condition grouped under “Other” due to the difficulty of description, which in turn is due to poor preservation involving possible termite activity. The defect involves the occipital bone, and consists of an amorphous, calcified or ossified object that was found embedded in the dirt which adhered to the endocranial surface of the occipital bone. The object is roughly ovate, with dimensions of 20.56mm, 13.99mm, and 15.55mm. Interestingly, it appears to have a small, circular, polished surface (5.49mm x 5.64mm) that may have been in continued contact with another bony surface. The occipital fragment associated with the object has two defects with smooth margins yet un-remodeled exposed diploë. It is not clear whether these defects are pathological (in which case the smooth margins would indicate healing) or the result of taphonomic process (in which case the smooth margins would indicate post-depositional erosion or termite activity). A search of

Traumatic conditions were observed on two adults (2 and 3) and one sub-adult (18), and all observations were confined to the cranium (tab. 7.3). However, description of both Skeletons 3 and 18 was confounded by taphonomic effects. Skeleton 3 exhibits the attempted trephinations and will be discussed at length below. Skeleton 18 is a

173

Archaeology of West African Mound-Clusters

the paleopathology literature presents only speculations of possible etiologies, the most likely of which may be neoplastic. The calcified or ossified mass could be the result of an intracranial meningioma, a benign tumor of the dura mater (Aufderheide and Rodriguez-Martin 1998). This soft-tissue tumor can cause changes in the surrounding bone, eroding the cranium’s inner table and/or causing hyperostosis. Unfortunately, a differential diagnosis for Skeleton 19’s condition is not attainable at this point. Future analysis may reach a more firm conclusion.

An irregular defect also occurs on the frontal bone to the right of the midline. Here, a wide, roughly linear section of the vault is missing, extending superiorly from the medial section of the right supraorbital margin. Just to the right of this defect, another circular area of the cranial vault is missing. Both defects have smooth margins and roughly triangular sections of apparently scraped outer table extending from these margins. However, the nature of these defects are less clear, since they do not resemble normal patterns of trephination and are obscured or possibly caused by post-depositional fracture and subsequent erosion of their margins. Rodent gnawing or termite activity could have further contributed to their observed morphologies. Other small defects scattered on the cranium point to taphonomic effects, and suggest that a mixture of pathological and pseudo-pathological defects is present on the cranium.

7.5 Attempted trephination at Tora-Sira-Tomo 7 The evidence of attempted trephination on the cranium of Skeleton 3 represents one of the most exciting aspects of the Mouhoun Bend osteological research. At the same time, it presents interesting paths of future research regarding cultural change in the Mouhoun Bend region.

Interestingly, a clear observation was made of a well-healed, irregular-shaped defect in the occipital bone. This defect has smooth margins and rough diameters of about 20mm and 26mm. It appears to be a pathological lesion, though of unknown etiology. It can be said with some confidence that it is not the result of trauma. Such a wound would be difficult to sustain either accidentally or in conflict, due to its inferior location on the occipital bone, an area of the cranium covered by thick muscles in life. Thus, it is possible that the lesion has a neoplastic or infectious origin.

The cranium described contains three loci of attempted trephination on the frontal, left and right parietal bones. All three defects are perimortem (there are no signs of a biological response), which means that we cannot be sure whether the modifications were performed shortly before or after death. Thus, the possibility cannot be ruled out that the modifications were performed postmortem, perhaps as a means of practice. We use the term “attempted” trephination because complete perforation of the inner table. of the cranium was not achieved in two of the observed loci. A third locus suggests complete perforation. Unfortunately, poor preservation, adhering sediment, and possible termite activity hinders our ability to confidently assess the relationship between the attempted trephinations and pathological conditions present on the cranium. Our analysis included extensive metric, descriptive, radiographic, and photographic documentation, as well as a review of the archaeological and ethnographic literature dealing with indigenous African trephination.

The observation of a cranial pathological condition lends an important aspect to our description of the attempted trephinations. Most researchers studying trephination in various regions of the globe propound the medicotherapeutic motivations for trephination, though some refer to more magical reasons for the surgery. In the vast majority of trepanned crania from the Andes region, for instance, some evidence of previous cranial injury or other pathology is almost always observed in conjunction with the modification. Most trephinations throughout time and space were probably performed in order to relieve the pain of intracranial pressure, which resulted from cranial fractures or other conditions. Since Skeleton 3 possessed a pathological lesion in his cranium, it may be more likely that the modification was performed ante-mortem.

The largest locus of modification is located on the left side of the frontal bone, about 8-10mm from the coronal suture. It is roughly circular with a maximum dimension of 52.00mm and a minimum dimension of 41.82mm. The outer table of the cranial vault has been scraped away, leaving a rough but slightly eroded area of diploë exposed. A smaller locus of modification is located on the left parietal, about 15mm anterior to the left lambdoidal suture. Part of the margins of the locus are damaged and missing, but may have extended onto the left temporal bone. The shape is ovoid and has a maximum dimension of 39.77mm and an estimated (due to damage) minimum dimension of 28mm. Here as well, the outer table has been scraped away, leaving a rough area of diploë exposed. The margins of both the loci just described appear beveled, having inward-sloping margins from the outer table. However, poor preservation and erosion have partially obscured this aspect of their morphology. A third poorly preserved and damaged area of the right parietal bone shows an angular linear defect that may either be a perimortem fracture or the edge of a third trephination.

The observation of attempted trephination from Tora-SiraTomo-7, regardless of whether the surgery was performed ante-mortem or postmortem, prompted a survey of the distribution of indigenous African trephination. Historical and archaeological reports indicate that trephination was a traditional medical practice in a number of African cultures. The earliest historical reference was made by Herodotus who reported the operation among the Tuareg of Libya. Prehistoric crania with evidence of trephination are known from sites throughout North Africa, precisely at Taforalt in Morocco and Afalou Bou Rhummel in Algeria (crubezy et al 2001). Historical reports of the practice among traditional surgeons extend across this area and into eastern and southern Africa. Aside from one historical

174

Human Skeletal Remains

account of trephination in southern Nigeria, West Africa is completely devoid of either historical or archaeological reports of trephination. This raises an obvious question: How did the people at Tora-Sira-Tomo learn the practice? It is equally plausible that the practice was independently invented or borrowed through contact with other groups. A definitive answer to this question is not in sight at this time. Perhaps a more interesting question involves the relationship among trephination, and the technological domains of early iron production and food production in the Mouhoun Bend. Dating of Skeleton 3 is badly needed, along with archaeological evidence for possible tools used in the practice of trephination.

section), and even occupational activity patterns. Before addressing these aspects, a discussion of the taphonomic processes that have adversely affected our analysis is necessary. Poor preservation among the Mouhoun Bend skeletons seems largely attributable to termite activity and the effects of water saturation of the surrounding burial matrix. As a result of these effects, many bone surfaces have become significantly altered, while many bony elements have completely disintegrated. As expected, sturdier bones such as the humerus, femur, tibia, and cranial vault bones are better preserved in this skeletal sample. While these bones enjoy greater representation in the Mouhoun Bend assemblage (at the expense of weaker bones, especially ribs and vertebrae), no bones have escaped the apparent marks of termites and water damage. This skeletally diffuse distribution of effects supports our feeling that termites are largely responsible, since rodents (which commonly affect other skeletal samples) often leave their marks only on certain bones that better fit the morphology of their gnawing apparatus (aufderheide & rodriguez-martin 1998). However, there are a few observed instances of rodent activity, and many more could have been subsequently obscured by the gross effects of termite activity.

An estimation of stature was attempted for Skeleton 3 after gross observation of the pronounced robusticity and length of the long bones. Unfortunately, no long bones were completely intact and available for precise measurement; however, the left tibia was sufficiently preserved to allow close estimation of the maximum length. The measurement is an estimate based on comparison of the incomplete tibia to the longest complete tibia in our laboratory’s comparative collection, which measured 408mm. An osteometric board was used to obtain measurements of the complete tibia and the broken specimen from Skeleton 3. Then, a simple proportional equation yielded an estimated maximum length of 483mm. This measurement was then plugged into a formula developed by Krogman and Isçan (1986), which converts the long bone measurement into an estimation of stature. The specific formula used was developed through measurements of stature among large samples of African American males. Accordingly, a rough estimate for Skeleton 3’s living stature was found to be 198.38cm, or 78.1in (6’6’’). We can speculate that his formidable body size was towards the upper extreme of variation among his local Holocene population, which probably was struggling with the nutritional deficiencies of developing agricultural production. His physical stature may have had some direct relationship to his social stature. This hypothesis may be supported by archaeological evidence such as burial characteristics and artifact associations, and it may further relate to the attempted trephinations performed on this individual. In some African cultures, trephination may have been performed without sex- or age-related biases, but it may have been available only to a disproportionate group of high-status individuals, if only through one’s ability to afford the price of the surgery (margetts 1967). Thus, the correlation between Skeleton 3’s large physical stature and cranial modification, while speculative, presents interesting questions regarding the social matrix that surrounds the biological entity, from which we draw our observations.

It is important for the paleopathologist to be familiar with the local taphonomic processes that could affect the bones he/she studies. Pseudo-pathological features are especially caused by the activity of insects. In fact, our first observations on some of the Mouhoun Bend skeletons generated some unwarranted excitement based on surprising patterns of pathologies and their would-be implications on the topic of the prehistoric epidemiology of syphilis. Henderson (1987 in aufderheide & rodriguezmartin 1998) recounts a case of pseudopathology among several Egyptian skeletons that appeared to have suffered the effects of cranial syphilis. It was subsequently found that the erosive effects of beetles were responsible for the lesions. Aufderheide and Rodríguez-Martin (1998: 16) advise the paleo-pathologists to “become acquainted with the effects of insects whose proteolytic enzyme secretions can liquefy the organic matrix of bones so that, when subsequent ground water action solubilizes the remaining bone mineral, the localized defects can simulate disease processes.” This sequence of events is very likely the cause behind the patterns of skeletal defects observed among the burials excavated from the Mouhoun Bend, a region plagued by termites and seasonal rains, and subject to the wanderings of the Mouhoun River, which has changed its course several times during the last few millennia. Despite limitations on observation due to poor preservation of skeletal elements, it is possible to discuss some aspects of the health status and life-ways of people buried within the settlement complexes recorded by the Mouhoun Bend Archaeological Project.

7.6 Global Assesment Within this assemblage, poor preservation most severely limits observation of skeletal pathological conditions. Because teeth are much more durable, they present a relative wealth of information regarding diet, health, biological affinities with other African groups (see next

Regarding skeletal pathology, it must first be noted that

175

Archaeology of West African Mound-Clusters

ingestion of significantly softer foods. The result is decreased use of the masticatory musculature, which has a direct effect on the size of the bony masticatory apparatus. Meanwhile, the teeth, which are relatively slow to respond to novel environments due to the strong genetic component of their development and eruption, experience no similar decrease in size, and dental crowding becomes a problem. The fact that only the mandibular dentition was affected among these five individuals supports this explanation. The mandible is the mobile component of the masticatory apparatus, which receives the insertions of both the masseter and temporalis muscles. Thus, it is possible that dental crowding in the Mouhoun Bend assemblage reflects the local transition from hunting and gathering to food production.

poor preservation of skeletal tissue may destroy many antemortem lesions and result in uncharacteristically low observed disease frequencies for a skeletal population (waldron 1987b in aufderheide & rodriguez-martin 1998). It must be reiterated that the skeletons excavated in 1999 and 2000 in no way represent a discrete population in either time or space, since they were located at different levels of occupation from distinct probes within three settlements, spanning several hundred years (ca. 700 BC to AD 1600). Still, the poor preservation of the skeletons should remind us to doubt how well their observed frequencies of pathological lesions represent the health status of the living population occupying the Mouhoun Bend during this time period. Degenerative and spinal joint disease (despite very poor preservation of vertebrae and articular surfaces) and trauma (in one or possibly three cases reflecting interpersonal violence) are the most commonly occurring pathologies among the Mouhoun Bend skeletons. We can only speculate that the true frequencies of these pathologies would be higher if the bones were preserved in good condition. However, even the observed frequencies among this poorly preserved sample suggest that osteoarthritis and cranial fractures were common among the skeletal population. Thus, the lives of the people inhabiting the site under study may have included heavy workloads (at least for some) and some occurrence of interpersonal violence. Unfortunately, it may be safer to say that no accurate representation of the health, workloads and prevalence of violence is attainable from this skeletal sample, due to poor preservation and heterogeneity of provenience.

Moderate to heavy calculus, or calcified plaque, was also observed at high frequencies in the sample. Heavy patterns suggest that diets did not include very gritty constituents that usually facilitate natural cleaning (Buikstra & Uberlaker 1994). Calculus samples may further serve to reconstruct the diet of the individuals in the sample, through microscopic and chemical analyses. This avenue of future research could generate interesting paleo-dietary evidence, comparable to the large amounts of botanical macro-remains recovered from excavations (holl & kote 2000). Both Tora-Sira-Tomo and Kerebe-Sira-Tomo yielded skeletons with abnormal patterns of attrition that may be the result of occupational use of the teeth. However, the two individuals with distinct patterns of wear among the upper and lower central incisors, described as forming a roughly circular area of abnormal wear when the anterior teeth of these individuals were centrically occlused, belong solely to KST. Archaeological evidence or ethnographic accounts may be able to shed light on the specific activities (e.g., using the anterior dentition as a “third hand” for stripping plant fiber and softening leather (aufderheide & rodriguezmartin 1998) that are responsible for the artificial wear observed among the Mouhoun Bend skeletons. Intentional shaping of the anterior dentition as a sign of social identity could be another factor, although microscopic observation of the affected teeth has revealed that this is not a likely explanation. It has been observed that chewing sticks as a method of dental hygiene persists to the present day in many parts of sub-Saharan West Africa. If this behavior was common in ancient times it may explain some of the wear patterns that we have observed. However, heavy calculus deposition among the sample suggests that this is not a likely explanation.

Dental pathology and patterns of attrition can provide information on the subsistence practices of the people occupying the Mouhoun Bend region. A relatively high frequency of occlusal crown caries among a skeletal population, as observed in this sample, suggests that the diet of the living population contained a high proportion of carbohydrates. For several investigators (cohen & armelagos 1984; Buikstra & Uberlaker 1994), increased occlusal caries frequencies among Holocene populations have served as evidence for local patterns of agricultural intensification and the dependence upon dietary staples. It is possible that the observed frequency of caries for the present sample reflects a growing intensification of food production. Unfortunately, the very small sample size of observable dentitions hinders a sound conclusion. Dental crowding was observed among five out of nine adolescent/adult dentitions in the Mouhoun Bend skeletal assemblage. In each case, crowding was evident only in the mandibular dentition. Dental crowding probably occurs most often as a result of a decrease in the size of the bony alveolus which houses the teeth. This condition has been observed among many Holocene skeletal populations that transitioned from hunting and gathering to food production (larsen 1997). Food production was often accompanied by greater efforts to process food, which resulted in the

One interesting pattern that becomes apparent is that cases of bilateral asymmetry of wear among the molars invariably attribute heavier wear or abnormal wear to the right molars. This perhaps supports an explanation involving occupational use of the dentition, in that right side dominance of the skeleton for occupational activities is understood as a human universal. For instance, random

176

Human Skeletal Remains

samples of any population will usually yield a frequency of 90% for favoring the use of the right arm in various occupational and leisure activities. This frequency of right side dominance may also hold for use of the dentition when occupational demands are placed upon the teeth. The disproportionate use of the right dentition may owe to its own dominance over the left dentition, or it may be an indirect result of the right arm’s dominance and its parallel use in conjunction with the teeth in occupational activities.

Dental pathological evidence suggests that food production was practiced during the period under study. However, no conclusions can be made regarding when agricultural intensification began or its patterns of progression. Further analysis of calculus samples may assist dietary reconstruction. Abnormal patterns of dental attrition suggest that occupational use of the teeth was practiced among at least some groups within the study area. Whether these activities received impetus from craft specialization and intensification of production is impossible to say at this point. However, it appears that different patterns of occupational wear may belong to different settlements, suggesting that spatially distinct groups manufactured distinct products.

Conclusion The excavation and analysis of twenty individuals from the Mouhoun Bend lends an indispensable aspect to the current investigation of emerging food production, iron technology, and craft specialization among a late Holocene sub-Saharan West African population. Although the skeletal sample is small and spans a large time period from ca. 700 BC to AD 1600, it provides clear evidence for, and some biological effects of, these major cultural changes. Unfortunately, diachronic analyses of these changes based on the skeletons are not possible with such a small sample. Our preliminary findings strongly encourage continued excavation within the Mouhoun Bend Archaeological Project.

In the case of the Mouhoun Bend skeletal analysis, what poor preservation takes away from the current study in terms of paleopathology, it in turn offers a separate avenue for future research involving tropical taphonomic processes. The Mouhoun Bend skeletons present useful material for future comparative and analytical studies of New World and Old World tropical taphonomy.

177

Chapter 8

Mound Dwelling Traditions of the Mouhoun Bend

Introduction

karité oil production, and burial. With the exception of the iron-smelting site and the cemetery, all the tested mounds have evidence of habitation and daily life activities. This suggests that some specialized craft people, depending on circumstances, had decided to build their residence in a separate spot.

After a Late Stone Age prelude of unknown duration, the Mouhoun bend was colonized by iron-using communities of mound dwellers. The initial phase of this colonization is still vague but appears to have started in the middle of the first millennium B.C. The climate was wetter. The Sudanian savanna and the Mouhoun River offered a diversified resource mix which allow for the stabilization and growth of these populations of mixed farming fishermen. The small sample of excavated sites offers an interesting entry into the dynamics of settlements that took place in the study area for a little more than 2000 years, from ca. 700/500 B.C. to AD 1600 (Annex B).

The archaeological evidence points to flexible social arrangements, specific to each of the “self-sustaining” and autonomous villages, which were the standard during the 2200 years of the Mouhoun Bend political economy under consideration. On the one hand, and as seen at KerebeSira-Tomo, craft people lived in the main village and had their iron-smelting and blacksmithing workshops in the southern periphery of the settlement. On the other hand, as indicated by the archaeological record from Tora-SiraTomo-/Gnambakouon-Sira-Tomo, craft people built their residences elsewhere, and supplied the rest of the settlement from their workshops. There was clearly no overarching prescribed policy of occupational and residential “segregation” as is the case in recent and present-day ethnography of the Mande world.

8.1 Variability of mounds clustering processes All the settlements recorded in the Mouhoun Bend study area are multi-mounds complexes. No single mound site has yet been discovered. The data collected from the excavation of four settlement complexes point to the existence of two main mound-clustering strategies, a tight and a loose one.

8.2 Patterns of Subsistence

The tight-clustering strategy is documented at KerebeSira-Tomo, a settlement complex which lasted from ca. 350 B.C. to AD 1450. All the mounds, with the exception of Kerebe-Sira-Tomo 4, are tightly packed within the same space of approximately 15 ha, delineating a densely settled village. Iron-working sites are located at a few hundred meters along the south edge of the village. None of the iron-working stations has any evidence of habitation. Consequently, the members of the working crews were all living in the village, commuting to their respective workshops during the iron-production season that may have taken place in the dry season.

The Mouhoun Bend iron-using mound-dwellers were mixed farming and fishing communities who took advantage of the relatively rich biome they settled in (D’andrea 2007; Dueppen 2008; Holl & Kote 2000). They were dry-grain farmers who cultivated millet (Pennisetum glaucum), sorghum (Sorghum sp.), fonio (Digitaria iburua), cowpeas (Vigna unguiculata), and Bambara groudnut (Voandzeia subterranea). They also exploited a wide range of wild plants, and protected a number of tree species which became a specific aspect of their land use patterns. The oleaginous karité (Butyrospermum parkii), protected for its fruits, became a key component of the agro-forestry system which evolved in the area.

The loose-clustering strategy is well represented by theToraSira-Tomo/ Gnambakouon-Sira-Tomo settlement complex with a total of 20 mounds. They are both comprised of well demarcated and individualized mounds, 17 for Tora-SiraTomo, and 3 for Gnambakouon-Sira-Tomo. Each cluster has a large and dominant mound, Tora-Sira-Tomo-3 in one case, and Gnambakouon-Sira-Tomo-3 in the other. All along the occupation history of the complexes, there has always been a number of special purpose sites, for iron-smelting, blacksmithing, cloth weaving and dyeing,

No intensive program geared toward the systematic recovery of plants macro-remains was carried out during the field part of the Mouhoun Bend Archaeological Project. Several attempts at flotation of selected ashy deposits were unsuccessful. The recorded plants remains are therefore purely indicative. They are divided into three groups. There are numerous pieces of undetermined burnt wooden

178

Mound Dwelling Traditions of the Mouhoun Bend

logs. Piles of burnt karité nuts kernels were found at the single occupation mound of Tora-Sira-Tomo-17 (AD 1280 - 1380) and Kerebe-Sira-Tomo-3, occupation I (AD 1000), in contexts indicative of karité oil production. And finally, the burnt food supplies from Kerebe-Sira-Tomo-4 occupation II (AD 1050 - 1380), which included several large storage jars of fonio (Digitaria iburua), and smaller amount of cowpeas (Vigna unguiculata) and Bambara groundnut (Voandzeia subterranea). The actual agricultural system may have involved some multi-cropping, with grain - millet, sorghum, and fonio - grown in association with cowpeas and/or Bambara groundnuts, on fields with scattered protected trees.

The largest part of the recorded faunal spectra is made of aquatic species, reptiles, fish, and shells, with frequencies varying from mound to mound. Reptiles are represented by two species, the crocodile (Crocodylus niloticus) and the monitor lizard (Varanus niloticus). The crocodile remains, three dermal plaques -, were found in occupation II at ToraSira-Tomo-3-West. Monitor lizard bones are present in all mounds with faunal samples. This species can be found in the cultivated fields, the bush, and the river. Instead of being considered as the product of planned hunting expeditions, the capture of monitor lizards was more likely the result of chance encounters during field preparation and fishing operations.

The developed agro-forestry system thus combined the protection and exploitation of a number of trees, the cultivation of a range of crops, and livestock herding (Kahlheber et al 2000).

Fish remains are present in all mounds with faunal samples. Considering the total of the recorded fish bones, a sizable amount is undetermined. With the significant exception of the Tilapia sp. bones found in Tora-Sira-Tomo-3-West occupation II and Nile perch (Lates niloticus) vertebras from Kerebe-Sira-Tomo-3 occupation I and III, bottom feeders from the Siluridae family are the most recurrent fish species found in all the studied samples. Gymnarchus niloticus and Synodontis niloticus are different kinds of catfish. They are well adapted to drastic seasonal variations and can survive in turbid and muddy ponds. They can be caught using different fishing techniques; line- and netfishing at high water and basket and dam-fishing at low water. However tempting it appears, these fish species are not good seasonality indicators. They are available and accessible all year-long.

Faunal remains were collected systematically during the excavations. Special purpose sites devoted to iron-working, cloth weaving and dyeing, pottery firing, and karité oil production were generally devoid of faunal remains. The most comprehensive faunal spectra were collected from Diekono 1 and 2, Tora--Sira-Tomo-3-East, Tora-Sira-Tomo3-West, Tora-Sira-Tomo-5, and Kerebe-Sira-Tomo-3. Domestic animals are represented in varying proportions in all the sites with faunal remains. Sheep/goat (Ovis/Capra sp.) bones are found in almost all the cases. The distribution of cattle (Bos taurus) is more restricted, with most of the bones found at Tora-Sira-Tomo-3-East, Tora-Sira-Tomo-3West, and Kerebe-Sira-Tomo-3. Interestingly enough, no cattle remains were found at Diekono-1 and 2 and sheep/ goat bones were confined exclusively to the foundation deposits of occupation I dated to AD 440-750 and AD 880 - 1185. It is as if the initial iron-using colonizers of the Mouhoun river valley settled with their sheep/goats herds, and shifted later to an exclusive reliance on the aquatic resources provided by the water course. Poultry (Gallus gallus) is represented by the egg-nest from Kerebe-SiraTomo-4 occupation II, and a rooster leg bone at Tora-SiraTomo-13 occupation I, dated to AD 1150-1280. The earliest evidence for the presence of chicken in the Mouhoun bend is dated to the middle of the first millennium AD, at Kirikongo in deposits dated to AD 500-600 (Dueppen 2008: 238).

Fresh water shells are also represented in all the recorded faunal samples. One species, the fresh water oyster - Etheria elliptica - is particularly dominant. There are a few shells of fresh water mussel - Aspartharia sp. - and land-snails Limicolaria sp. -. In summary, the faunal remains collected in the excavated probes point to the existence of two main strategies for access to animal resources. One is based on livestock husbandry indicated by the recurrent presence of cattle and sheep/goat bones. Poultry, represented by the chicken egg-nest from Kerebe-Sira-Tomo-4 occupation II, and a rooster leg bone, can be added to this component. The other is articulated around the exploitation of the Mouhoun River resources, through fishing, gathering of mollusks, and opportunistic capture of reptiles, crocodiles and monitor lizards. As suggested by the evidence from Kirikongo (Dueppen 2008), available wild mammals may have been hunted but did not enter the archaeological record in the sampled areas.

Wild mammals hunting does not appear to have played any significant role in the subsistence of the mounds dwellers from this part of the Mouhoun Bend. Remains of wild cat and undetermined Bovidae sp. were recorded at Tora-Sira-Tomo-3-West and Kerebe-Sira-Tomo-3. Guinea fowl (Numida meleagris) while not ubiquitous is represented in the faunal samples from Tora-Sira-Tomo-5 and Kerebe-Sira-Tomo-3. This species is still part of the local wild fauna. The domestic specimens turn feral very often, making the determination of signs of domestication extremely difficult.

The archaeological record is particularly sketchy as far as faunal remains are concerned. There is no apparent timetrend pointing to a significant shift in the procurement of animal resources during the 2000 years long occupation sequence under investigation.

179

Archaeology of West African Mound-Clusters

8.3 Mortuary practices

of the right leg is however puzzling and tends to point to the practice of secondary burial.

Mortuary practices are good indicators of cultural commonalities and their variations are significant social traits. It is generally easy to outline differences in how the deceased are treated in burials. It is however much more difficult to grasp the meanings of the recorded variations. The analysis of mortuary practices which follows is articulated on the concept of mortuary program, understood as the patterned ways with which the deceased are taken care of (Holl 1994, 2002). In the archaeological contexts of the Mouhoun Bend settlements, the analysis is based exclusively on the arrangement of the body in the grave and the presence/absence of grave goods. Grave pits and burial mounds were not visible in the tested archaeological probes. Almost all the times, there was no indication of burial prior to the exposure of the skeletal remains themselves. The longitudinal perspective adopted in this case helps identify chronological trends, if any.

It is interesting that primary and secondary burials are represented right at the early appearance of the Mouhoun Bend mortuary record. In the recorded primary burials, the deceased bodies were oriented along the south-north axis, facing east. The secondary burials include skulls in two cases and leg bones in the third. 8.3.2 AD 1000 - 1200 burials Two burials are dated to the very beginning of the 2nd millennium AD, from 1000 to 1200. One was found at Kerebe-Sira-Tomo-1B occupation II, and the other at Kerebe-Sira-Tomo-5 occupation III. The former, also known as feature 15, is a complete, relatively well preserved primary burial. The deceased was buried in a tightly flexed position, lying on the left side, oriented southeast-northwest, facing southwest, hands before the face, and wearing a “clay bowl-helmet”. Two additional offering bowls were found in the grave. The latter - the Kerebe-Sira-Tomo-5 occupation III burial - was partly excavated with the upper part of the body, from pelvis to skull unexposed. The deceased was an adult individual lying on the right side, oriented west-east, legs flexed, and probably facing south.

Twenty nine burials were excavated during the three field seasons of the Mouhoun Bend Archaeological Project. Their distribution between settlements and chronological periods is far from even. There are very few early burials dated to the earlier phases of settlement. In fact, the 0 -500 AD time range is devoid of any burial at all and the total number of burials from 500 BC to AD 1200 amounts to only 4. Are there any indications of long lasting changes in mortuary programs which may have had important social and economic implications?

8.3.3 AD 1200 - 1400 burials There are many more burials in the AD 1200 - 1400 time interval, found in five mounds in two settlement complexes. Seven of the burials were recorded at Tora-Sira-Tomo in four mounds, and the remaining 5 were uncovered at Kerebe-Sira-Tomo-4.

8.3.1 The earliest burials (ca. AD 500 - 1000) Burials dated to this period were found in all three settlement complexes, at Tora-Sira-Tomo, Kerebe-SiraTomo, and Diekono. The earliest human remains from Tora-Sira-Tomo-3-East, were found in occupation I at 2.25 m below the surface. The deposit is not dated directly but considering that occupation IV at 1.00 - 1.40 m is dated to ca. AD 1195 - 1400, the second half of the 1st millennium AD is a very reasonable option. The evidence consists of secondary burials, each with a child skull, buried under the floor level. One of the skulls was held in a hole-mouth pot and the other buried simply in the ground. The recorded children are 4 to 7 years old, and their skulls were probably part of the performance of a foundation rite.

The burial from Tora-Sira-Tomo-3-East occupation IV, was partially excavated. Only the legs were exposed in the excavation probe. The deceased, an adult, was buried in a slightly flexed position, oriented south-north, lying on the left side, and facing east. The grave from Tora-Sira-Tomo-4 occupation I was dug in the laterite crust at the very bottom of the mound sequence. The deceased, a 35 years old female was forced in a too small burial pit. She was oriented northsouth, lying on the left side, and facing west. The right arm and leg were folded at right angle, and the left arm and leg were extended.

There are two burials firmly dated to the second half of the first millennium AD, ca. AD 500-1000. The Diekono-2 occupation I grave is dated to AD 890-1190. The deceased was buried in a tightly flexed position, oriented southeastnorthwest, lying on the left side, facing southwest, with hands before the face. The body was covered with large sherds and a very large clay vessel cut longitudinally into half was used as a sort of coffin. The Kerebe-Sira-Tomo occupation I burial consists of a slightly flexed left leg, oriented south-north. A poorly preserved pelvis fragment was protruding from the probe south section, suggesting the possibility of a complete primary burial. The absence

Three skeletons were exposed at Tora-Sira-Tomo-7 occupation I, in what could have been a collective burial monument. Legs bones only were preserved from the burial of a 15 -17 years old individual. The grave pit and the bones were oriented west-east, associated with an elaborate sherds pavement and three clay vessels, two bowls and one pot. The second burial located at the center of the probe was that of a middle adult male, 45+ years old. He was oriented west-east, lying on his back, arms folded on both sides, facing south, with legs tightly flexed. The upper half of the third burial was exposed in the southwest corner of the excavated probe. The deceased was a middle aged

180

Mound Dwelling Traditions of the Mouhoun Bend

adult female, 35 - 50 years old, oriented north-south, laid on the right side, facing west, with the hands before the face. The sherd pavement, the cluster of three pots, and a series of additional clay vessels delineate a semi-circular arrangement that may indicate the presence of small burial mound. Both man and woman present evidence of trauma, with the former having gone through at least two distinct episodes of head surgery (trephination). They were clearly victims of violence, very likely from a raid on their village, and were honored and buried in the same grave.

of a small round hut, or more probably had a circular low mound delineated by a circle of large laterite blocks. The deceased, an adult individual, was laid on the left side, oriented southeast-northwest, legs slightly flexed, facing southwest, with the hands under the head. The grave from Tora-Sira-Tomo-7, occupation III, contained the remains of a 5-7 years old child, laid on the right side, oriented west-east, facing north, with legs slightly flexed. Most of the recorded burials, 10 out of 12, were part of the “restricted access” cemetery of Tora-Sira-Tomo-9. They include 3 secondary and 7 primary burials. Feature 14, 15, and 18 are secondary burials found in different areas of the cemetery. Feature 14 contains poorly preserved remains of an adult skull. Feature 15 includes a series of adult long bones. And finally feature 18 has the skull of a 9-11 years old child. The excavated primary burials can be distributed into three sub-sets according to their overall orientation. Burial 16, 20, 21, and 22 are oriented southwest-northeast, burial 10 northeast-southwest, burial 11 southeast-northwest, and burial 12 south-north. Burial 16 and 20 are part of the same cluster. Both deceased are adult. The former was laid on the left side, facing northwest with legs and arms flexed. Burial 21 and 22 are also part of another cluster. The former is 18+ years old and the latter an adult male, 25-50 years old.. Both were buried exactly the same way, lying on the right side, legs slightly flexed, facing southeast with hands before the face. Burial 10 and 11 are a special case of two adult females of virtually the same age, 35-40 and 33-38 years old, who were fit more or less in the same grave. Individual 10 was buried first, laid on her back, oriented northeast-southwest, arms “raised” on both side of the head, with extended legs. Individual 11 was buried second, laid on her right side, oriented southeastnorthwest, legs flexed, face down, and arms on the chest. The femurs of individual 10 were removed from their initial position and piled along the back of Individual 11. Finally, burial 12, part of the cluster which includes burial 10 and 11, contains poorly preserved remains of a 18 years old individual, lying on the right side, oriented south-north, legs slightly flexed, and facing east.

Two additional graves were excavated in occupation I at Tora-Sira-Tomo-8. One is a secondary burial containing the skull of an adult individual. The other is a poorly preserved partially excavated grave of a 18+ years old individual. The neck and the skull were still in the probe section but all the long bones were exposed. The deceased was oriented west-east, laid on the back, legs flexed on the left, with arms extended along the body. An iron arm-ring and two broken iron artifacts were found in the grave. Four burials were excavated at Kerebe-Sira-Tomo-4. The recorded material and the context suggest the use of this small mound as a burial ground after its abandonment when three graves were dug in the older deposit. One of the grave, burial H2, dates from occupation II and includes the body of an individual laid to rest on the floor of house 2, without any evidence of burial pit. The house was burnt by a fire that had destroyed most of the dwelling installations from occupation II. Individual H-2 may have been a victim of the raid. Whatever the initial circumstances, individual H-2 was laid on the right side on the floor of house 2, oriented west-east, facing south, the legs slightly flexed, the right arm extended and the left arm folded. The body was covered with sediment containing a large amount of sherds and a large pot was set above the legs. Burial E2, located at the center of the probe contains the skeletal remains of an adult individual. The deceased was laid on the left side, oriented north-south, facing east, with legs and arms flexed. Burial E4, in the same area of the excavated unit, is a secondary burial containing two poorly preserved adult long bones associated with one miniature iron axe and an iron ring. Burial E6 is located in the same area. It is a complete partially exposed primary burial. The lower legs of an adult individual were exposed. The deceased appears to have been oriented southeast-northwest and laid on the right side. Finally, burial E8 in the center-east of the excavation is a complete primary burial with the remains of an adult individual. The deceased was laid on the right side, oriented southeast-northwest, the legs slightly flexed, facing northeast, with the hands before the face. The base of a three-legged bowl was found in the burial. 8.3.4 AD 1400 - 1600 burials

With the exception of the secondary burial 14 which is relatively isolated, all the graves from Tora-Sira-Tomo-9 are arranged in clusters with at least two generations represented: a young 18+ years old individual associated with a 25-50 years old adult male; a 9-11 years old child in the company of three adults of unknown sex in the Burial 15 - 20 cluster; and finally, a young 18 years old associated with two middle adult females aged 33 - 40 years in the burial 10 - 12 cluster. As far as the orientation of the deceased bodies are concerned, virtually all the individuals but one (burial 10) are buried with their head positioned in the south, southeast, and southwest.

The sample of AD 1400 - 1600 burials consists of 12 tombs found in three mounds, all from Tora-Sira-Tomo settlement complex. The burial from Tora-Sira-Tomo-3East occupation VI, may have been located under the floor

The development of cemeteries appears to have taken place at the very end of the mound-dwelling traditions in the Mouhoun bend. The data from Kerebe-Sira-Tomo-4 occupation IV are particularly ambiguous and difficult to

181

Archaeology of West African Mound-Clusters

interpret. The uninhabited mound could well have been used as a cemetery. The data are clearer and more straightforward for the “restricted access” cemetery of Tora-Sira-Tomo 9.

series of selected long bones. Primary burials are essentially single with interesting variations in the orientations and positions of the bodies. The southeast-northwest orientation is represented in all the cases from ca AD 500-1000 to 1400 - 1600. It is followed in frequency terms by the westeast orientation recorded from ca AD 1000-1200 to AD 1400-1600 with however a peak in the ca. AD 1200-1400 period. North-south orientations are found exclusively in burials from the ca. AD 1200-1400 period. In summary and based on the sample of excavated burials, a trend toward increasing diversification of burials orientation appears to have been kicked off in around ca AD 1200 - 1400 to peak in ca. AD 1400 - 1600, during the last two centuries of the Mouhoun Bend mound-dwelling traditions. This last period also witnessed the emergence of formal disposal areas, as shown by the “restricted access” cemetery of Tora-SiraTomo-9.

8.3.4 Patterns of mortuary variability Grave-goods and ritual offerings are particularly rare in the sample of excavated burials and consist of two bowls from Kerebe-Sira-Tomo-IB occupation II burial, one iron arm-ring and two broken iron artifacts from Tora-SiraTomo-8 occupation I, and two iron artifacts and the base of a three-legged bowl from Kerebe-Sira-Tomo-4 occupation IV burial E4. The practice of grave-goods took place at the beginning of the second millennium AD and lasted up to AD 1400. Burials orientation and the pattern of body arrangement are not gender or age-specific. The number of graves per period is critical for the evaluation of the range of variation in the key characteristics of the mortuary program described so far. Data from ca. AD 500 - 1000 to 1000-1200 are particularly limited with only six cases of burials. The secondary burials with children skulls at ToraSira-Tomo-3-East occupation I, do not present any specific orientation. The two burials dated to ca AD 500 - 1000 are oriented southeast-northwest at Diekono-2 occupation I, and south-north at Kerebe-Sira-Tomo-5 occupation II. The burials from ca AD 1000 - 1200 which also amount to two cases also display two orientations: southeastnorthwest for Kerebe-Sira-Tomo-1B occupation II, and west-east for Kerebe-Sira-Tomo-5 occupation III. There are many more burials dated to ca AD 1200-1400. But the range of variation in orientation is surprisingly narrow with only 3 options for 12 graves. Three burials found, at Tora-Sira-Tomo-3-East occupation IV and Kerebe-SiraTomo-4 occupation IV, are oriented southeast-northwest. Four inhumations, from Tora-Sira-Tomo-5 occupation I, Tora-Sira-Tomo-7 occupation I, and Kerebe-Sira-Tomo-4 occupation IV (burial E-2 and E-4), are oriented northsouth. And finally, west-east orientation was recorded for four tombs exposed at Kerebe-Sira-Tomo-4 occupation IV, Tora-Sira-Tomo-7 occupation I (2 burials), and ToraSira-Tomo-8 occupation I. The later phase of the mound dwelling traditions, dated to ca AD 1400 – 1600, is the period of the emergence of formal disposal areas, as indicated by the “restricted access” cemetery of Tora-SiraTomo-9. Two additional tombs found at Tora-Sira-Tomo3-East occupation IV and Tora-Sira-Tomo-7 occupation III are respectively oriented southeast-northwest and west-east. The burials from the Tora-Sira-Tomo-9 cemetery present a relatively broad range of orientations. Burial 16, 20, 21, and 22 are oriented southwest-northeast. Burial 10 is oriented northeast-southwest, burial 11 southeast-northwest, and burial 12 south-north. The southwest-northeast and northeast-southwest orientations are exclusively represented in the cemetery of Tora-Sira-Tomo-9 and dated to the late phase of the Mouhoun Bend mound dwelling traditions.

8.4 Craft specialization Many facets of more or less specialized crafts have been documented in the archaeological record of the tested settlement complexes. Pottery production is ubiquitous and probably the most widespread. Besides its products, its production areas are nonetheless difficult to trace in the archaeological record. Iron metallurgy is clearly massively present, with evidence of all the steps of the iron artifacts production process - the chaine operatoire -. Cloth weaving and dyeing, masonry, and karité oil production are also attested. 8.4.1 Pottery production Pottery is by far the most frequent and abundant element of material culture found in the region. Good quality clay is also ubiquitous and the local production of clay objects and containers cannot be questioned. But archaeological features which can be unequivocally linked to pottery production are most of the time absent from the excavation units sunk in the selected settlement complexes. The firing of small quantities of clay vessels using the bonfire method can be done in any open space. Such burnt patches, if found in an archaeological probe, are generally difficult to interpret. It is the case at Kerebe-Sira-Tomo-5 occupation I, where an irregular shaped burnt red pink surface was exposed. It is not dated directly but was probably in use during the second half of the 1st millennium AD. The circular feature from Kerebe-Sira-Tomo-1A could have been a pottery firing “kiln” dated to the second half of the 1st millennium BC even if another interpretation is equally plausible. The manufactured vessels range from small service bowls to very large storage jars and portable hearths. It is very likely that each of the settlement complexes had a number of potters servicing the community. The shapes, sizes, and decorative repertoires of the recorded pottery point to the same global cultural universe, with however subtle variations from one settlement complex to the next.

The mortuary programs implemented in all the settlement complexes excavated in the Mouhoun Bend include primary and secondary burials. Secondary burials consist of skulls or

182

Mound Dwelling Traditions of the Mouhoun Bend

8.4.2 Iron metallurgy

Each village thus appears to have had its iron-smelters and blacksmiths who supplied their respective communities with the needed iron tools and objects.

All the tested settlement complexes include at least one quarry from which laterite was extracted for iron production and the flooring of houses and courtyards. The quarries from Tora-Sira-Tomo-1 and Kerebe-Sira-Tomo-2 were described. That from Diekono was visited but not studied in detail. In all the cases, they are large and relatively deep excavations resulting from long lasting exploitation. The raw material most suitable for iron production comes from the deeper stratum of the laterite crust (Coulibaly 2006; Kiethega 2006). The collected iron ore was processed to enhance its suitability for smelting. This may have included the breaking into small pieces, and sun-drying or roasting if wet. The smelting requires careful preparations. The furnace has to be built. The fuel, essentially charcoal, has to be ready. The blow-pipes have to be available. A number of iron-smelting sites were excavated at Tora-Sira-Tomo-2, Kerebe-Sira-Tomo-8, 9, and 11. The furnaces were of the natural draft variant in which the constant dry season wind - the Harmattan - is harnessed to run the smelting operations. They are cylindrical in shape, measuring 1.3 to 2 m in maximum diameter at the ground level with an upper chimney of unknown length, and an underground combustion chamber where tuyeres were fired. Each of the tested iron-smelting sites had one and only one furnace. In fact, all the recorded furnaces were of the semi-subterranean variant, with the underground combustion chamber 1 to 1.4 m deep (Coulibaly 2006; Kiethega 2006, 2009).

8.4.3 Masonry: house and courtyard flooring Evidence of highly skilled masonry is documented at almost all the sites as carefully made and maintained house floors and courtyards. It is difficult to know if such skills were under the exclusive control of a “guild” of masons as is the case today for iron-workers in most of West Africa, or if they were accessible to everyone without gender prescription. In any case, the skills and techniques required include the ability to locate the adequate kind of laterite, to quarry it and process it into good flooring material. There may have been some division of labor. The procurement and first stage of processing of the raw material could have been operated by one group. Their product could have been exchanged or sold to other individuals with more “flooring”expertise”. Or “buyers” of the processed raw material could have hired experts’ masons to take care of the flooring of the dwelling installations. The visual examination of floor sections shows that re-flooring was a frequent and common practice. The laterite had to be pounded and ground into fine powder. This powder was mixed with clay and organic temper material like straw and/or grass and left to soak in water for a number of days. The goal is to manufacture a paste as homogeneous as possible. An appropriate apportionment of each of these components - ground laterite, clay, straw or grass, and water -, is critical for the production of this kind of dark-red “plastering” material. It is then laid in thin layers over the surface to be floored and left to dry slowly to avoid premature cracks. The frequent use of fire and the heat help in hardening the floored areas.

Tuyeres production was streamlined at such sites as ToraSira-Tomo-1, Kerebe-Sira-Tomo-8 and 10. A number of blow-pipes at different stage of the manufacturing process were uncovered at these sites at the bottom of the smelting furnace. The specimens from Kerebe-Sira-Tomo-8 and 10 made of white clay, were set vertically and still unfired. Those from Tora-Sira-Tomo-1 were fired and arranged leaning against the furnace wall. Iron slag was found virtually in all the tested sites, with the largest accumulation documented at Tora-Sira-Tomo-1. Forges installations, in this case forge furnaces, indicative of blacksmiths workshops, were uncovered at Tora-Sira-Tomo-3-East occupation II and Kerebe-Sira-Tomo-7. None is dated directly. The former workshop can however be dated to the end of the 1st millennium BC/first half of the 1st millennium AD, and the latter was probably in operation in the second half of the 1st millennium AD. There is the possibility of two additional blacksmith workshops at Tora-Sira-Tomo-12 and 15 where diverse sets of “unused” iron artifacts that may have been on “sale” were found. Both sites are more or less contemporaneous, dated to AD 1160 - 1430 for the first and AD 1025 - 1320 / 1230 - 1430 for the second. A relative broad range of iron artifacts was recorded in the tested sites. Weaponry is largely predominant with arrowheads, spears, and knives. Elements of personal adornment follow with arm-rings, finger-rings, “nose-rings”, as well as buckles used for clothing purposes. Tools like hooks, hoes, and many undetermined items are also represented.

8.4.4 Weaving and Cloth dyeing. Cloth dyeing probably associated with weaving is documented at Tora-Sira-Tomo-4 occupation I. One spindle-whorl was found among the scattered and crushed vessels. However, most of the evidence consists of 12 medium to large size hole-mouth vessels used to soak the fabric in the dye. During occupation I, the site was used exclusively for cloth processing activities. The weavers/ cloth dyers did not reside on Tora-Sira-Tomo-4 but used it as a special purpose workshop. During occupation II, the mound was turned into a standard habitation site. 8.4.5 Karité oil production Karité oil can be produced in small quantities within the domestic unit without any special installation. The nuts are cracked and dehusked, put in a large pot with water, boiled, pounded, ground, and the paste heated to release all the oil. This oil - also called shea butter - is an extremely versatile produce used in ritual, medicinal contexts, and daily food consumption.

183

Archaeology of West African Mound-Clusters

Direct evidence for the processing of karité nuts was recorded at two or three sites, possibly at Kerebe-SiraTomo-1A, and certainly at Tora-Sira-Tomo-17 and KerebeSira-Tomo-3 occupation I. In the last two cases, the evidence consists of piles of burnt karité nuts kernels. These deposits dated to AD 1280-1380 at Tora-Sira-Tomo-17 and AD 1040-1380 at Kerebe-Sira-Tomo-3, are contemporaneous. There is a number of artifacts which were more or less related to the karité oil production but are difficult to interpret when found out of a coherent activity area. It is the case for flat and elongated grindstones, and crescentshape and worn-out pear-shaped grinders. The latter are the discarded remains of large crescent-shape grinders used to grind karité nuts. The circular 2-m in diameter thick walled installation from Kerebe-Sira-Tomo-1A may also have been part of a karité oil production system. In that case, it was a karité nuts grinding platform dated to the later part of the 1st millennium BC/ beginning of the 1st millennium AD, (ca 350 BC - 110 AD).

This category of vessels appears to have been used in different contexts, in burials as well as daily life circumstances. The dated specimens belong to the first half of the 2nd millennium AD but earlier specimens cannot be completely ruled out. Footed bowls have been recorded along the upper and middle Niger, from Niani in the south to Tombouctou in the north, in the Bandiagara cliff and the Inland Niger delta (Bedaux 1980; Filipowiak 1979; Mcintosh & Mcintosh 1980). They are dated from the 6th to the 17th century AD and were used in different contexts and circumstances. At Niani, the footed bowls were found in both domestic and funerary contexts. They could have been used to serve beverage - palm wine, millet beer - during funerary meals in honor of the deceased (Filipowiak 1979: 278). The specimens from Goundam-Niafounke and Kouga further north were interpreted as karité oil lamps (Mauny 1961: 351). The bowls from the “Tellem civilization” found in the Sanga - Bandiagara cliff area and dated to the 11th - 13th century AD, were all collected from burial contexts (Bedaux 1972, 1980). The Jenne-jeno specimens were collected from the mound fill mostly in domestic contexts (Mcintosh & Mcintosh 1980). Despite their considerable versatility, Filipowiak (1979: 279) considers the footed bowls to have been basically ritual vessels marking the expansion of Mande speaking communities along the Niger River and its tributaries.

8.5 Long distance interaction The Mouhoun bend is connected to the Inland Niger delta by the Sourou (formerly known as White Volta) River which flows north and south depending on seasons. With this link, one would expect the development of long distance exchange network between the Niger valley and the territories on its periphery. The Inland Niger delta is essentially an alluvial and sedimentary environmental context devoid of rocks and iron ore. Rocks, iron bars, and other artifacts could have been imported from the Mouhoun bend and river valley. The impressive iron ore mine site of Tissi could have been part of that plausible exchange network.

The Mouhoun bend specimens share the main characteristics of those from the rest of the upper and middle Niger sites. They date from the beginning of the 2nd millennium AD, the major phase of expansion of Mande speakers and the rise of the Mali Empire. It is however not yet known if they were obtained through exchange, copied by local potters, or more likely both. 8.5.2 The Cowry shells

There are three series of evidence pointing to long distance interaction between the mound-dwellers of the Mouhoun bend and the rest of West Africa. One of the series is made of a number of remains of three-legged footed bowls. The other includes cowry shells (Cyprea monetaria), and the last one consists copper arm-rings.

The second series of remains with direct connection to long distance interaction consists of two cowry shells, cyprea monetaria. One was found in occupation III deposit at Tora-Sira-Tomo-3-West and the other at Tora-Sira-Tomo-13 occupation I. The former is dated to AD 1040 -1240 and the latter to AD 1150-1280. The shells were not associated to any specific context or feature but were found in the sedimentary matrix of the mounds. These cowry shells were very likely lost, unrestrained, and entered the archaeological record accordingly. Similar evidence were found at Kirikongo where “the fill above Burial Episode 3 contained eight cowry shells” (Dueppen 2008: 252). The deposit is dated to AD 1100-1300, overlapping significantly with the data from Tora-Sira-Tomo. Most of the inhumations found in the reported cemetery are those of infants and children but the association between the cowry shells and any specific burial is far from clear. Cowry shells dated to the 6th - 7th century AD were found in two burials in the cemetery of Kissi 3 in the Oudalan province of northern Burkina Faso (Pelzer & Magnavita-Santos 2000). They were part of a head-gear, a head-band or a hat, and as such were an element of personal adornment and dressing style.

8.5.1 The three-legged footed bowls The recorded sample of footed bowls is made of four specimens, found in each of the tested settlement complex. Three of specimens are footed bowls bases, found at Diekono-1, Tora-Sira-Tomo-8, and Kerebe-Sira-Tomo-4. The Diekono-1 specimen, made of white clay was found in occupation IV dated to AD 1000-1200. The Kerebe-SiraTomo-4 piece comes from occupation IV burial E8 and is dated to AD 1200-1400. Finally, the Tora-Sira-Tomo-8 fragment was collected from occupation III deposit. It is unfortunately not dated directly. The fourth specimen, a complete large three-legged footed bowl, was found in occupation I deposit at Kerebe-Sira-Tomo 6. It is decorated with twisted roulette impression on the body with red slip on the rim.

184

Mound Dwelling Traditions of the Mouhoun Bend

Another sample of cowry shells dated to the 7th - 9th century AD was collected from the Mema (Togola 1996).

8.6 The Evolution of the Mouhoun Bend settlement complexes

In summary, the cowry shells found in the Mouhoun bend settlement complexes, at Tora-Sira-Tomo-3-West, Tora-Sira-Tomo-13, and Kirikongo Mound 3, all dated to the early part of the second millennium AD point to the southward expansion of this commodity that may have been introduced from northeastern Africa, via the Garamantes of the Fezzan (Liverani 2000a, b, 2005). It is particularly difficult, partly because of the small sample size and the versatility in the use of cowry shells, to pinpoint what these items were used for in the Mouhoun bend in the first half of the 2nd millennium AD. At Kissi 3, the cowry shells were part of the deceased dressing code. Were these caps or headbands worn as part of these individuals’ garments in their daily interaction with the other community members or were they especially made for the performance of funerary and mortuary performance? The monetary value the cowry shells happened to have acquired later after the massive influx implemented by the Portuguese from the Maldive Islands does not have to be pushed backward in time. Cowry-shells still have ceremonial and ritual “values” in many present-day African societies. They were probably not a currency at the beginning of the 2nd millennium AD. Interesting and exciting as the find may be, eight cowry shells in the fill above infants and children burials at Kirikongo mound 3 (Dueppen 2008: 252) can hardly be considered as convincing evidence for social ranking and ascribed status.

Groups of Late Stone Age foragers settled along the Mouhoun River during the later part of the Holocene. Very few LSA sites have been recorded so far and their presence is indicated by small scatters of stone tools, destroyed, rearranged, or exposed by erosion agencies. The Late Stone Age “tradition” documented in the Mouhoun bend lacked pottery, providing no evidence for comparisons with the later mound-dwellers material culture. In the middle of the first millennium BC, small groups of iron-using mixed farmers and fishing communities started to settle in the Mouhoun bend. Were they of the earlier LSA foragers’ descent who adopted mixed farming and iron technology or new immigrants? Their presence is recorded at Tora-Sira-Tomo, Kerebe-Sira-Tomo, and possibly at Diekono. At Tora-Sira-Tomo, they started the exploitation of the quarry (TST-1), smelted iron ore at Tora-Sira-Tomo-2, with a domestic unit and blacksmith workshop at Tora-SiraTomo-3-East occupation I and II. At Kerebe-Sira-Tomo, they also started the exploitation of a quarry at KST-2 and built a karité oil “press“at KST-1A. The inhabited area was increased with the foundation of an additional domestic unit in the first half of the 1st millennium AD. The low-lying areas within the river bend and along the valley were settled first. The relative higher land, away from the river course, started to be settled at the beginning of the first millennium as shown by the foundation of the settlement of Kirikongo (Dueppen 2008). Three mounds, Mound I occupation 1, Mound III occupation 1, and Mound IV occupation 1 and 2, were then inhabited in this new complex. These small hamlets remained the social norm for many centuries. Settlement growth was slow but sustained during the second half of the 1st millennium AD. Two growth patterns leading to the formation of larger villages emerged at that time. Some mounds, like the one excavated at Diekono, Kerebe-Sira-Tomo, and Kirikongo grew in size. Others like Tora-Sira-Tomo spread out with numerous distinct mounds. The study area witnessed a phenomenal growth spur during the first few centuries of the 2nd millennium AD. All the tested settlement complexes became large autonomous, probably self-sustaining villages, leading to rivalry and conflict as will be discussed later. Craft specialization peaked and cultural items from elsewhere, like three-legged bowls and cowry shells, found their way to the Mouhoun Bend.

8.5.3 The Copper arm-rings Two copper arm-rings dated to the beginning of the second millennium AD were found concealed in a cache under the floor of occupation V at Tora-Sira-Tomo-3-West. While iron metallurgy is ubiquitous in the study area, no traces of copper working have been found in any of the tested sites. It is thus highly probable that the uncovered armrings were imported from elsewhere. The nearest sources of copper ore are found in the Eghazzer basin in Niger and the region of Nioro du Sahel in southwestern Mali. Copper, alloyed-copper, and brass were already known in West Africa and circulated through intricate long distance exchange networks linked with the Trans-Saharan trade. In summary, three-legged bowls, cowry shells, and copper arm-rings seem to have spread from the north to the south at the beginning of the second millennium AD. They could have been part of exchange packages that may have included semi-finished iron and iron artifacts as the northbound component of the inter-regional exchange system. More likely however, the presence of these categories of objects in the Mouhoun bend could have been the result of random individual’s decisions unrelated to any organized inter-regional trade network.

In the middle of the 2nd millennium AD, all of the four tested mounds from Kirikongo were still inhabited. Settlement size started to shrink in all the other tested sites. Diekono was left uninhabited. More than half of the twenty mounds of the Tora-Sira-Tomo / Gnambakouon-Sira-Tomo settlement complex were abandoned. And Kerebe-SiraTomo-4 was used as an ad-hoc burial place. The mounddwelling traditions collapsed in the middle of the second millennium AD. This collapse coincided with the expansion

185

Archaeology of West African Mound-Clusters

of the Marka, speakers of eastern Mande languages, who may have migrated from the core of Manden in the upper reaches of the Niger River. It is not known if this population movement was the cause or the consequence of the demise of the Mouhoun bend mound-dwelling traditions, traditions which lasted for more than 2000 years, from ca. 700/500 BC to 1600 AD.

Tomo mound. The targeting of newly founded habitation could have been a strategy to curtail the growth of ToraSira-Tomo. Two cases of burnt houses were recorded at Kerebe-SiraTomo. One at Kerebe-Sira-Tomo-3 occupation I, with a thick ash deposit and a large piece of burnt wooden log, and the other at Kerebe-Sira-Tomo-4 occupation II, where a large and complex habitation unit was completely burnt down. Both cases are dated to AD 1050 - 1380. Comparable cases of burnt houses dated to AD 1300-1450 were recorded at Kirikongo, in Mound III, level 8 and Mound IV, level 7 (Dueppen 2008).

Concluding Remarks Most of the study area is relatively flat. The locational model outlined in chapter 2 highlights the pattern of distribution of settlements in the landscape. The recorded settlement complexes are more or less evenly distributed. In site-catchment analysis terms, each village is surrounded by rings of cultivated fields, fallow zones, and bush (Chisholm 1979; Vita-Finzi & Higgs 1970). At their peak, during the first centuries of the 2nd millennium AD, each of the recorded settlement complexes was a large autonomous and self-sustaining village. Some, like Kerebe-Sira-Tomo and Kirikongo, were compact villages with a few outlying mounds and special purpose sites. Others, as was the case for Tora-Sira-Tomo/Gnambakouon-Sira-Tomo and Diekono, were spread out with a multiplicity of distinct mounds.

These events took place during a period of rapid growth, circumstances in which competition and rivalry can easily result in confrontation. The clustering of burnt down domestic units within a very short segment of the settlement history of the Mouhoun bend makes the warfare explanation the most parsimonious. 2 Evidence of violence and surgery Even if there is no direct correlation, the violence suggested by the burnt houses presented above is partly corroborated by evidence of traumatic injuries recorded on some of the deceased. Individual 18, a 9-11 years old child from the cemetery of Tora-Sira-Tomo-9, had multiple peri-mortem depressed cranial features made by a sharp object (Maes & Walker, Annex I). This battering was certainly the main cause of death.

The archaeological record does not point to any significant wealth differentials. The practice of grave-goods and burial offerings was too marginal to be significant. However, some individuals, male, female, children, and infants, were buried in cemeteries while others were put to rest in their domestic space. Such variations may have resulted from subtle differences in social statuses without durable material correlates. Demographic cohorts, be they agesets or other kinds of associations, were very likely key structural components in the organization of these villages’ communities. Cycles of initiations and rites of passage from childhood to adulthood, adulthood to elder-hood, and elder-hood to ancestor-hood, were the key institutions of the villages communities social life. In such social framework, one can expect cohorts of initiated young males to be trained to protect and defend their respective communities. The accelerated growth of the beginning of the 2nd millennium AD triggered peer-village rivalry and violence, signs of which are found in the archaeological record.

An adult female, 35 - 50 years old from Tora-Sira-Tomo-7 had a series of four well healed cranial fractures on the frontal, left parietal, and occipital. A male subject from the same site, 45+ years old in age, 1.95 m tall, is particularly interesting. He had a well healed circular defect in the occipital and presented multiple trephinations with no signs of healing on the frontal, left and right parietals. In summary, he had recovered from a first blow but the second one was fatal. It is worth emphasizing that both adult individuals present comparable patterns of blows. They were hit on the frontal, occipital, left and right parietals, as if the aggressors were well trained fighters. The individual with evidence of trephination found at TST7 is not dated directly. It is however very likely that he belonged to the phase III (AD 1000 - 1200) when the area witnessed a significant growth acceleration. The surgery may have taken place after a violent raid. The new case from Tora-Sira-Tomo-7, even if comparatively recent, is the first ever documented in West Africa. Traces of trephination are known in different part of the Old World (Crubezy et al 2001). The oldest known cases, recorded at Taforalt and Afalou-bou-Rhummel in North Africa, Muge in Portugal, Zawi Chemi in Iraq, and Vasiliyevka II and Vovnigi in Ukraine, date to the Epipalaeolithic and Mesolithic periods. Early Neolithic cases are also documented in Europe, at Vedrovice in Czech Republic, Pendimoun and Ensisheim

1 Fire: accidents or acts of warfare? A number of instances of burnt houses have been recorded in some of the excavated mounds. Such instances were found at Tora-Sira-Tomo-3-West in occupation I and II, both dated to AD 1000-1150. All five totally and partially exposed round huts from occupation I were completely burnt down. The hut exposed in occupation II was also burnt, with a large wooden log, probably part of the supporting central pole, found on the floor. In both cases, the burnt installations were part of the initial extension of habitation to the western confine of the large Tora-Sira-

186

Mound Dwelling Traditions of the Mouhoun Bend

in France, Trasano and Grotta Patrizi in Italy (Crubezy et al 2001: 420). One can but speculate on the origins of the surgical practices documented at Tora-Sira-Tomo. More systematic research has to be conducted on West African skeletal remains to collect more information.

and cloths dyers, as well as part-time warriors, and healers [surgeons] as suggested by the data on attempted trephination. The government of such communities is generally based on the participation of all the family heads and operates in the context of deliberative assemblies. A “council of elders” can be a complementary or alternative model depending on circumstances, and the nominal ruling individual is simply the primus inter pares - the first among equals -.

In summary, there are convincing sets of evidence pointing to conflict and inter-village warfare during the first half of the second millennium AD in the Mouhoun Bend. The nature and characteristics of the kind of warfare which may have developed in the area during this phase of accelerated growth are difficult to decipher. The tactic involved may have consisted of surprise raiding and counter-raiding with the aim of seizing supplies and hostages.

The Mouhoun Bend peer-villages were autonomous and self-sustaining but not autarkic. Political and matrimonial alliances allowed for the circulation of people, goods, and ideas. A general compatibility of “worldviews” is suggested by the strong coherence of mortuary practices. Despite variations from mound to mound and settlement complex to settlement complex, pottery decoration techniques and syntax suggest a shared cultural framework. If the villages were constituted of exogamic groups with viri-local postmarital residence, young women and among them potters from generation to generation were constantly redistributed among the peer-villages throughout the different matrimonial networks. Alliances may have shifted; former allies becoming enemies; and former enemies becoming allies, depending on circumstances. Peace agreements could have been negotiated by iron-smelters/blacksmiths.

3 The Dynamics of Peer-Village interaction Each of the large autonomous and self-sustaining villages from the study area controlled a near catchment zone of approximately 5 kilometers in diameter where cultivated fields and grazing lands are located (Chisholm 1979; Vita-Finzi & Higgs 1970). This prime economic zone is surrounded by an outer-ring with wild resources, plants as well as animal, accessible to any one. In situation of accelerated growth, more land is needed to feed the growing population, triggering an outward movement toward the appropriation of the no-man’s land previously accessible to all the communities. This trend combined with rapid soil exhaustion and longer fallow periods could have been the source of peer-villages conflicts.

The speculative character of this conclusion is intentional. It is based on the findings of decades of ethnological research on so-called a-cephalous societies. Such social organizations are extremely resilient and robust of their own. The mound-dwellers of the Mouhoun Bend developed an original socio-political system which, for approximately two centuries, from AD 1200 to 1400, existed on the periphery of the powerful Empire of Mali.

Each village had its autonomous system of government, with a horizontally differentiated society. These mixed farming communities included a number of craft specialists, potters, iron-smelters and blacksmiths, masons, weavers

187

References Connah, G. 1981 Three Thousand Years in Africa. Cambridge; Cambridge University Press. Connah, G. 1984 An archaeological exploration in southern Bornu. The African Archaeological Review 2: 153-171. Cook, A. C., J.R. Southon, and J. Wadsworth 2003 Using Radiocarbon Dating to establish the Age of Iron-based Artifacts. A Hypertext Enhanced Article. http://www. tms.org/pubs/journals/JOM/0305/Cook-0305.html. Crubezy, E., J. Bruzek, J. Guilaine, E. Cunha, D. Rouge, J. Jelinek 2001 The Antiquity of cranial surgery in Europe and the Mediterranean basin. Compte Rendus de l’Academie des Sciences: Sciences de la Terre et des Planetes 332: 417-423. D’ Andrea, A.C., S. Kahlheber, A.L. Logan, and D.J. Watson 2007 Early domesticated cowpea (Vigna unguiculata) fromCentral Ghana. Antiquity 81: 686-698. Devisse, J. Editor 1983 Tegdaoust III: Recherches sur Aoudaghost: Campagnes 1960-1965, Enquetes generales. Paris; Editions Recherches sur les Civilisations. Diallo, M. 2000 Le marka dans l’ensemble dialectal Mandingue. In Kulturentwicklung und Sprach-geschichte im Naturraum Westafrikanische Savanne. Berichte des Sonderforshungsbereichs 268. Band 14. Pp. 379-384. Frankfurt am Main. Dueppen, S.A. 2008 Re-Inventing Equality: The archaeology of Kirikongo, Burkina Faso. PhD Dissertation, Ann Arbor; the University of Michigan. Fikipowiak, W. 1979 Etudes archeologiques sur la capitale medieval du Mali. Szczecinie; Museum Narodowe. Gado, B. 1993 Un “Village des morts” a Bura en republique du Niger. In J. Dveisse, editor. Vallees du Niger. Pp. 365-75. Paris; Reunions de Musees Nationaux. Gado, B. 2004 Les systemes des sites a statuaire en terre cuite et en pierre dans la vallee du Moyen Niger entre le Gorouol et la Mekrou. In A. Bazzana and H. Bocoum, editors. Du Nord au sud du Sahara: 50 Ans d’Archeologie Francaise. Pp. 154 - 181. Paris; Editions Sepia. Gjerdrum, T., A.F.C. Holl, K.C. Maes, and P.L. Walker 2002 West African trephination: Skeletal evidence from Tora Sira Tomo, Burkina Faso. Poster presented at the 14th Biennial European Members Meeting of the Paleopathology Association, Aug. 28-31, Coimbra, Portugal. Grebenart, D. 1985 La Region d’In Gall – Tegidda n Tesemt (Niger) II: Le Neolithique Final et les Debuts de la Metallurgie. Niamey, Etudes Nigeriennes 49. Haaland, R. and P. Shinnie (editors) 1985 Aftican Iron Working: Ancient and Traditional. Bergen, Norwegian University Press.

Albert, K.-D., M. Hallier, S. Kahlheber, and C. Pelzer 2000 Montee et Abandon des collines d’occupation de l’Age de Fer au Nord du Burkina Faso. In Kulturentwicklung und Sprachgeschichte im Naturraum Westafrikanische Savanne. Berichte des Sonderforshungsbereichs 268. Band 14. Pp. 335-351. Frankfurt am Main. Albert, K.-D. And S. Kahlheber 2002 Review of and outlook on an interdisciplinary Research project. In Kulturentwicklung und Sprachgeschichte im Naturraum Westafrikanische Savanne. Berichte des Sonderforshungs-bereichs 268. Band 17. Pp. 121-128. Frankfurt am Main. Aufderheide, A.C. and C. Rodríguez-Martin. 1998 The Cambridge Encyclopedia of Human Paleopathology. Cambridge; Cambridge University Press. Ballouche, A. 2001 Un diagramme polonium de la mare de Kissi (Oudalan, Burkina Faso). Nouveaux elements pour l’histoire anthropique de la vegetation sahelienne. In Kulturentwicklung und Sprachgeschichte im Naturraum West-afrikanische Savanne. Berichte des Sonderforshungsbereichs 268. Band 17. Pp. 129-135. Frankfurt am Main Bedaux, R. M.A. 1972 Tellem, Reconnaissance Archeologique d’une culture de l’Ouest Africain au Moyen Age, Recherches Architectoniques. Journal de la Societe des Africanistes 13 (2): 105-1985. Bedaux, R.M.A. 1980 The geographic distribution of footed bowls in the Upper and Middle Niger Region. In West African Culture Dynamics. Edited by B.K. Swartz and R.A. Dumett. Pp. 247-258. The hague/Paris; Mouton Publishers. Bedaux, R., K. MacDonald, A. Person, J. Polet, K. Sanogo, A Schmidt, and S. Sidibe 2001 The Dia Archaeological Project: rescuing cultural heritage in the inland Niger delta (Mali). Antiquity 75: 837-48. Berthier, S. 1997 Recherches Archeologiques sur la Capitale de l’Empire de Ghana. Oxford; British Archaeological Report. Buikstra, J.E. and D.H. Ubelaker (eds.). 1994 Standards for Data Collection from Human Skeletal Remains. Fayetteville: Arkansas Archaeological Research Series 44. Casey, J. 1998 The Ecology of Food Production in West Africa. In G. Connah, editor, Transformations in Africa. Pp. 46-70. London; Leicester University Press. Chisholm, M. 1979 Rural Settlement and Land Use: An Essay in Location. London; Hutchinson. Cohen, N.N. and G.J. Armelagos (eds.). 1984 Paleopathology at the Origins of Agriculture. New York: Academic Press.

188

References

Hallier, M. 2001 Recherches archeologiques sur le site d’Oursi. In Kulturentwicklung und Sprachgeschichte im Naturraum Westafrikanische Savanne. Berichte des Sonderfor-schungsbereichs 268. Band 17. Pp. 137-144. Frankfurt am Main. Henderson, J. 1987 Factors determining the state of preservation of human remains. In Death, Decay and Reconstruction: Approaches to Archaeology and Forensic Science, ed. A. Boddington, A.N. Garland & R.C. Janaway, pp.43-54. Manchester: Manchester University Press. Higgs. E.S. And C. Vita-Finzi 1970 Prehistoric Economy of the Mount Carmel Area of Palestine: Site-catchment Analysis. Proceedings of the Prehistoric Society 36: 1-37. Holl, A. 1993 Transition from Late Stone Age to Iron Age in the Sudano-Sahelian zone: A case study from the Perichadian Plain. In The Archaeology of Africa: Food, Metals and Towns. Edited by T. Shaw, P. Sinclair, B. Andah, and A. Okpoko, pp. 330-343. London, Routledge Holl, A.F.C. 1994 The Cemetery of Houlouf in Northern Cameroon (A.D. 1500 -1600): Fragments of a past social system. African Archaeological Review 12: 133-170. Holl, A. F. C. 2002 The Land of Houlouf: Genesis of a Chadic Polity ca 1900 BC - AD 1800. Ann Arbor, University of Michigan Museum of Anthropology. Holl, A.F.C. 2004 Holocene Saharans: An Anthropological Perspective. London/New York; Continuum. Holl, A.F.C. 2006 West African Early Towns: Archaeology of Households in urban landscapes. Ann Arbor; University of Michigan Museum of Anthropology. Holl, A.F.C. 2009a West African early metallurgies: New Evidence and old orthodoxy. Journal of World Prehistory. 22 (4): 415 - 438. Holl, A.F.C. 2009b Iron Metallurgy in West Africa: An Early Iron Smelting Site in the Mouhoun Bend, Burkina Faso. In Crossroads/Carrefour Sahel: Cultural and Technological Developments in the first millennium BC/ AD West Africa.. Edited by S. Magnavita, L. Kote, P. Breunig, and O. A. Ide. Pp. 59-68. Frankfurt; Journal of African Archaeology Monographs Series, vol. 2 Holl, A.F.C. and Lassina, K. 2003 Settlement patterns, food production and craft specialization in the Mouhoun Bend (NW Burkina Faso). West African Journal of Archaeology 30(1): 69-107. Insoll, T. 1996 Islam, Archaeology, and History: Gao Region (Mali) ca. AD 900-1250. Oxford; British Archaeological Reports. Insoll, T. 2000 Urbanism, Archaeology and Trade: Further observations on the Gao Region (Mali). Oxford; British Archaeological Reports. Irish, J.D. (1993) Biological affinities of late Pleistocene through modern African aboriginal populations: the dental evidence. Ph.D. dissertation, Arizona State University. Irish, J.D. 1997 Characteristic high- and low-frequency dental traits in sub-Saharan African populations. American Journal of Physical Anthropology 102:455467.

Irish, J.D. 1998 Ancestral dental traits in recent SubSaharan Africans and the origins of modern humans. Journal of Human Evolution 34:81-98. Kahlheber, S., K.-D. Albert, and A. Hohn 2001 A Contribution to the palaeoenvironment of the archaeological site of Oursi in north Burkina Faso. In Kulturentwicklung und Sprachgeschichte im Natur-raum Westafrikanische Savanne. Berichte des Sonderforshungsbereichs 268. Band 17. Pp. 145-159. Frankfurt am Main. Kiethega, J.B. 2006 La Metallurgie Lourde du Fer au Burkina Faso. Une Technologie a l’Epoque Precoloniale. Paris, Karthala. Kiethega, J.B. 2009 La Métallurgie lourde du fer au Burkina Faso. Paris; Karthala. Krogman, W.M. and M.Y. Isçan. 1986 The Human Skeleton in Forensic Medicine. Second Edition. Springfield, IL: C.C. Thomas. Larsen, C.S. 1997 Bioarchaeology: Interpreting Behavior from the Human Skeleton. Cambridge: Cambridge University Press. Lebeuf, J.P. 1969 Carte Archeologique des Abords du Lac Tchad. Paris; Editions du Centre National de la Recherche Scientifique. Lebeuf, J.P. 1981 Supplement a la Carte Archeologique des Abords du Lac Tchad. Paris; Editions du Centre National de la Recherche Scientifique. Levtzion, N. And J.F.P. Hopkins editors 1981 Corpus of Early Arabic Sources for West African History. Cambridge; Cambridge University Press. Liverani, M. 2000a The Garamantes: A Fresh Approach. Libyan Studies 31: 17-28. Liverani, M. 2000b Looking for the southern Frontier of the Garamantes. Sahara 12: 31-44. Liverani, M. 2005 editor Aghram Nadharif: Tha Barkat Oasis (Sha’abiya of Ghat, Libyan Sahara) in Garamantian Times. Arid Zone Archaeology Monograph 5. Firenze; Edizioni All’Insegna del Giglio. Okafor, E.E. 1993 New evidence on early iron-smelting from southeastern Nigeria. In The Archaeology of Africa: Food, Metals and Towns. Edited by T. Shaw, P. Sinclair, B. Andah, and A. Okpoko, pp. 432-448. London, Routledge. Maes, K.C., J.D. Irish, A.F.C. Holl, P.L. Walker, and G.J. Armelagos. 2004 Preliminary analysis of dental morphology and identity of an early iron producing population in the Mouhoun Bend, Burkina Faso. Poster presented at the 73rd Annual Meetings of the American Assocation of Physical Anthropologists, April 14-17, 2004. Magnavita, S. 2004 Das Grab des Kriegers. In Leben in West Afrika. Pp. 44-49. SFB 268, J.W. Goethe Universitat, Frankfurt a/Main/ Deutsche Forschungs-gemeinschaft. Magnavita, S. 2009 Sahelian Crossroads: Some aspects of the Iron Age sites of Kissi, Burkina Faso. In Crossroads/ Carrefour Sahel. Edited by S. Magnavita, L. Kote, P. Breunig, and O. A. Ide, pp. 79-97. Journal of African Archaeology Monograph Series 2. Frankfurt; Africa Magna Verlag.

189

Archaeology of West African Mound-Clusters

Natur-raum Westafrikanische Savanne. Berichte des Sonderforshungsbereichs 268. Pp. 367-373. Frankfurt am Main. Phenice, T.W. 1969 A newly developed visual method of sexing the Os Pubis. American Journal of Physical Anthropology 30:297-301. Polet, J. 2004 L’emergence de la ville en Afrique Subsaharienne. In A Bazzana and H. Bocoum, editors. Du Nord au Sud du Sahara: Cinquante Ans d’Archeologie Francaise. Pp. 249-259. Paris; Editions Sepia. Raimbault, M. and K. Sanogo (editors) 1991 Recherches Archeologiques au Mali. Paris; ACCT – Karthala. Roberts, C. and K. Manchester 1995 Archaeology of Disease. Second Edition. Ithaca, NY: Cornell University Press. Rose, J.C., M.K. Marks, and L.L. Tieszen 1991 Bioarchaeology and subsistence in the central and lower portions of the Mississippi Valley. In What Mean These Bones?, eds. M. Powell, P. Bridges, and A. Mires, pp. 7-21. Tuscaloosa: University of Alabama Press. Sanogo, K. 2004 La Problematique des Modes d’inhumation dans la Boucle du Niger (Mali): du Lieutenant Deplanes a nos jours. In A Bazzana and H. Bocoum, editors. Du Nord au Sud du Sahara: Cinquante Ans d’Archeologie Francaise. Pp. 189-194. Paris; Editions Sepia. Scott, E.C. 1979 Dental wear scoring technique. American Journal of Physical Anthropology 51:213-218. Schmidt, P. ed. 1996 The Culture and Technology of African Iron Age Production. Gainesville, University of Florida Press. Togola, T. 1993 Archaeological Investigations of Iron Age sites in the Mema Region, Mali (West Africa). PhD Dissertation. Rice University. Ann Arbor; University Microfilms International. Togola, T. 1996 Iron Age Occupation in the Méma region, Mali. African Archaeological Review 13: 91-110. Treinen-Claustre, F. 1982 Sahara et sahel a l’Age du Fer: Borkou, Tchad. Paris; Societe des Africanistes. Turner, C.G., C. Nichol, and G. Scott. 1991 Scoring procedures for key morphological traits of the Permanent Dentition: The Arizona State University Dental Anthropology System. In Advances in Dental Anthropology, eds. M. Kelley and C.S. Larsen, pp. 1331. New York: Wiley-Liss. Waldron, T. 1987b The relative survival of the human skeleton: Implications for paleopathology. In Death, Decay and Reconstruction: Approaches to Archaeology and Forensic Science, ed. A. Boddington, A.N. Garland & R.C. Janaway, pp.43-54. Manchester: Manchester University Press. Walker, P.L., J.R. Johnson, and P.M. Lambert. 1988 Age and sex biases in the preservation of human remains. American Journal of Physical Anthropology 80: 313323.

Margetts, E.L. 1967 Trephination of the skull by the medicine-men of primitive cultures with particular reference to present-day native East African practice. In Diseases in Antiquity. Edited by. D. Brothwell & A.T. Sandison, pp. 673-701. Springfield, IL: C.C. Thomas. Martinelli, B. 2008 Patrimoine siderurgique traditionel en Centrafrique.Revue Centre Africain d’Anthro-pologie Volume I. McIntosh, R.J. 1999 Peoples of the Middle Niger. Oxford; Blackwell. McIntosh. S.K editor 1995 Excavations at Jenne-Jeno, Hambarketolo, and Kaniana (Inland Niger Delta, Mali), the 1981 field season. Berkeley; University of California Press. McIntosh. S.K. 1999 Modeling political organization in large scale settlement clusters: A case study from the Inland Niger Delta. In Beyond Chiefdoms: Pathways to Complexity in Africa. Edited by S.K. McIntosh, pp. 6679; Cambridge, Cambridge University Press. McIntosh, R. J. And S.K. McIntosh 1980 Prehistoric Investigations at Jenne, Mali. Oxford, British Archaeological Report. McIntosh, R. J. And S.K. McIntosh 1984 The Early City in West Africa: Toward an Understanding. The African Archaeological Review 2: 73-98. McIntosh, R.J. And S.K. McIntosh 1993 Cities without citadels: Understanding urban origins along the Middle Niger. In T. Shaw, P. Sinclair, B.W. Andah, and A. Okpoko, editors. The Archaeology of Africa: Food, Metals, and Towns. Pp. 622-641. London and New York; Routledge.s Mauny, R. 1961 Tableau geographique de l’Afrique de l’Ouest au Moyen-Age. Dakar; Institut Fondamental de l’Afrique Noire. Millogo, A.K. And L. Kote 2000 Recherches Archeologiques a Gandefabou. In Kulturentwicklung und Sprachgeschichte im Naturraum West-afrikanische Savanne. Berichte des Sonder-forshungsbereichs 268. Pp. 353-365. Frankfurt am Main Neumann, K. 2000 Le Sahel du Burkina Faso: Paleoenvironnement et developpement culturel. In Kulturentwicklung und Sprachgeschichte im Natur-raum Westafrikanische Savanne. Berichte des Sonderforshungsbereichs 268. Pp. 323-325. Frankfurt am Main. Neumann, K., P. Breunig, and S. Kahlheber 2000 Early Food Production in the Sahel o Burkina Faso. In Kulturentwicklung und Sprachgeschichte im Naturraum Westafrikanische Savanne. Berichte des Sonderforshungsbereichs 268. Pp. 327-334. Frankfurt am Main. Pelzer, C. And S. Magnavita-Santos 2000 La Necropole de Kissi et ses implications historiquea. In Kulturentwicklung und Sprachgeschichte im

190

Annex 1: Mouhoun Bend Skeleton Summaries K. C. Maes and P. L. Walker

Skeleton 1

Age: Middle-Old Adult (45+ years) Sex: Male—pelvic, cranial, and mandibular characteristics Preservation: 14% (Fair) Condition: Poor Dental Pathology: Malocclusion (overbite) causing abnormal wear patterns; heavy calculus; possible abscess (RM1); carious lesion (RM2); antemortem loss of RI2; crowding of anterior mandibular dentition; moderate to heavy attrition Skeletal Pathology: Cultural modification: multiple trephinations with no signs of healing on frontal, left and right parietals; Other: well-healed circular defect in occipital Taphonomic Processes: Termite activity; sediment adhering; possible rodent gnawing Skeletal Non-metric traits: None noted Associations: None noted

Site: TST-7 Occupation (Level): I Burial: 1 Year excavated: 1999 Age: Subadult (15-17 years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Poor (Maxilla and mandible = fair) Dental Pathology: Calculus flecks; carious lesion (RM1) Skeletal Pathology: None noted Taphonomic Processes: Rodent gnawing or termite activity; sediment adhering Skeletal Non-metric traits: None noted Associations: None noted Skeleton 2

Skeleton 4

Site: TST-7 Occupation (Level): I Burial: 2 Year excavated: 1999 Age: Middle Adult (35-50 years) Sex: Female—cranial and mandibular characteristics Preservation: 7% (Poor) Condition: Fair Dental Pathology: Heavy calculus; moderate to heavy attrition; 90 degree rotation of LC1 Skeletal Pathology: Trauma: multiple (four) well-healed depressed cranial fractures on frontal, left parietal, and occipital; Non-specific Stress Indicator: possible healed porotic hyperostosis Taphonomic Processes: Possible termite activity, especially on cranium Skeletal Non-metric traits: None noted Associations: Ceramic pottery fragments

Site: TST-7 Occupation (Level): III Burial: N/A Year excavated: 1999 Age: Subadult (5-7 years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Fair Dental Pathology: Slight to moderate deciduous attrition Skeletal Pathology: None noted Taphonomic Processes: Termite activity Skeletal Non-metric traits: None noted Associations: Goat or Sheep teeth fragments Skeleton 5 Site: TST-5 Occupation (Level): N/A Burial: N/A Year excavated: 1999 Age: Middle-Old Adult (35+ years) Sex: Female—cranial and mandibular characteristics Preservation: 36% (Fair) Condition: Fair

Skeleton 3 Site: TST-7 Occupation (Level): I Burial: 3 Year excavated: 1999

191

Archaeology of West African Mound-Clusters

Dental Pathology: Calculus flecks; carious lesions (LM1 and RM1); moderate to heavy attrition; possible antemortem tooth loss with alveolar resorption (LM2); crowding of anterior mandibular dentition Skeletal Pathology: None noted Taphonomic Processes: Rodent gnawing (on frontal); termite activity Skeletal Non-metric traits: None noted Associations: None noted

Preservation: 0% (Poor) Condition: Poor Dental Pathology: No teeth Skeletal Pathology: Joint Disease: Osteophyte growth and porosity on un-sided proximal ulna. Taphonomic Processes: None noted Skeletal Non-metric traits: None noted Associations: Thin, white shell fragments Skeleton 9

Skeleton 6 Site: TST-8 Occupation (Level): N/A Burial: N/A Year excavated: 1999 Age: Adult-Other (18+ years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Poor Dental Pathology: No teeth Skeletal Pathology: None noted Taphonomic Processes: None noted Skeletal Non-metric traits: None noted Associations: None noted

Site: KST-5 Occupation (Level): 80-100cm Burial: 1 Year excavated: 1999 Age: Adult-Other (18+ years) Sex: Probable Female—pelvic characteristic and lack of robusticity of available long bone material Preservation: 0% (Poor) Condition: Poor Dental Pathology: No teeth Skeletal Pathology: None noted Taphonomic Processes: None noted Skeletal Non-metric traits: None noted Associations: None noted

Skeleton 7

Skeleton 10

Site: KST-4 Occupation (Level): II Burial: N/A Year excavated: 1999 Age: Old Adult (50+ years) Sex: Male—pelvic, cranial, and mandibular characteristics Preservation: 43% (Good) Condition: Good Dental Pathology: Hypoplastic lesions on all third molars; carious lesions (LM2 and RM2); moderate calculus; light to moderate attrition; alveolar abscesses (RP1 and RM1); abscesses into maxillary sinus (RM1 and RM2) Skeletal Pathology: Joint Disease: enthesophyte on left patella; minor osteophyte growth on axis (C2) Taphonomic Processes: Possible termite activity Skeletal Non-metric traits: Septal aperture on left humerus; supracondyloid tubercle on right humerus; wormian bone in right lambdoidal suture Associations: None noted

Site: KST-5 Occupation (Level): 100-120cm Burial: 2 Year excavated: 1999 Age: Adult-Other (18+ years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Poor Dental Pathology: No teeth Skeletal Pathology: None noted Taphonomic Processes: None noted Skeletal Non-metric traits: None noted Associations: None noted Skeleton 11 Site: DIE-2 Occupation (Level): N/A Burial: N/A Year excavated: 1999 Age: Middle Adult (35-50 years) Sex: Probable Male—pelvic, cranial, and mandibular characteristics Preservation: 64% (Good) Condition: Good Dental Pathology: Possible hypoplastic lesions on RM2 and LM2; heavy calculus; crowding of anterior mandibular dentition; moderate attrition

Skeleton 8 Site: KST-4 Occupation (Level): 20-40cm Burial: N/A Year excavated: 1999 Age: Adult-Other (18+ years) Sex: Indeterminate

192

Annex 1

Skeletal Pathology: Joint disease: minor to moderate osteophyte growth on C4 through T2, and T12; minor to heavy osteophyte growth on two lumbar vertebrae; minor osteophyte growth on left carpals, right carpals, left pedal phalanges, right pedal phalanges, and right ulna (distal articular surface) Taphonomic Processes: Sediment adhering Skeletal Non-metric traits: None noted Associations: None noted

Skeletal Pathology: Non-specific infection: periostitis on left fibula Taphonomic Processes: Possible termite activity; water damage Skeletal Non-metric traits: None noted Associations: None noted Skeleton 15

Site: TST-9 Occupation (Level): N/A Burial: 1 Year excavated: 1999 Age: Young-Middle Adult (25-50 years) Sex: Probable male—mandibular characteristics Preservation: 0% (Poor) Condition: Poor Dental Pathology: Calculus flecks; antemortem loss of LI1; moderate attrition Skeletal Pathology: None noted Taphonomic Processes: Sediment adhering Non-metric traits: Dental: Peg-shaped upper lateral incisor (unsided) Associations: Charcoal pieces

Site: KST-4 Occupation (Level): II Burial: E.8 Year excavated: 2000 Age: Young-Middle Adult (25-50 years) Sex: Female—cranial and mandibular characteristics Preservation: 21% (Poor) Condition: Poor Dental Pathology: Supernumerary tooth (between RP3 and RP4); carious lesions (supernumerary tooth, RM3, RM1, RM2); Light to heavy calculus; rotation of LM2; crowding of anterior mandibular dentition; moderate to heavy attrition; possible occupational wear on upper and lower central incisors Skeletal Pathology: None noted Taphonomic Processes: Termite activity Non-metric traits: Dental: supernumerary tooth Associations: None noted

Skeleton 13

Skeleton 16

Site: TST-9 Occupation (Level): N/A Burial: 2 Year excavated: 1999 Age: Adult-Other (18+ years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Poor Dental Pathology: No teeth Skeletal Pathology: None noted Taphonomic Processes: Termite activity Skeletal Non-metric traits: None noted Associations: None noted

Site: KST-4 Occupation (Level): II Burial: “Feature 4” Year excavated: 2000 Age: Adult-Other (18+ years) Sex: Probable male—robust muscle attachments on tibia and fibula Preservation: 14% (Poor) Condition: Good Dental Pathology: No teeth Skeletal Pathology: None noted Taphonomic Processes: None noted Skeletal Non-metric traits: None noted Associations: None noted

Skeleton 12

Skeleton 14

Skeleton 17

Site: KST-4 Occupation (Level): N/A Burial: E.2 Year excavated: 2000 Age: Young Adult (20-25 years) Sex: Probable male cranial characteristic and robus-ticity of long bone elements Preservation: 0% (Poor) Condition: Poor Dental Pathology: Moderate attrition

Site: KST-4 Occupation (Level): IV Burial: N/A Year excavated: 2000 Age: Adult-Other (18+ years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Poor Dental Pathology: No teeth

193

Archaeology of West African Mound-Clusters

Skeletal Pathology: None noted Taphonomic Processes: Termite activity Skeletal Non-metric traits: None noted Associations: None noted

Year excavated: 2000 Age: Subadult (12-14 years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Poor Dental Pathology: slight attrition Skeletal Pathology: Other: possible neoplastic condition of non-specific infection—endocranial defects on occipital associated with an irregular-shaped, free-floating ossified mass Taphonomic Processes: Probable post-depositional missing of skeletal/dental elements; possible termite activity Skeletal Non-metric traits: None noted Associations: None noted

Skeleton 18 Site: TST-9 Occupation (Level): I Burial: Est. 18 Year excavated: 2000 Age: Subadult (9-11 years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Fair Dental Pathology: moderate attrition on deciduous teeth Skeletal Pathology: Trauma: Possible multiple perimortem depressed cranial fractures (sharp object) on right parietal Taphonomic Processes: Possible rodent gnawing; excavation scars; post-depositional breakage of cranium, possible due to soil pressure Skeletal Non-metric traits: None noted Associations: None noted

Skeleton 20 Site: TST-9 Occupation (Level): I Burial: (Mixed elements sorted from Est. 20) Year excavated: 2000 Age: Adult-Other (18+ years) Sex: Indeterminate Preservation: 0% (Poor) Condition: Poor Dental Pathology: moderate attrition Skeletal Pathology: No skeletal material Taphonomic Processes: None noted Skeletal Non-metric traits: None noted Associations: None noted

Skeleton 19 Site: TST-9 Occupation (Level): I Burial: Est. 20

194

Annex 2: Radiocarbon Dates from the Mouhoun Bend Archaeological Project

ISGS NO

SAMPLE

C-14 AGE (RCYBP)

Cal BC/AD (1 sigma)

Cal BC/AD (2 sigma)

KST-1 (20 cm)

2050+/-70

- 9.7

167 BC – 47 AD

348 BC – 117 AD

KST-2 (20 cm) KST-2 (60 cm) KST-2 (140 cm)

930+/-70 880+/-70 1660+/-70

- 25.3 - 25.7 - 25.5

1021 – 1209 AD 1037 – 1241 AD 261 – 434 AD

983 – 1260 AD 1018 – 1281 AD 237 – 556 AD

KST-3 (20-40 cm) KST-3 (40-60) KST-3 (60-80) KST-3 (80-100) KST-3 (100-120)

800+/-70 770+/-70 800+/-70 770+/-70 790+/-70

- 25.0 - 24.7 - 25.5 - 25.7 - 26.6

1165 – 1283 AD 1215 – 1291 AD 1165 – 1283 AD 1215 – 1291 AD 1191 – 1284 AD

1039 – 1376 AD 1070 – 1385 AD 1039 – 1376 AD 1070 – 1385 AD 1042 – 1379 AD

KST-4 (0-20) KST-4 (40) KST-4 (50) KST-4 (40-60) KST-4 (60) KST-4 (75-80) KST-4 (100)

710+/-70 550+/-70 660+/-70 1830+/-50 720+/-70 790+/-70 720+/-70

- 25.1 - 26.2 - 24.6 - 26.8 - 25.2 - 24.7

1262 – 1382 AD 1313 – 1435 AD 1282 – 1396 AD 123 – 235 AD 1259 – 1379 AD 1191 – 1284 AD 1259 – 1379 AD

1195 – 1401 AD 1292 – 1453 AD 1244 – 1418 AD 67 – 291 AD*1 1191 – 1399 AD 1042 – 1379 AD 1191 – 1399 AD

KST-5(0-20) KST-5(40-60) KST-5(60-80)

810+/-70 850+/-70 840+/-70

- 25.4 - 25.5 - 25.5

1163 – 1281 AD 1061 – 1263 AD 1070 – 1274 AD

1037 – 1298 AD 1022 – 1288 AD 1024 – 1291 AD

TST-1(50)

2360+/-70

- 26.7

501 – 386 BC

761 – 212 BC

TST-3 E (20-40) TST-3 E (80-100) TST-3 E(120-140) TST-3 W(20-40) TST-3 W(20-40) TST-3 W(40-60) TST-3 W(60-80) TST-3 W(100) TST-3 W(120)

460+/- 70 780+/- 70 710+/- 70 790+/- 70 740+/- 40 740+/- 70 880+/- 70 960+/- 70 980+/- 70

- 22.5 - 26.4 - 25.8 - 24.8 - 24.6 - 25.5 - 25.4 - 24.9

1412 – 1477 AD 1211 – 1287 AD 1262 – 1382 AD 1191 – 1284 AD 1212 – 1300 AD 1223 – 1298 AD 1037 – 1241 AD 1003 – 1162 AD 997 – 1158 AD

1326 – 1631 AD 1059 – 1382 AD 1195 – 1401 AD 1042 – 1379 AD 1220 – 1380 AD*2 1163 – 1393 AD 1018 – 1281 AD 904 – 1222 AD 899 – 1216 AD

Kerebe-Sira-Tomo KST-1 1 - 3941 KST-2 2 - 3937 3 - 3932 4 - 3936 KST-3 5 - 4302 6 - 4310 7 - 4311 8 - 4321 9 - 4319 KST-4 10 - 4340 11 - 4343 12 - 4320 13 – LLNL 14 - 4341 15 - 4344 16 - 4345 KST-5 17 - 4346 18 - 4347 19 - 4348 Tora Sira Tomo TST 1 20 - 4349 TST 3 21 - 3927 22 - 3931 23 - 3928 24 - 4586 25 – LLNL 26 - 4587 27 - 4588 29 - 4585 30 - 4590

195

Archaeology of West African Mound-Clusters

ISGS NO

SAMPLE

C-14 AGE (RCYBP)

Cal BC/AD (1 sigma)

Cal BC/AD (2 sigma)

TST-4 (30)

570+/- 70

- 25.7

1302 – 1426 AD

1286 – 1445 AD

TST-5 (20-40) TST-5 (80) TST-5(80-100)

400+/-70 440+/-70 560+/-70

- 24.8 - 25.1 - 24.6

1436 – 1627 AD 1420 – 1486 AD 1304 – 1431 AD

1408 – 1649 AD 1331 – 1637 AD 1289 – 1449 AD

TST-6 (40-60) TST-6 (80-100) TST-6(100-120) TST-6(120-140) TST-6(140-160)

630+/-90 770+/-70 820+/-70 950+/-70 960+/-70

- 24.7 - 24.9 - 25.4 - 25.6 - 24.2

1284 – 1409 AD 1215 – 1291 AD 1161 – 1279 AD 1018 – 1183 AD 1003 – 1162 AD

1223 – 1441 AD 1070 – 1385 AD 1033 – 1296 AD 978 – 1243 AD 904 – 1222 AD

TST-9 SE (40)

330+/- 70

- 25.8

1467 – 1649 AD

1436 – 1945 AD

TST-10 (20-40)

700+/- 70

- 25.2

1268 – 1385 AD

1214 – 1404 AD

TST-12 (40) TST-12(40-60)

720+/- 70 570+/- 30

- 25.7 -

1259 – 1379 AD 1280 – 1400 AD

1191 -1399 AD 1300 – 1430 AD*3

TST-13 (00-20) TST-13 (20-40) TST-13 (60)

570+/- 70 830+/- 70 820+/- 70

- 25.6 - 25.9 - 25.2

1302 – 1426 AD 1160 – 1277 AD 1161– 1279 AD

1286 - 1445 AD 1207 – 1294 AD 1033 – 1296 AD

TST-14(20-40) TST-14 (40-60)

870+/- 70 670+/- 70

- 25.5 - 25.7

1040 – 1256 AD 1280 – 1393 AD

1019 – 1283 AD 1223 – 1413 AD

TST-15 (20) TST-15 (40)

660+/- 70 820+/- 70

- 24.4 - 25.9

1282 – 1396 AD 1161 – 1279 AD

1244 – 1418 AD 1033 – 1296 AD

TST-16 (20) TST-16 (60)

980+/- 70 670+/- 70

- 25.1 - 25.8

997 – 1158 AD 1280 – 1393 AD

899 – 1216 AD 1223 – 1413 AD

TST-17 (60)

650+/- 70

- 26.1

1284 – 1398 AD

1258 – 1423 AD

DIE-1(100-120) DIE-1(130)

1380+/-70 1430+/-70

-25.6 -26.0

618 – 687 AD 561 – 662 AD

541 – 777 AD 440 – 759 AD

DIE-2(0-20) DIE-2 (20-40) DIE-2 (40-60) DIE-2 (60-80) DIE-2 (80-100) DIE-2(100-120)

940+/-70 930+/-70 960+/-70 960+/-70 950+/-70 1020+/-70

-25.0 -24.9 -25.2 -26.1 -25.4 -25.5

1020 – 1205 AD 1021 – 1209 AD 1003 – 1162 AD 1003 – 1162 AD 1018 – 1183 AD 979 – 1148 AD

981 – 1256 AD 983 – 1260 AD 904 – 1222 AD 904 – 1222 AD 978 – 1243 AD 889 – 1185 AD

TST 4 31 - 4591 TST 5 32 - 4350 33 - 4356 34 - 4357 TST 6 35 - 4358 36 - 4361 37 - 4360 38 - 4359 39 - 4362 TST 9 40 - 4593 TST 10 41 - 4595 TST 12 42 - 4596 43 – LLNL TST 13 44 - 4592 45 - 4594 46 - 4597 TST 14 47 - 4598 48 - 4599 TST 15 49 - 4600 50 - 4601 TST 16 51 - 4603 52 - 4604 TST 17 53 - 4605 Diekono DIE 1 54 - 4295 55 - 4296 DIE 2 56 - 4281 57 - 4282 58 - 4283 59 - 4284 60 - 4293 61 - 4294

Key: LLNL = Lawrence Livermore National Laboratory: * Radiocarbon dates from iron objects processed by Dr. Andrea C. Cook at Lawrence Livermore National Laboratory, UC Berkeley. *1 = from an iron ring *2 = from a small iron spear *3 = from a large iron spear

196